IDP Library

This IDP Library (flows_sdk.implementations) includes the latest performant blocks and functions for building intelligent document processing (IDP).

The IDP Library contains:

  • Processing Blocks, which are provided as convenient Python subclasses for plug-and-play use.

  • Trigger and Output Blocks, which require use of base classes.

  • Helper functions for ease of development.

Note

Most solutions will also require use of base classes, such as Flow or CodeBlock, found in the Source Documentation.

V39.1

Processing Blocks (v39.1)

Core Blocks are powerful processing blocks necessary to build intelligent document processing solutions on Hyperscience. They include both manual and machine processing blocks.

Note

Classes in the idp_v39_1 implementation have parameters with ambiguous typing. For example, a field of type integer could also be represented as a string reference (e.g. that integer gets provided at runtime). The types introduced in the example are therefore incomplete and are likely to be revisioned in a future version.

class flows_sdk.implementations.idp_v39_1.idp_blocks.IDPCoreBlock(idp_wf_config, reference_name='__idp_core__', title='Start Document Processing Subflow', description='Calls the Document Processing Subflow with the specified settings.', compatibility_spec=CompatibilitySpec(filter_roles=['idp_core'], filter_schema=None, label=None), identifier='IDP_CORE_V39_1', additional_inputs=None)

Bases: Block

IDP core block includes the steps to process the passed documents into machine-readable output.

Mandatory parameters:

Parameters:

idp_wf_config (IdpWorkflowConfig) – IdpWorkflowConfig object with all flow-level configuration settings for IDP

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique reference name within a flow, a unique one will be generated when not provided

  • compatibility_spec (Optional[CompatibilitySpec]) – defines a filter for flows that could be selected in the UI in place of “IDP Core” for the block. Defaults to flows with the role ‘idp_core’. The user will be able to pick any flow that matches the compatibility_spec. This serves to make upgrading to a new compatible version of “IDP Core” (e.g. IDP_CORE_V39) easier in future releases.

  • identifier (str) – The identifier of the flow. Defaults to the IDP_CORE_VXX identifier of the out-of-the-box flow that comes with the system.

  • additional_inputs (Optional[Dict[str, Any]]) – Dictionary containing key-value pairs that will be added to the inputs of the block. If a key is contained in the default inputs, its value will be overwritten by the value provided here. You generally do not need to provide additional_inputs, unless you are creating a custom derivative of the IDP_CODE_Vxx flow and are modifying or extending the list of inputs it accepts.

Example usage:

idp_core = IDPCoreBlock(idp_wf_config=idp_wf_config)
REFERENCE_NAME = '__idp_core__'
class flows_sdk.implementations.idp_v39_1.idp_blocks.SubmissionBootstrapBlock(reference_name=None, layout_release_uuid='${workflow.input.layout_release_uuid}', s3_config='${workflow.input.submission_bootstrap_s3_config}', s3_endpoint_url='${workflow.input.submission_bootstrap_s3_endpoint_url}', ocs_config='${workflow.input.submission_bootstrap_ocs_config}', http_config='${workflow.input.submission_bootstrap_http_config}', notification_workflow='${workflow.input.notification_workflow_initialization}', workflow_uuid='${workflow.input.workflow_uuid}', workflow_name='${workflow.input.workflow_name}', workflow_version='${workflow.input.workflow_version}', submission='${workflow.input.submission}')

Bases: Block

Submission bootstrap block initializes the submission object and prepares external images or other submission data if needed. This block should be called at the START of every flow that uses blocks included in the library.

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • layout_release_uuid (str) – Layout Release UUID at submission creation time, defaults to workflow_input(Settings.LayoutReleaseUuid)

  • s3_config (str) – Specify Amazon S3 credentials for submission retrieval store. Expects a json expression: {"aws_access_key_id": "X", "aws_secret_access_key": "Y"}, defaults to workflow_input(Settings.SubmissionBootstrapS3Config)

  • s3_endpoint_url (str) – Endpoint URL for S3 submission retrieval store, defaults to workflow_input(Settings.SubmissionBootstrapS3EndpointUrl)

  • ocs_config (str) – OCS Configuration for downloading submission data, defaults to workflow_input(Settings.SubmissionBootstrapOCSConfig)

  • http_config (Optional[str]) – An optional system secret field expressing the HTTP Configuration for downloading submission data. The content is expected to be a json formatted as: {"username": "X", "password": "Y", "ssl_cert": "CA bundle filename"}, defaults to workflow_input(Settings.SubmissionBootstrapHTTPConfig)

  • notification_workflow (Optional[str]) – Notification flow name to run when the submission is initialized, defaults to workflow_input(Settings.SubmissionBootstrapNotificationWorkflow)

  • workflow_uuid (str) – UUID of the triggered workflow version, defaults to workflow_input(Inputs.WorkflowUuid)

  • workflow_name (str) – Name of the triggered workflow, defaults to workflow_input(Inputs.WorkflowName)

  • workflow_version (str) – Version of the triggered workflow, defaults to workflow_input(Inputs.WorkflowVersion)

  • submission (Any) – Submission object

Example usage:

submission_bootstrap = SubmissionBootstrapBlock(reference_name='submission_bootstrap')
IDENTIFIER = 'SUBMISSION_BOOTSTRAP_2'
output(key=None)

Submission bootstrap ends with a CodeBlock, which nests the actual outputs under the ‘result’ key. This method overrides the default output in the same manner that flows_sdk.blocks.Routing.CodeBlock does for convenience.

Parameters:

key (Optional[str]) –

Optionally provide a key to directly get a nested property. When not provided, the entire ‘result’ will be returned. For example, we have:

{
    "result": {
        "a": {
            "b": 42
        }
    }
}
  • skipping the key output()` will return ``{"a": {"b": 42}}

  • calling with output("a") will result in {"b": 42}

  • calling with output("a.b") will return 42

Return type:

str

Returns:

the value under the provided key

class flows_sdk.implementations.idp_v39_1.idp_blocks.MachineCollationBlock(submission, cases, dedupe_files='${workflow.input.machine_collation_dedupe_files}', refresh_retention_period='${workflow.input.machine_collation_refresh_retention_period}', remove_from_cases='${workflow.input.machine_collation_remove_from_cases}', retention_period='${workflow.input.machine_collation_retention_period}', reference_name=None)

Bases: Block

Machine collation block groups files, documents and pages (from the submission) into cases.

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • cases (Any) –

    This parameter accepts an array of JSON objects that contain information on how Cases should be created. There are two ways to collate a case: by the filenames specified in the current Submission, or by the ids of the Documents or Pages submitted in previous submissions.

    This parameter follows the given format:

    {
        'cases': [
            'external_case_id': 'HS-1',
            'filename': 'file1.pdf',
            'documents': [42],
            'pages': [43]
        ]
    }
    

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • dedupe_files (Union[bool, str]) – Enabling this setting will replace case data from repeated file names within the same case. Cases will retain data from the most recently submitted version of a particular file. Note that this setting does not delete the old data, it just removes it from the case. Keep in mind this option is only relevant when adding to cases. Defaults to workflow_input(Settings.CollationDedupeFiles)

  • remove_from_cases (Optional[Any]) –

    This parameter accepts an array of JSON objects that contains information on how documents or pages should be removed from a pre-existing case. Currently, the only way to de-collate from a case is by the ids of the Documents or Pages within a given case, defaults to workflow_input(Settings.CollationRemoveFromCases)

    This parameter follows the given format:

    {
        'remove_from_cases': [
            'external_case_id': 'HS-1',
            'documents': [42],
            'pages': [43]
        ]
    }
    

  • retention_period (Union[int, str, None]) – The number of days to retain cases after they are updated by this block. If passed None, no changes to the deletion date will be made. Defaults to workflow_input(Settings.CollationRetentionPeriod)

  • refresh_retention_period (Union[bool, str]) – If True, (re)applies Retention Period parameter to all cases passed into the block. If False, only updates those without a pre-existing deletion date. Defaults to workflow_input(Settings.CollationRefreshRetentionPeriod)

Example usage:

machine_collation = MachineCollationBlock(
    reference_name='machine_collation',
    submission=submission_bootstrap.output('result.submission'),
    cases=submission_bootstrap.output('result.api_params.cases'),
    dedupe_files=True,
    remove_from_cases=custom_supervision.output('remove_from_cases')
)
IDENTIFIER = 'MACHINE_COLLATION_2'
class flows_sdk.implementations.idp_v39_1.idp_blocks.MachineClassificationBlock(submission, api_params, reference_name=None, rotation_correction_enabled='${workflow.input.machine_classification_rotation_correction_enabled}', mobile_processing_enabled='${workflow.input.machine_classification_mobile_processing_enabled}', file_transcription_enabled='${workflow.input.machine_classification_file_transcription_enabled}', layout_release_uuid='${workflow.input.layout_release_uuid}', vpc_registration_threshold='${workflow.input.structured_layout_match_threshold}', nlc_enabled='${workflow.input.semi_structured_classification}', nlc_target_accuracy='${workflow.input.semi_target_accuracy}', nlc_doc_grouping_logic='${workflow.input.semi_doc_grouping_logic}')

Bases: Block

Machine classification block automatically matches documents to structured, semi-structured or additional layouts

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • api_params (Any) –

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • rotation_correction_enabled (Union[bool, str]) – Identifies and corrects the orientation of semi-structured images, defaults to workflow_input(Settings.RotationCorrectionEnabled)

  • mobile_processing_enabled (Union[bool, str]) – Improves the machine readability of photo captured documents, defaults to workflow_input(Settings.MobileProcessingEnabled)

  • file_transcription_enabled (Union[bool, str]) – Improves the transcription speed of pdf files. Make sure that the “Image Correction” option is disabled and the pdfs you submit do not contain any rotated images defaults to workflow_input(Settings.FileTranscriptionEnabled)

  • layout_release_uuid (str) – Layout Release UUID at submission creation time, defaults to workflow_input(Settings.LayoutReleaseUuid)

  • vpc_registration_threshold (str) – Structured pages above this threshold will be automatically matched to a layout variation, defaults to workflow_input(Settings.StructuredLayoutMatchThreshold)

  • nlc_enabled (str) – Enables workflow to manage a model for automated classification of semi-structured and additional layout variations, defaults to workflow_input(Settings.SemiStructuredClassification)

  • nlc_target_accuracy (str) – defaults to workflow_input(Settings.SemiTargetAccuracy)

  • nlc_doc_grouping_logic (str) – Logic to handle multiple pages matched to the same layout variationin a given submission, defaults to workflow_input(Settings.SemiDocGroupingLogic)

Example usage:

machine_classification = MachineClassificationBlock(
    reference_name='machine_classification',
    submission=case_collation.output('submission'),
    api_params=submission_bootstrap.output('result.api_params'),
    rotation_correction_enabled=idp_wf_config.rotation_correction_enabled,
)
IDENTIFIER = 'MACHINE_CLASSIFICATION_6'
class flows_sdk.implementations.idp_v39_1.idp_blocks.ManualClassificationBlock(submission, api_params, reference_name=None, layout_release_uuid='${workflow.input.layout_release_uuid}', manual_nlc_enabled='${workflow.input.manual_nlc_enabled}', task_restrictions='${workflow.input.manual_classification_task_restrictions}', notification_workflow='${workflow.input.notification_workflow_classification}', nlc_enabled='${workflow.input.semi_structured_classification}', nlc_target_accuracy='${workflow.input.semi_target_accuracy}', mobile_processing_enabled='${workflow.input.machine_classification_mobile_processing_enabled}', vpc_registration_threshold='${workflow.input.structured_layout_match_threshold}')

Bases: Block

Manual classification block allows keyers to manually match submissions to their layouts. Keyers may perform manual classification if machine classification cannot automatically match a submission to a layout with high confidence

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • api_params (Any) –

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • layout_release_uuid (str) – Layout Release UUID at submission creation time, defaults to workflow_input(Settings.LayoutReleaseUuid)

  • manual_nlc_enabled (str) – Enables manual classification when applicable, defaults to workflow_input(Settings.ManualNlcEnabled)

  • task_restrictions (Union[List[Any], str, None]) – Defines what users can access Supervision tasks created by this block, defaults to workflow_input(Settings.ManualClassificationTaskRestrictions)

  • notification_workflow (str) – Notification flow name to run when the submission enters Manual Classification, defaults to workflow_input(Settings.ManualClassificationNotificationWorkflow)

  • nlc_enabled (Union[bool, str, None]) – Enables workflow to manage a model for automated classification of semi-structured and additional layout variations, defaults to workflow_input(Settings.SemiStructuredClassification)

  • nlc_target_accuracy (Union[float, str, None]) – Classification Target Accuracy, defaults to workflow_input(Settings.SemiTargetAccuracy)

  • mobile_processing_enabled (Union[bool, str, None]) – Improves the machine readability of photo captured document, defaults to workflow_input(Settings.MobileProcessingEnabled)

  • vpc_registration_threshold (Optional[str]) – Structured pages above this threshold will be automatically matched to a layout variation when performing light vpc registration in the doc-org task, defaults to workflow_input(Settings.StructuredLayoutMatchThreshold)

Example usage:

manual_classification = ManualClassificationBlock(
    reference_name='manual_classification',
    submission=machine_classification.output('submission'),
    api_params=submission_bootstrap.output('result.api_params'),
)
IDENTIFIER = 'MANUAL_CLASSIFICATION_3'
class flows_sdk.implementations.idp_v39_1.idp_blocks.MachineIdentificationBlock(submission, api_params, layout_release_uuid='${workflow.input.layout_release_uuid}', reference_name=None, manual_field_id_enabled='${workflow.input.manual_field_id_enabled}', field_id_target_accuracy='${workflow.input.field_id_target_accuracy}', table_id_target_accuracy='${workflow.input.table_id_target_accuracy}')

Bases: Block

Machine identification automatically identify fields and tables in the submission

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • api_params (Any) –

Parameters with defaults:

Parameters:
  • layout_release_uuid (Any) – Layout Release UUID at submission creation time, defaults to workflow_input(Settings.LayoutReleaseUuid)

  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • manual_field_id_enabled (str) – Enables manual identification when applicable, defaults to workflow_input(Settings.ManualFieldIdEnabled)

  • field_id_target_accuracy (str) – Field ID Target Accuracy, defaults to workflow_input(Settings.FieldIdTargetAccuracy)

  • table_id_target_accuracy (str) – Table ID Target Accuracy, defaults to workflow_input(Settings.TableIdTargetAccuracy)

Example usage:

machine_identification = MachineIdentificationBlock(
    reference_name='machine_identification',
    submission=manual_classification.output('submission'),
    api_params=submission_bootstrap.output('result.api_params'),
)
IDENTIFIER = 'MACHINE_IDENTIFICATION_7'
class flows_sdk.implementations.idp_v39_1.idp_blocks.ManualIdentificationBlock(submission, api_params, reference_name=None, task_restrictions='${workflow.input.manual_identification_task_restrictions}', manual_field_id_enabled='${workflow.input.manual_field_id_enabled}', layout_release_uuid='${workflow.input.layout_release_uuid}', notification_workflow='${workflow.input.notification_workflow_identification}')

Bases: Block

Manual identification allows keyers to complete field identification or table identification tasks, where they draw bounding boxes around the contents of certain fields, table columns or table rows. This identification process ensures that the system will be able to transcribe the correct content in the upcoming transcription process

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • api_params (Any) –

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • task_restrictions (Union[List[Any], str, None]) – Defines what users can access Supervision tasks created by this block, defaults to workflow_input(Settings.ManualFieldIdTaskRestrictions)

  • manual_field_id_enabled (str) – Enables manual identification when applicable, defaults to workflow_input(Settings.ManualFieldIdEnabled)

  • layout_release_uuid (str) – Layout Release UUID at submission creation time, defaults to workflow_input(Settings.LayoutReleaseUuid)

  • notification_workflow (str) – Notification flow name to run when the submission enters Manual Identification, defaults to workflow_inputSettings.ManualFieldIdNotificationWorkflow

Example usage:

manual_identification = ManualIdentificationBlock(
    reference_name='manual_identification',
    submission=machine_identification.output('submission'),
    api_params=submission_bootstrap.output('result.api_params'),
    task_restrictions=idp_wf_config.manual_identification_config.task_restrictions,
)
IDENTIFIER = 'MANUAL_IDENTIFICATION_5'
class flows_sdk.implementations.idp_v39_1.idp_blocks.MachineTranscriptionBlock(submission, api_params, reference_name=None, layout_release_uuid='${workflow.input.layout_release_uuid}', transcription_model='${workflow.input.transcription_model}', finetuning_only_trained_layouts='${workflow.input.finetuning_only_trained_layouts}', transcription_overrides='${workflow.input.transcription_overrides}', structured_text_target_accuracy='${workflow.input.structured_text_target_accuracy}', structured_text_confidence_threshold='${workflow.input.structured_text_threshold}', structured_text_acceptable_confidence='${workflow.input.structured_min_leg_threshold}', semi_structured_text_target_accuracy='${workflow.input.semi_structured_text_target_accuracy}', semi_structured_text_confidence_threshold='${workflow.input.semi_structured_text_threshold}', semi_structured_text_acceptable_confidence='${workflow.input.semi_structured_min_leg_threshold}', semi_structured_table_target_accuracy='${workflow.input.semi_structured_table_target_accuracy}', semi_structured_table_confidence_threshold='${workflow.input.semi_structured_table_threshold}', semi_structured_table_acceptable_confidence='${workflow.input.semi_structured_table_min_leg_threshold}', structured_checkbox_target_accuracy='${workflow.input.structured_checkbox_target_accuracy}', structured_checkbox_confidence_threshold='${workflow.input.structured_checkbox_threshold}', structured_checkbox_acceptable_confidence='${workflow.input.checkbox_min_leg_threshold}', structured_signature_target_accuracy='${workflow.input.structured_signature_target_accuracy}', structured_signature_confidence_threshold='${workflow.input.structured_signature_threshold}', structured_signature_acceptable_confidence='${workflow.input.signature_min_leg_threshold}', semi_structured_checkbox_target_accuracy='${workflow.input.semi_structured_checkbox_target_accuracy}', semi_structured_checkbox_confidence_threshold='${workflow.input.semi_structured_checkbox_threshold}', semi_structured_checkbox_acceptable_confidence='${workflow.input.semi_structured_checkbox_min_leg_threshold}')

Bases: Block

Machine transcription automatically transcribes the content of your submission

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • api_params (Any) –

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • layout_release_uuid (str) – Layout Release UUID at submission creation time, defaults to workflow_input(Settings.LayoutReleaseUuid)

  • transcription_model (str) – The transcription model that will be used for this flow, defaults to workflow_input(Settings.TranscriptionModel)

  • finetuning_only_trained_layouts (str) – defaults to workflow_input( Settings.FinetuningOnlyTrainedLayouts )

  • transcription_overrides (str) – defaults to workflow_input( Settings.TranscriptionOverrides )

  • structured_text_target_accuracy (str) – defaults to workflow_input( Settings.StructuredTextTargetAccuracy )

  • structured_text_confidence_threshold (str) – defaults to workflow_input( Settings.StructuredTextThreshold )

  • structured_text_acceptable_confidence (str) – defaults to workflow_input( Settings.StructuredTextMinLegThreshold )

  • semi_structured_text_target_accuracy (str) – defaults to workflow_input( Settings.SemiStructuredTextTargetAccuracy )

  • semi_structured_text_confidence_threshold (str) – defaults to workflow_input( Settings.SemiStructuredTextThreshold )

  • semi_structured_text_acceptable_confidence (str) – defaults to workflow_input( Settings.SemiStructuredTextMinLegThreshold )

  • structured_checkbox_target_accuracy (str) – defaults to workflow_input( Settings.StructuredCheckboxTargetAccuracy )

  • structured_checkbox_confidence_threshold (str) – defaults to workflow_input( Settings.StructuredCheckboxThreshold )

  • structured_checkbox_acceptable_confidence (str) – defaults to workflow_input( Settings.StructuredCheckboxMinLegThreshold )

  • structured_signature_target_accuracy (str) – defaults to workflow_input( Settings.StructuredSignatureTargetAccuracy )

  • structured_signature_confidence_threshold (str) – defaults to workflow_input( Settings.StructuredSignatureThreshold )

  • structured_signature_acceptable_confidence (str) – defaults to workflow_input( Settings.StructuredSignatureMinLegThreshold )

  • semi_structured_checkbox_target_accuracy (str) – defaults to workflow_input( Settings.SemiStructuredCheckboxTargetAccuracy )

  • semi_structured_checkbox_confidence_threshold (str) – defaults to workflow_input( Settings.SemiStructuredCheckboxThreshold )

  • semi_structured_checkbox_acceptable_confidence (str) – defaults to workflow_input( Settings.SemiStructuredCheckboxMinLegThreshold )

Example usage:

machine_transcription = MachineTranscriptionBlock(
    reference_name='machine_transcription',
    submission=manual_identification.output('submission'),
    api_params=submission_bootstrap.output('result.api_params'),
)
IDENTIFIER = 'MACHINE_TRANSCRIPTION_9'
class flows_sdk.implementations.idp_v39_1.idp_blocks.ManualTranscriptionBlock(submission, api_params, reference_name=None, supervision_transcription_masking='${workflow.input.manual_transcription_supervision_transcription_masking}', table_output_manual_review='${workflow.input.manual_transcription_table_output_manual_review}', always_supervise_blank_cells='${workflow.input.manual_transcription_always_supervise_blank_cells}', task_restrictions='${workflow.input.manual_transcription_task_restrictions}', manual_transcription_enabled='${workflow.input.manual_transcription_enabled}', notification_workflow='${workflow.input.notification_workflow_transcription}', title='Manual Transcription', description='Manually transcribe fields that could not be automatically transcribed')

Bases: Block

Manual transcription lets your keyers manually enter the text found in fields or tables that could not be automatically transcribed

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • api_params (Any) –

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • supervision_transcription_masking (Union[bool, str]) – Prevents users from inputting invalid characters during Supervision Transcription tasks, defaults to workflow_input(Settings.ManualTranscriptionSupervisionTranscriptionMasking)

  • table_output_manual_review (Union[bool, str]) – Always generates a table transcription task if the layout contains a table. If disabled, a table transcription task will only be generated if one or more cells have transcribed values below the defined thresholds, defaults to workflow_input(Settings.ManualTranscriptionTableOutputManualReview)

  • always_supervise_blank_cells (Union[bool, str]) – If enabled, always send blank cells in transcribed tables to be supervised manually, regardless of confidence defaults to workflow_input(Settings.ManualTranscriptionAlwaysSuperviseBlankCells)

  • task_restrictions (Union[List[Any], str, None]) – Defines what users can access Supervision tasks for submission from a particular source or submissions matching a specific layout, defaults to workflow_input(Settings.ManualTranscriptionTaskRestrictions)

  • manual_transcription_enabled (str) – Enables manual transcription when applicable, defaults to workflow_input(Settings.ManualTranscriptionEnabled)

  • notification_workflow (str) – Notification flow name to run when the submission enters Manual Transcription, defaults to workflow_input(Settings.ManualTranscriptionNotificationWorkflow)

Example usage:

manual_transcription = ManualTranscriptionBlock(
    reference_name='manual_transcription',
    submission=machine_transcription.output('submission'),
    api_params=submission_bootstrap.output('result.api_params'),
    supervision_transcription_masking=(
        idp_wf_config.manual_transcription_config.supervision_transcription_masking
    ),
    table_output_manual_review=(
        idp_wf_config.manual_transcription_config.table_output_manual_review
    ),
    task_restrictions=idp_wf_config.manual_transcription_config.task_restrictions,
)
IDENTIFIER = 'MANUAL_TRANSCRIPTION_5'
class flows_sdk.implementations.idp_v39_1.idp_blocks.FlexibleExtractionBlock(submission, api_params, reference_name=None, supervision_transcription_masking='${workflow.input.flexible_extraction_supervision_transcription_masking}', task_restrictions='${workflow.input.flexible_extraction_task_restrictions}', layout_release_uuid='${workflow.input.layout_release_uuid}', notification_workflow='${workflow.input.notification_workflow_flexible_extraction}')

Bases: Block

Flexible extraction manually transcribes fields marked for review

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • api_params (Any) –

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • supervision_transcription_masking (Union[bool, str, None]) – Prevents users from inputting invalid characters during the Flexible Extraction task, defaults to workflow_input(Settings.FlexibleExtractionSupervisionTranscriptionMasking)

  • task_restrictions (Union[List[Any], str, None]) – Defines what users can access Supervision tasks created by this block, defaults workflow_input(Settings.FlexibleExtractionTaskRestrictions)

  • layout_release_uuid (str) – Layout Release UUID at submission creation time, defaults to workflow_input(Settings.LayoutReleaseUuid)

  • notification_workflow (str) – Notification flow name to run when the submission enters Flexible Extraction, defaults to workflow_input(Settings.FlexibleExtractionNotificationWorkflow)

Example usage:

flexible_extraction = FlexibleExtractionBlock(
    reference_name='flexible_extraction',
    submission=manual_transcription.output('submission'),
    api_params=submission_bootstrap.output('result.api_params'),
    supervision_transcription_masking=(
        idp_wf_config.flexible_extraction_config.supervision_transcription_masking
    ),
    task_restrictions=idp_wf_config.flexible_extraction_config.task_restrictions,
)
IDENTIFIER = 'FLEXIBLE_EXTRACTION_4'
class flows_sdk.implementations.idp_v39_1.idp_blocks.ReprocessingBlock(reference_name, submission, parent_flow='${workflow.definition.identifier}', idp_wf_config={'auto_qa_sampling_rate_enabled': '${workflow.input.auto_qa_sampling_rate_enabled}', 'checkbox_min_leg_threshold': '${workflow.input.checkbox_min_leg_threshold}', 'exclude_auto_transcribe_fields_from_qa': '${workflow.input.exclude_auto_transcribe_fields_from_qa}', 'field_id_qa_enabled': '${workflow.input.field_id_qa_enabled}', 'field_id_qa_sample_rate': '${workflow.input.field_id_qa_sample_rate}', 'field_id_target_accuracy': '${workflow.input.field_id_target_accuracy}', 'finetuning_only_trained_layouts': '${workflow.input.finetuning_only_trained_layouts}', 'flexible_extraction_supervision_transcription_masking': '${workflow.input.flexible_extraction_supervision_transcription_masking}', 'flexible_extraction_task_restrictions': '${workflow.input.flexible_extraction_task_restrictions}', 'layout_release_uuid': '${workflow.input.layout_release_uuid}', 'machine_classification_file_transcription_enabled': '${workflow.input.machine_classification_file_transcription_enabled}', 'machine_classification_mobile_processing_enabled': '${workflow.input.machine_classification_mobile_processing_enabled}', 'machine_classification_rotation_correction_enabled': '${workflow.input.machine_classification_rotation_correction_enabled}', 'manual_field_id_enabled': '${workflow.input.manual_field_id_enabled}', 'manual_nlc_enabled': '${workflow.input.manual_nlc_enabled}', 'manual_transcription_always_supervise_blank_cells': '${workflow.input.manual_transcription_always_supervise_blank_cells}', 'manual_transcription_enabled': '${workflow.input.manual_transcription_enabled}', 'manual_transcription_supervision_transcription_masking': '${workflow.input.manual_transcription_supervision_transcription_masking}', 'manual_transcription_table_output_manual_review': '${workflow.input.manual_transcription_table_output_manual_review}', 'manual_transcription_task_restrictions': '${workflow.input.manual_transcription_task_restrictions}', 'notification_workflow_classification': '${workflow.input.notification_workflow_classification}', 'notification_workflow_flexible_extraction': '${workflow.input.notification_workflow_flexible_extraction}', 'notification_workflow_identification': '${workflow.input.notification_workflow_identification}', 'notification_workflow_initialization': '${workflow.input.notification_workflow_initialization}', 'notification_workflow_transcription': '${workflow.input.notification_workflow_transcription}', 'qa': '${workflow.input.qa}', 'semi_doc_grouping_logic': '${workflow.input.semi_doc_grouping_logic}', 'semi_qa_sample_rate': '${workflow.input.semi_qa_sample_rate}', 'semi_structured_checkbox_max_training_records': '${workflow.input.semi_structured_checkbox_max_training_records}', 'semi_structured_checkbox_min_leg_threshold': '${workflow.input.semi_structured_checkbox_min_leg_threshold}', 'semi_structured_checkbox_min_training_records': '${workflow.input.semi_structured_checkbox_min_training_records}', 'semi_structured_checkbox_target_accuracy': '${workflow.input.semi_structured_checkbox_target_accuracy}', 'semi_structured_checkbox_threshold': '${workflow.input.semi_structured_checkbox_threshold}', 'semi_structured_classification': '${workflow.input.semi_structured_classification}', 'semi_structured_improved_transcr_threshold_accuracy': '${workflow.input.semi_structured_improved_transcr_threshold_accuracy}', 'semi_structured_min_leg_threshold': '${workflow.input.semi_structured_min_leg_threshold}', 'semi_structured_qa_sample_rate': '${workflow.input.semi_structured_qa_sample_rate}', 'semi_structured_table_max_training_records': '${workflow.input.semi_structured_table_max_training_records}', 'semi_structured_table_min_leg_threshold': '${workflow.input.semi_structured_table_min_leg_threshold}', 'semi_structured_table_min_training_records': '${workflow.input.semi_structured_table_min_training_records}', 'semi_structured_table_target_accuracy': '${workflow.input.semi_structured_table_target_accuracy}', 'semi_structured_table_threshold': '${workflow.input.semi_structured_table_threshold}', 'semi_structured_text_max_training_records': '${workflow.input.semi_structured_text_max_training_records}', 'semi_structured_text_min_training_records': '${workflow.input.semi_structured_text_min_training_records}', 'semi_structured_text_target_accuracy': '${workflow.input.semi_structured_text_target_accuracy}', 'semi_structured_text_threshold': '${workflow.input.semi_structured_text_threshold}', 'semi_structured_transcription_training': '${workflow.input.semi_structured_transcription_training}', 'semi_structured_transcription_training_period': '${workflow.input.semi_structured_transcription_training_period}', 'semi_target_accuracy': '${workflow.input.semi_target_accuracy}', 'signature_min_leg_threshold': '${workflow.input.signature_min_leg_threshold}', 'structured_checkbox_max_training_records': '${workflow.input.structured_checkbox_max_training_records}', 'structured_checkbox_min_training_records': '${workflow.input.structured_checkbox_min_training_records}', 'structured_checkbox_qa_sample_rate': '${workflow.input.structured_checkbox_qa_sample_rate}', 'structured_checkbox_target_accuracy': '${workflow.input.structured_checkbox_target_accuracy}', 'structured_checkbox_threshold': '${workflow.input.structured_checkbox_threshold}', 'structured_entry_qa_sample_rate': '${workflow.input.structured_entry_qa_sample_rate}', 'structured_improved_transcr_threshold_accuracy': '${workflow.input.structured_improved_transcr_threshold_accuracy}', 'structured_layout_match_threshold': '${workflow.input.structured_layout_match_threshold}', 'structured_min_leg_threshold': '${workflow.input.structured_min_leg_threshold}', 'structured_signature_max_training_records': '${workflow.input.structured_signature_max_training_records}', 'structured_signature_min_training_records': '${workflow.input.structured_signature_min_training_records}', 'structured_signature_qa_sample_rate': '${workflow.input.structured_signature_qa_sample_rate}', 'structured_signature_target_accuracy': '${workflow.input.structured_signature_target_accuracy}', 'structured_signature_threshold': '${workflow.input.structured_signature_threshold}', 'structured_text_max_training_records': '${workflow.input.structured_text_max_training_records}', 'structured_text_min_training_records': '${workflow.input.structured_text_min_training_records}', 'structured_text_target_accuracy': '${workflow.input.structured_text_target_accuracy}', 'structured_text_threshold': '${workflow.input.structured_text_threshold}', 'structured_transcription_training': '${workflow.input.structured_transcription_training}', 'structured_transcription_training_period': '${workflow.input.structured_transcription_training_period}', 'submission_bootstrap_http_config': '${workflow.input.submission_bootstrap_http_config}', 'submission_bootstrap_ocs_config': '${workflow.input.submission_bootstrap_ocs_config}', 'submission_bootstrap_s3_config': '${workflow.input.submission_bootstrap_s3_config}', 'table_cell_transcription_qa_enabled': '${workflow.input.table_cell_transcription_qa_enabled}', 'table_cell_transcription_qa_sample_rate': '${workflow.input.table_cell_transcription_qa_sample_rate}', 'table_id_qa_enabled': '${workflow.input.table_id_qa_enabled}', 'table_id_qa_sample_rate': '${workflow.input.table_id_qa_sample_rate}', 'table_id_target_accuracy': '${workflow.input.table_id_target_accuracy}', 'transcription_model': '${workflow.input.transcription_model}', 'workflow_name': '${workflow.input.workflow_name}', 'workflow_uuid': '${workflow.input.workflow_uuid}', 'workflow_version': '${workflow.input.workflow_version}'}, trigger='${workflow.input.$trigger}')

Bases: Block

Reprocessing block enables a second extraction attempt if documents were initially confidently mis-classified by the machine. When an end-user specifies that a document layout is incorrect during Manual Identification, Manual Transcription, or Flexible Extraction, the flow will reach the Reprocessing block which will mark correctly classified documents as “complete” status, and misclassified documents with “reprocess” status will be processed in a second execution of the flow. This block should be placed before the SubmissionCompleteBlock to enable reprocessing.

Mandatory parameters:

Parameters:

submission (Any) – Submission object

IDENTIFIER = 'REPROCESSING'
class flows_sdk.implementations.idp_v39_1.idp_blocks.SubmissionCompleteBlock(submission, is_reprocessing=False, reference_name=None, nlc_qa_sampling_ratio='${workflow.input.semi_qa_sample_rate}', field_id_qa_enabled='${workflow.input.field_id_qa_enabled}', field_id_qa_sampling_ratio='${workflow.input.field_id_qa_sample_rate}', table_id_qa_enabled='${workflow.input.table_id_qa_enabled}', table_id_qa_sampling_ratio='${workflow.input.table_id_qa_sample_rate}', transcription_qa_enabled='${workflow.input.qa}', transcription_structured_entry_qa_sample_rate='${workflow.input.structured_entry_qa_sample_rate}', transcription_structured_signature_qa_sample_rate='${workflow.input.structured_signature_qa_sample_rate}', transcription_structured_checkbox_qa_sample_rate='${workflow.input.structured_checkbox_qa_sample_rate}', transcription_semi_structured_qa_sample_rate='${workflow.input.semi_structured_qa_sample_rate}', table_cell_transcription_qa_enabled='${workflow.input.table_cell_transcription_qa_enabled}', table_cell_transcription_qa_sample_rate='${workflow.input.table_cell_transcription_qa_sample_rate}', auto_qa_sampling_rate_enabled='${workflow.input.auto_qa_sampling_rate_enabled}', exclude_auto_transcribe_fields_from_qa='${workflow.input.exclude_auto_transcribe_fields_from_qa}', payload=None)

Bases: Block

Submission complete block finalizes submission processing and updates reporting data. This block should be called at the END of every flow that uses blocks included in the flows_sdk.implementations.idp_v34 library. The block requires both the submission object and a payload to be passed in.

Mandatory parameters:

Parameters:
  • submission (str) – Submission object

  • payload (Optional[Any]) – Object to pass to downstream systems

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • nlc_qa_sampling_ratio (Union[str, int]) – Defines the percentage of documents sampled for Classification QA, defaults to workflow_input(Settings.SemiQaSampleRate)

  • field_id_qa_enabled (Union[str, bool]) – defaults to workflow_input(Settings.FieldIdQaEnabled)

  • field_id_qa_sampling_ratio (Union[str, int]) – Defines the percentage of documents sampled for Identification QA, defaults to workflow_input(Settings.FieldIdQaSampleRate)

  • table_id_qa_enabled (Union[str, bool]) – Allows users to verify the location of table cells, defaults to workflow_input(Settings.TableIdQaEnabled)

  • table_id_qa_sampling_ratio (Union[str, int]) – Defines the percentage of tables the system samples for QA, defaults to workflow_input(Settings.TableIdQaSampleRate)

  • transcription_qa_enabled (Union[str, bool]) – defaults to workflow_input(Settings.TranscriptionQaEnabled)

  • transcription_qa_sampling_ratio – Defines the percentage of fields sampled for Transcription QA, defaults to workflow_input(Settings.TranscriptionQaSampleRate)

  • table_cell_transcription_qa_enabled (Union[str, bool]) – defaults to workflow_input(Settings.TableCellTranscriptionQaEnabled)

  • table_cell_transcription_qa_sample_rate (Union[str, int]) – Defines the percentage of cells the system samples for QA. This value is likely to be lower than “Transcription QA Sample Rate” since there are more table cells than fields on any given page, defaults to workflow_input(Settings.TableCellTranscriptionQaSampleRate)

  • exclude_auto_transcribe_fields_from_qa (Union[str, bool]) – Defines fields that are marked as auto-transcribe in layout should be excluded from QA or not. defaults to workflow_input(Settings.ExcludeAutoTranscribeFieldsFromQA)

  • is_reprocessing (Union[str, bool]) – boolean to handle execution when reprocessing defaults to False

Example usage:

submission_complete = SubmissionCompleteBlock(
    reference_name='complete_submission',
    submission=flexible_extraction.output('submission')
)
IDENTIFIER = 'SUBMISSION_COMPLETE_4'
class flows_sdk.implementations.idp_v39_1.idp_blocks.SetTransformedOutputBlock(submission, transformed_output, reference_name=None)

Bases: Block

Set Transformed Output block sets the transformed output for the provided submission. This can later be used to send the transformed data downstream using output blocks. For more information see https://docs.hyperscience.com/#retrieving-transformed-submissions.

Mandatory parameters:

Parameters:
  • submission (str) – Submission object

  • transformed_output (str) – Object that will be set as the transformed output for this submission

IDENTIFIER = 'SET_TRANSFORMED_OUTPUT'
class flows_sdk.implementations.idp_v39_1.idp_blocks.IdpCustomSupervisionBlock(submission, task_purpose, supervision_template, reference_name=None, supervision_transcription_masking=True, task_restrictions=None, page_ids=None, notification_workflow='IDP_SUBMISSION_NOTIFY_V39_1', title='IDP Custom Supervision', description='IDP Custom Supervision')

Bases: Block

An IDP wrapper for the Custom Supervision block. It has the same functionality, but handles reading/writing data from the IDP database.

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • task_purpose (str) – the custom task name that’s given to all supervision tasks that run through a particular instance of a custom supervision block. This can be used to filter tasks in the task queue.

  • supervision_template (Any) – JSON used for configuration of the Custom Supervision Task.

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • supervision_transcription_masking (Union[str, bool]) – Prevents users from inputting invalid characters during the Custom Supervision task, defaults to True

  • task_restrictions (Union[str, List[Any], None]) – Defines what users can access Supervision tasks created by this block, defaults to []

  • page_ids (Union[str, List[int], None]) – This is an optional parameter. If left empty, the IDP wrapper will, by default, pull in all pages in the submission. Adding a list of page IDs will override the default list of pages. defaults to []

  • notification_workflow (str) – Notification flow name to run when the submission enters Custom Supervision (which is called within the IDP Custom Supervision block), defaults to IDP_SUBMISSION_NOTIFY_NAME

  • title (Optional[str]) – UI-visible title of the block. Defaults to ‘IDP Custom Supervision’

  • description (Optional[str]) – Description of the block. Both useful for documentation purpose and visible to users in the Flow studio. Defaults to ‘IDP Custom Supervision’

Example usage:

idp_custom_supervision = IdpCustomSupervisionBlock(
    reference_name='idp_custom_supervision',
    submission=manual_transcription.output('submission'),
    task_purpose='make_custom_decision',
    supervision_template=[
        {
            'name': 'three_column_template',
            'version': '1.0',
            'thumbnail': {
                'group_by_document': True,
                'group_by_case': True,
            },
            'action': [
                {
                    'name': 'my_tab',
                    'display': 'SSN Info',
                    'input': [
                        {
                            'name': 'ssn_field',
                            'type': 'transcription',
                            'data_identifier': '2afd7a67-ddee-484f-ab6b-2d292752e5ee',
                            'title': 'SSN',
                        },
                    ],
                    'ui': {
                        'groups': [
                            {
                                'title': 'My Group',
                                'fields': ['ssn_field'],
                            },
                        ],
                    },
                },
            ],
        },
    ],
    supervision_transcription_masking=False,
)
IDENTIFIER = 'IDP_CUSTOM_SUPERVISION_2'
class flows_sdk.implementations.idp_v39_1.idp_blocks.IDPFullPageTranscriptionBlock(submission, reference_name=None, title='Full Page Transcription (Submission)', description='Transcribes the documents included in a submission', app_metadata=None)

Bases: Block

IDP Full Page Transcription Block machine-transcribes the documents contained in a submission

Mandatory parameters:

Parameters:

submission (str) – submission object that includes the raw documents that need to be transcribed

Optional Parameters:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • app_metadata (Optional[List[str]]) – used for capturing information about the flow

Example usage:

# This is one way to obtain the submission object (using SubmissionBootstrapBlock)
submission_bootstrap = SubmissionBootstrapBlock(reference_name='submission_bootstrap')
# IDPFullPageTranscriptionBlock transcribes the documents contained in the submission object
idp_fpt_block = IDPFullPageTranscriptionBlock(
    reference_name='idp_fpt',
    submission=submission_bootstrap.output(),
)
# reference to the output of IDPFullPageTranscriptionBlock which can be passed as input to
# the next block. For more information on output() and output references, please take a
# look at class Block and its output() method
output_ref = idp_fpt_block.output()
IDENTIFIER = 'FULL_PAGE_TRANSCRIPTION_SUBMISSION_2'
class flows_sdk.implementations.idp_v39_1.idp_blocks.IDPImageCorrectionBlock(submission, reference_name=None, rotation_correction_enabled=True, mobile_processing_enabled=False, title='Image Correction (Submission)', description='Rotate and de-skew documents included in a submission', app_metadata=None)

Bases: Block

IDP Image Correction Block rotates and de-skews tilted/skewed images so that they are more transcription-friendly

Mandatory parameters:

Parameters:

submission (str) – submission object that includes the raw documents that need to be image-corrected

Optional Parameters:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • app_metadata (Optional[List[str]]) – used for capturing information about the flow

Example usage:

# This is one way to obtain the submission object (using SubmissionBootstrapBlock)
submission_bootstrap = SubmissionBootstrapBlock(reference_name='submission_bootstrap')
# IDPImageCorrectionBlock takes in the submission object and outputs in the same format with
# its images
idp_image_correct_block = IDPImageCorrectionBlock(
    reference_name='idp_image_correct',
    submission=submission_bootstrap.output(),
)
# reference to the output of IDPImageCorrectionBlock which can be passed as input to the
# next block. For more information on output() and output references, please take a look at
# class Block and its output() method
output_ref = idp_image_correct_block.output()
IDENTIFIER = 'IMAGE_CORRECTION_SUBMISSION'

Additional Blocks (v39.1)

class flows_sdk.implementations.idp_v39_1.additional_blocks.JSONOutputsBlock(inputs, blocks=None)

Bases: Outputs

Output block allows users to send JSON data extracted by an IDP flow to other systems for downstream processing Mandatory parameters:

Parameters:

inputs (Dict[str, Any]) – Used by the UI to automatically pre-populate the input of newly added blocks. Extended in this class to include 'enabled': True to visualize an enabled/disabled trigger that is controllable via Flow Studio.

Usage when parameters are to be defined in Flow Studio:

outputs = JSONOutputBlock(
    inputs={'result': }
)
class flows_sdk.implementations.idp_v39_1.additional_blocks.ImageCorrectionBlock(pages, reference_name=None, rotation_correction_enabled=True, mobile_processing_enabled=False, title='Image Correction (Pages)', description='Rotate and de-skew documents', app_metadata=None)

Bases: Block

Image Correction Block rotates and de-skews tilted/skewed images so that they are more transcription-friendly Mandatory parameters:

Parameters:

pages (str) – list of page object that includes the raw documents that need to be image-corrected

Optional Parameters:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • app_metadata (Optional[List[str]]) – used for capturing information about the flow

Example usage:

# This is one way to obtain the submission object (using SubmissionBootstrapBlock)
submission_bootstrap = SubmissionBootstrapBlock(reference_name='submission_bootstrap')
# ImageCorrectionBlock takes in the submission object and outputs in the same format with
# its images
image_correct_block = ImageCorrectionBlock(
    reference_name='image_correct',
    pages=submission_bootstrap.output(submission.unassigned_pages),
)
# reference to the output of IDPImageCorrectionBlock which can be passed as input to the
# next block. For more information on output() and output references, please take a look
# at class Block and its output() method
output_ref = image_correct_block.output()
IDENTIFIER = 'IMAGE_CORRECTION_PAGES'
class flows_sdk.implementations.idp_v39_1.additional_blocks.FullPageTranscriptionBlock(pages, reference_name=None, title='Full Page Transcription (Pages)', description='Transcribes documents', app_metadata=None)

Bases: Block

Full Page Transcription Block machine-transcribes the documents contained in a submission

Mandatory parameters:

Parameters:

pages (str) – list of pages that includes the raw documents that need to be transcribed

Optional Parameters:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • app_metadata (Optional[List[str]]) – used for capturing information about the flow

Example usage:

# This is one way to obtain the submission object (using SubmissionBootstrapBlock)
submission_bootstrap = SubmissionBootstrapBlock(reference_name='submission_bootstrap')
# FullPageTranscriptionBlock transcribes the documents contained in the submission object
fpt_block = FullPageTranscriptionBlock(
    reference_name='fpt',
    submission=submission_bootstrap.output(submission.unassigned_pages),
)
# reference to the output of FullPageTranscriptionBlock which can be passed as input to the
# next block. For more information on output() and output references, please take a look at
# class Block and its output() method
output_ref = fpt_block.output()
IDENTIFIER = 'FULL_PAGE_TRANSCRIPTION_PAGES_2'
class flows_sdk.implementations.idp_v39_1.additional_blocks.CustomSupervisionBlock(submission, task_purpose, data, supervision_template, reference_name=None, supervision_transcription_masking=True, task_restrictions=None, notification_workflow='IDP_SUBMISSION_NOTIFY_V39_1', title='Custom Supervision', description='Custom Supervision')

Bases: Block

The CustomSupervisionBlock uses a JSON schema as a form builder to create customer-driven UIs for Supervision tasks.

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • task_purpose (str) – the custom task name that’s given to all supervision tasks that run through a particular instance of a custom supervision block. This can be used to filter tasks in the task queue.

  • data (Any) – JSON structure that supplies the data used to populated the supervision_template

  • supervision_template (Any) – JSON used for configuration of the Custom Supervision Task.

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • supervision_transcription_masking (bool) – Prevents users from inputting invalid characters during the Custom Supervision task, defaults to True

  • task_restrictions (Optional[List[Any]]) – Defines what users can access Supervision tasks created by this block, defaults to []

  • notification_workflow (str) – Notification flow name to run when the submission enters Custom Supervision, defaults to IDP_SUBMISSION_NOTIFY_NAME

  • title (Optional[str]) – UI-visible title of the block. Defaults to ‘IDP Custom Supervision’

  • description (Optional[str]) – Description of the block. Both useful for documentation purpose and visible to users in the Flow studio. Defaults to ‘IDP Custom Supervision’

Example usage:

custom_supervision = CustomSupervisionBlock(
    reference_name='custom_supervision',
    submission=manual_transcription.output('submission'),
    task_purpose='make_custom_decision',
    supervision_template=[
        {
            'name': 'three_column_template',
            'version': '1.0',
            'thumbnail': {
                'group_by_document': True,
                'group_by_case': True,
            },
            'action': [
                {
                    'name': 'my_tab',
                    'display': 'SSN Info',
                    'input': [
                        {
                            'name': 'ssn_field',
                            'type': 'transcription',
                            'data_identifier': '2afd7a67-ddee-484f-ab6b-2d292752e5ee',
                            'title': 'SSN',
                        },
                    ],
                    'ui': {
                        'groups': [
                            {
                                'title': 'My Group',
                                'fields': ['ssn_field'],
                            },
                        ],
                    },
                },
            ],
        },
    ],
    data={
        'fields': [
            {
                'id': 967,
                'uuid': 'fe8d4e7a-a82c-448c-85a1-0a3cbbe53b61',
                'value': '111-22-3333',
                'page_id': 15,
                'layout_field_id': '2afd7a67-ddee-484f-ab6b-2d292752e5ee',
                'validation_overridden': False,
                'bounding_box': [
                    0.09360783305186686,
                    0.25792617589433436,
                    0.6358913805295097,
                    0.28862414740388187,
                ],
                'occurrence_index': 1,
            },
        ],
        'template_fields': [
            {
                'uuid': '2afd7a67-ddee-484f-ab6b-2d292752e5ee',
                'notes': '',
                'type': 'entry',
                'data_type_uuid': 'b600a4e4-758f-4c4b-a2f4-cbb5dd870a0c',
                'name': 'SSN',
            },
        ],
        'pages': [
            {
                'id': 15,
                'file_page_number': 1,
                'document_page_number': 1,
                'image_url': '/image/2ed3d989-b64a-44bd-a9fb-6e0dcc5ccdb8',
                'form_id': 2,
                'submission_id': 1,
                'external_case_ids': [],
                'read_only': False,
                'filename': '1 page.pdf',
                'form_page_id': 1
            },
        ],
        'documents': [
            {
                'id': 2,
                'layout_uuid': 'e31c3cbf-dcea-44fd-a052-902463c6040f',
                'layout_name': 'layout_name',
                'page_count': 2,
                'read_only': False,
            },
        ],
        'cases': [],
        'selected_choices': [],
    },
    supervision_transcription_masking=False,
)
IDENTIFIER = 'CUSTOM_SUPERVISION_2'
class flows_sdk.implementations.idp_v39_1.additional_blocks.SoapRequestBlock(method, reference_name=None, endpoint='', body_namespace=None, headers_namespace=None, headers=None, params=None, use_wsdl=False, wsdl_uri='', wsdl_service_name='', wsdl_port_name='', title='HTTP REST Block', description='HTTP REST block')

Bases: Block

Soap Block allows user to perform SOAP requests Mandatory parameters:

Parameters:
  • method (str) – SOAP method to invoke

  • endpoint (str) – absolute url including schema to HTTP endpoint for this request

Optional Parameters:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • body_namespace (Optional[str]) – SOAP request body namespace url

  • headers_namespace (Optional[str]) – SOAP request headers namespace url

  • headers (Optional[Dict[Any, Any]]) – SOAP headers to use

  • params (Optional[Dict[Any, Any]]) – key-value pair used as query parameters in the request to be inserted as SOAP body

Example usage:

# minimum viable instantiation, the rest of the fields can later be added in the UI
soap_block = SoapRequestBlock(
    reference_name='soap_request',
    title='Soap Request',
    description='Perform soap request',
    method='ExampleMethod',
)
soap_block = SoapRequestBlock(
    reference_name='soap_request',
    title='Soap Request',
    description='Perform soap request',
    method='ExampleMethod',
    endpoint='http://example.org/abc.wso',
    body_namespace='http://www.example.org/abc.countryinfo',
    params={'example': 'stuff'},
    headers={}
)
# reference to the output of SoapRequestBlock which can be passed as input to the next block
# for more information on output() and output references, please take a look at class Block
# and its output() method
output_ref = soap_block.output()
IDENTIFIER = 'SOAP_REQ'
class flows_sdk.implementations.idp_v39_1.additional_blocks.HttpRestBlock(method, endpoint, reference_name=None, headers=None, params=None, payload=None, json=None, authorization_type=None, handled_error_codes=None, auth_params=None, title='HTTP REST Block', description='HTTP REST block')

Bases: BaseHttpRestBlock

Http Rest Block allows user to perform HTTP requests Mandatory parameters:

Parameters:
  • method (str) – HTTP method to use

  • endpoint (str) – absolute url including schema to HTTP endpoint for this request

Optional Parameters:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • headers (Optional[Dict[str, str]]) – HTTP headers to use

  • params (Optional[Dict[str, Any]]) – key-value pair used as query parameters in the request

  • payload (Optional[Dict[str, str]]) – key-value pair to be used as x-www-form-urlencoded data. Mutually exclusive with json

  • json (Optional[Dict[str, Any]]) – key-value pair to be used as json data. Mutually exclusive with payload

  • authorization_type (Optional[str]) – authorization type (none, http_header, oauth_2_client_credentials)

  • handled_error_codes (Optional[List[Union[int, str]]]) – specify HTTP error codes that will not fail the task, allowing for wildcards with the ‘x’ char. Example: [400, “5xx”]

  • auth_params (Union[HttpHeaderAuthorizationParams, OAuth2AuthorizationParams, None]) – authorization parameters to be passed when the authorization_type is set to “http_header” or “oauth_2_client_credentials”

Example usage:

# minimum viable instantiation, the rest of the fields can later be added in the UI
http_rest_block = HttpRestBlock(
    reference_name='http_request',
    title='Http Request',
    description='Perform http request',
    method='GET',
)
http_rest_block = HttpRestBlock(
    reference_name='http_request',
    title='Http Request',
    description='Perform http request',
    method='GET',
    endpoint='https://sheets.googleapis.com/v4/spreadsheets/example',
)
# reference to the output of HttpRestBlock which can be passed as input to the next block
# for more information on output() and output references, please take a look at class Block
# and its output() method
output_ref = http_rest_block.output()
class flows_sdk.implementations.idp_v39_1.additional_blocks.HyperscienceRestApiBlock(method, app_endpoint, reference_name=None, headers=None, params=None, payload=None, json=None, handled_error_codes=None, title='Hyperscience HTTP REST Block', description='Hyperscience HTTP REST block')

Bases: BaseHttpRestBlock

Make Http requests to the Hyperscience platform APIs. This block will automatically take care of prepending the base url and will include the appropriate authentication headers with the request. Mandatory parameters:

Parameters:
  • method (str) – HTTP method to use

  • app_endpoint (str) – relative url, beginning with a forward slash, to an HTTP endpoint served by the Hypersciecne platform. For reference on the available endpoints, check out https://docs.hyperscience.com/

Optional Parameters:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • headers (Optional[Dict[str, str]]) – HTTP headers to use

  • params (Optional[Dict[str, Any]]) – key-value pair used as query parameters in the request

  • payload (Optional[Dict[str, str]]) – key-value pair to be used as x-www-form-urlencoded data. Mutually exclusive with json

  • json (Optional[Dict[str, Any]]) – key-value pair to be used as json data. Mutually exclusive with payload

  • handled_error_codes (Optional[List[Union[int, str]]]) – specify HTTP error codes that will not fail the task, allowing for wildcards with the ‘x’ char. Example: [400, “5xx”]

Example usage:

hs_rest_block = HyperscienceRestApiBlock(
    reference_name='http_request',
    app_endpoint='/api/healthcheck',
    method='GET',
)
class flows_sdk.implementations.idp_v39_1.additional_blocks.DatabaseAccessBlock(reference_name=None, db_type='', database='', host='', username='', password='', query='', port=None, options=None, timeout=None, query_params=None, title='DB Block', description='DB access block')

Bases: Block

Database Block allows user to access and perform queries on the specified database Mandatory parameters:

Parameters:
  • db_type (str) – database type (mssql, oracle, postgres)

  • host (str) – URL/IP address of the server database is hosted on

  • database (str) – database name

  • username (str) – database username

  • password (str) – database password

  • query (str) – parameterized query

Optional parameters:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • port (Optional[int]) – database port number

  • options (Optional[Dict[Any, Any]]) – dictionary of additional connection string options

  • query_params (Optional[Dict[Any, Any]]) – dictionary of values for query placeholders

  • timeout (Optional[int]) – timeout in seconds (mssql and postgres only)

Example usage:

# minimum viable instantiation, the rest of the fields can later be added in the UI
db_lookup = DatabaseAccessBlock(
    reference_name='db_lookup',
    title='Database Lookup',
    description='Perform database lookup',
)
# note that storing password in plain text form is not recommended
# password can later be entered in the UI or defined as a flow secret.
db_lookup = DatabaseAccessBlock(
    reference_name='db_lookup',
    title='Database Lookup',
    description='Perform database lookup',
    db_type='mssql',
    database='cars',
    host='example.com',
    username='user',
    # we recommend skipping the password field or defining it as a system secret
    # e.g. system_secret('{your_secret_identifier}') (both allow for secret input via. UI)
    password='pw'
    port=1433,
    timeout=200,
    query='SELECT * from CAR',
)
# reference to the output of DBAccessBlock which can be passed as input to the next block
# for more information on output() and output references, please take a look at class Block
# and its output() method
output_ref = db_lookup.output()
# content of output_ref:
{
    "result": [
        {
            "id": 1,
            "brand": "Aston Martin",
            "model": "Vanquish"
        },
        {
            "id": 2,
            "brand": "Jaguar",
            "model": "XE 2018"
        }
    ]
}

Database output for the same query (SELECT * FROM CAR)

../_images/db_access_query_example.png
IDENTIFIER = 'DB_ACCESS'
class flows_sdk.implementations.idp_v39_1.additional_blocks.LlamaBlock(model_id, prompt, reference_name=None, consent=False, tokenize_only=False, predict_extra_kwargs=None, title='Llama Text Generation', description='Text generation using the Llama 2 model')

Bases: Block

Text generation using the Llama 2 model by Meta.

This block utilizes Llama 2 by Meta. Llama 2 is licensed under the LLAMA 2 Community License (https://ai.meta.com/llama/license/). By using the Llama Block, you acknowledge that you have received a copy of the LLAMA 2 Community License.

Note: This block is still experimental and not ready for production use.

Mandatory parameters:

Parameters:
  • model_id (str) – UUID, identifier of the asset with the ML model. Model assets are not part of the Hypersicence bundle and must first be downloaded and imported into the system before they can be used. Please contact your Hyperscience representative to discuss the usecase and the appropriate model and to get instructions on downloading the models. Once downloaded, here are the available model asset UUIDs: b3da05c7-061c-4a58-af6d-4c905d2e618f - Llama 2, 13B, 8-bit quantized

  • prompt (str) – The prompt to generate text from. Use a CodeBlock ahead of the LlamaBlock in order to build the prompt. The context window of the model is limited to 4096 tokens. If the prompt has more tokens, the block would return an error. The total input and output length also cannot exceed 4096 tokens, which means that if the input is too long, the output may be truncated. Also see tokenize_only for a more detailed explanation of how to handle longer texts.

Optional parameters:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • consent (Optional[bool]) – Has the business user read and acknowledged the Llama 2 legal notice. Unless the flow would be used in CI automation, this should be set to False so that a business user can read and acknowledge the notice in the Hyperscience UI. Setting this to True in the Python definition of the flow is also treated as acknowledgement of the notice on behalf of the entity that will be using the flow.

  • tokenize_only (Optional[bool]) – In the context of LLMs, tokens correspond roughly to words, partial words and punctuation. A rough way to compute the number of tokens in a text is number_of_tokens = 1.25 * number_of_words. The Llama block works with a context window of 4096 tokens. If a longer prompt is passed in, it would return an error. Also, if the prompt is shorter, but still close to 4096, the output may be truncated. Since the length of the prompt in tokens cannot be known ahead of time, sometimes the flow may need to make dynamic decisions whether to pass the text to the Llama block, or to split it into multiple chunks, or to use some alternative method for processing it (e.g. a custom supervision manual task). The tokenize_only option enables flows to take such dynamic decisions based on the length of the prompt in tokens. The default is false - it instructs the block to run the ML model inference over the prompt, i.e. to generate a text response. When true is passed, the model will simply tokenize the prompt without running inference and return the list of tokens - this will work regardless of the length of the prompt. A subsequent Decision task can then route longer prompts to an alternative path in the flow, while routing shorter prompts to the LlamaBlock for actual text generation.

  • predict_extra_kwargs (Optional[Dict[Any, Any]]) – Extra keyword arguments used for model inference. (Used to fine-tune the model prediction. Not recommended to change this unless you know what you are doing)

Example usage:

llama = LlamaBlock(
    reference_name='llama',
    model_id='b3da05c7-061c-4a58-af6d-4c905d2e618f',
    prompt='${previous_block.output.llama_prompt}',
)
IDENTIFIER = 'LLAMA'
class flows_sdk.implementations.idp_v39_1.additional_blocks.EntityRecognitionBlock(pages, regex_map=None, keyword_map=None, enhanced_output=False, reference_name=None, title='Entity Recognition', description='Recognizes entities of different types in documents')

Bases: Block

Entity Recognition Block detects entities of different types within the documents contained in a submission

Mandatory parameters:

Parameters:

pages (str) – list of pages that includes the raw documents that need to be transcribed

Optional Parameters:

Parameters:
  • regex_map (Optional[Dict[str, str]]) – a map that specifies different regex patterns that the block should look for, a default one is present

  • keyword_map (Optional[Dict[str, List[str]]]) – a map that specifies different keyword patterns that the block should look for, a default one is present

  • enhanced_output (bool) – a boolean flag that enables toggling between simple and enhanced output formats. Default value is False.

  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

Example usage:

# This is one way to obtain the submission object (using SubmissionBootstrapBlock)
submission_bootstrap = SubmissionBootstrapBlock(reference_name='submission_bootstrap')

# EntityRecognitionBlock detects entities within the documents contained in the submission
entity_recognition_block = EntityRecognitionBlock(
    reference_name='entity_recognition',
    pages=submission_bootstrap.output(submission.unassigned_pages),
    regex_map=DEFAULT_PII_REGEX_MAP,
    keyword_map=DEFAULT_PII_KEYWORD_MAP,
    enhanced_output=False,
)

# reference to the output of EntityRecognitionBlock which can be passed as input to the
# next block. For more information on output() and output references, please take a look at
# class Block and its output() method
output_ref = entity_recognition_block.output()

Example content of output_ref:

{
    "documents": [
        {
            "file_uuid": "da2c6117-43c4-4b30-84c4-cdc59cc3fa4f",
            "entities": [
                {
                    "text": "John Smith",
                    "type": "Person Name",
                    "positions": [
                        [0.5, 0.1, 0.55, 0.13],
                        [0.6, 0.2, 0.65, 0.23],
                        [0.7, 0.3, 0.75, 0.33],
                    ],
                    "page_ids": [0, 0, 1],
                    "origin": "NER",
                },
                {
                    "text": "<SIGNED>",
                    "type": "signature",
                    "positions": [[0.701171875, 0.732421875, 0.92578125, 0.8681640625]],
                    "page_ids": [0],
                    "origin": "Signature",
                },
            ]
        },
        {
            "file_uuid": "4aa38f84-e5af-4cdc-ba5c-edd60f0a6b89",
            "entities": [
                ...
            ]
        }
    ],
    "redaction_annotations": {
        "ner": [ ... ],
        "ced": [ ... ],
    }
}
IDENTIFIER = 'ENTITY_RECOGNITION'

Trigger Blocks (v39.1)

Note

“Triggers” are referred to as “Input Blocks” in our non-technical documentation.

Warning

Flows are built to allow full definition of triggers either in code or manually in the Flow Studio. We recommend defining triggers manually in the Flow Studio. Instructions on how to do so are in our non-technical documentation.

To enable trigger definition via the Flow Studio UI, you need to instantiate the IDPTriggers() convenience class and pass it to the triggers property of the Flow

return Flow(
   ...
   triggers=IDPTriggers()
)

If instead you wish to define specific triggers in Python, use the trigger definitions below.

Folder Listener (v39.1)

Identifier: FOLDER_TRIGGER

Parameters:

Name

Type

Default Value

Title

Description

folder

Path

N/A

Folder to scan for submissions

N/A

file_extensions

string

N/A

File extensions

List of file extensions to monitor for (e.g.: ‘png, jpg, pdf’)

has_meta

boolean

False

Enable metadata

Select this if a metadata file is to be expected along with document files (in XXX_index.txt file)

poll_interval

integer

10

Poll interval

Poll interval in seconds

warmup_interval

integer

15

Warm-up interval

Seconds to wait past document last modified time before processing it

folder_cleanup_delay

integer

86400

Folder cleanup delay

Seconds to wait before cleaning up subfolders

api_params

object

{}

API Parameters

N/A

folder_listener = IOBlock(
    identifier='FOLDER_TRIGGER',
    reference_name='folder_trigger',
    title='Folder listener',
    enabled=True,
    input={
        'folder': '/var/www/forms/forms/folder/',
        'file_extensions': 'png, pdf, jpg',
    },
)

triggers = IDPTriggers(blocks=[folder_listener])

S3 Listener (v39.1)

Identifier: S3_TRIGGER Parameters:

Name

Type

Default Value

Title

Description

bucket_uri

string

N/A

S3 source URI

S3 URI, where the files will be uploaded (e.g. s3://hs_bucket/prefix1/)

archive_bucket_uri

string

N/A

S3 archive URI

S3 URI to move the processed files to (e.g. s3://hs_bucket/prefix1/)

file_extensions

array

[]

File extensions

Types of file extensions for which to monitor

other_file_extensions

string

N/A

Other file extensions

Comma separated list of file extensions for which to monitor (e.g. ‘png, jpg, pdf’)

has_meta

boolean

false

Include submission level parameters

Select this to ingest JSON files along with document files and submission s3 prefixes. These JSON files can contain information such as metadata, cases and external_id. These JSON file names should match the related files or s3 prefixes (e.g., XXX.jpg.json for XXX.jpg)

aws_access_key_id

string

N/A

AWS ACCESS KEY ID

Access Key used for authentication

aws_secret_access_key

password

N/A

SECRET ACCESS KEY

Secret Key used for authentication

use_ec2_instance_credentials

boolean

true

Use AWS EC2 Instance IAM Role Credentials

If selected, credentials are obtained from the EC2 instance directly, and the AWS Access Key ID and Secret Key are not used

poll_interval

integer

10

Polling interval (in seconds)

How often the S3 Listener will check the S3 bucket for submissions

warmup_interval

integer

15

Warm-up interval (in seconds)

How long the S3 Listener will wait to process the document after it was first detected

api_params

object

{}

API Parameters

N/A

s3_listener = IOBlock(
    identifier='S3_TRIGGER',
    reference_name='s3_trigger',
    title='S3 Listener',
    enabled=True,
    input={
        'bucket_uri': 's3://hs_bucket/prefix1/',
        'archive_bucket_uri': 's3://hs_bucket/prefix2/',
        'file_extensions': ['png', 'pdf', 'jpg'],
        'aws_access_key_id': 'admin',
        'aws_secret_access_key': system_secret('s3_secret_access_key'),
        'use_ec2_instance_credentials': False,
    },
)
triggers = IDPTriggers(blocks=[s3_listener])

Email Listener (v39.1)

Identifier: IMAP_TRIGGER

Parameters:

Name

Type

Default Value

Title

Description

provider [1]

string

imap

Email Provider

N/A

poll_interval

integer

60

Polling interval in seconds

The frequency at which the connection will monitor the email folder for submissions

folder

Path

N/A

Folder to scan for emails

The email folder to retrieve submissions from (e.g., “inbox”)

email_body_treatment

string

“ignore”

Email body treatment

What to do with the body of an email that is ingested. Available options are “process” and “ignore”.

post_process_action

string

“move”

Post process action

What to do with the email after it is processed. Available options are “move” and “delete”.

post_process_move_folder

string

“”

Post process archive folder

Folder to move emails to once they are processed. It will be included only if post_process_action is set to “move”.

api_params

object

{}

API Parameters

N/A

render_headers [2]

array

[]

Headers to include

Headers to render at the top of the email body. Headers will be included only if email_body_treatment is set to “process”.

[1] one of:

imap
graph

[2] any of:

To
From
Subject
Date

Additional inputs when "imap" is the selected provider:

Name

Type

Default Value

Title

Description

host

string

N/A

IMAP server address

N/A

port

integer

993

Port Number

N/A

ssl

boolean

True

Use SSL connection

N/A

username

string

N/A

Username

N/A

password

Password

N/A

Password

N/A

Additional inputs when "graph" is the selected provider:

Name

Type

Default Value

Title

Description

client_id

string

N/A

Client id

Application (client) ID for the corresponding application registered with the Microsoft identity platform.

tenant_id

string

N/A

Tenant id

Also called directory ID. Unique identifier of the tenant in which the corresponding application is registered with.

graph_cloud_endpoint [1]

string

Global

Microsoft Graph Cloud Endpoint

Determines base URL used for accessing cloud services through Microsoft Graph.

azure_ad_endpoint [2]

string

AZURE_PUBLIC_CLOUD

Azure AD Endpoint

Determines base URL for the Azure Active Directory (Azure AD) endpoint to acquire token for each national cloud.

client_secret

Password

N/A

Client secret

Client secret for the corresponding application registered with the Microsoft identity platform.

[1] one of:

Global
US_GOV
US_DoD
Germany
China

[2] one of:

AZURE_PUBLIC_CLOUD
AZURE_CHINA
AZURE_GERMANY
AZURE_GOVERNMENT
imap_trigger = IOBlock(
    identifier='IMAP_TRIGGER',
    reference_name='imap_trigger',
    title='IMAP trigger',
    enabled=True,
    input={
        'host': 'example@mail.com',
        'folder': '/var/www/forms/forms/imap/',
        'username': 'admin',
        'password': 'pass',
    }
)

triggers = IDPTriggers(blocks=[imap_trigger])

Box Folder Listener (v39.1)

Identifier: BOX_FOLDER_TRIGGER

Parameters:

Name

Type

Default Value

Title

Description

source_folder_id

string

N/A

Folder to scan for submissions

Use the Box Folder ID found in the URL

target_folder_id

string

N/A

Folder to move completed files

Use the Box Folder ID found in the URL

file_extensions

array

[]

File extensions

Types of file extensions for which to monitor

custom_file_extensions

string

N/A

Other file extension types

Comma separated list of file extensions for which to monitor (e.g. ‘png, jpg, pdf’)

poll_interval

integer

10

Polling interval (in seconds)

How often the connector will check the base folder for submissions

warmup_interval

integer

15

Warm-up interval (in seconds)

How long the connector will wait to process the document after it was last modified

public_key_id

string

N/A

Public-Key ID

ID of the Public-key used for authentication with Box

private_key

Password

N/A

Private-Key

Private key used for authentication with Box

passphrase

Password

N/A

Passphrase

Passphrase used for authentication with Box

client_id

string

N/A

Client ID

Client Id used for authentication with Box

client_secret

Password

N/A

Client Secret

Client Secret used for authentication with Box

enterprise_id

string

N/A

Enterprise ID

Enterprise Id used for authentication with Box

api_params

object

{}

API Parameters

N/A

box_folder_trigger = IOBlock(
    identifier='BOX_FOLDER_TRIGGER',
    reference_name='box_folder_trigger',
    title='Box folder',
    enabled=True,
    input={
        'file_extensions': ['png', 'pdf', 'jpg'],
        'source_folder_id': '24',
        'target_folder_id': '42',
        'public_key_id': 'admin',
        'private_key': 'secret',
        'passphrase': 'password',
        'client_id': 'admin',
        'client_secret': 'secret',
        'enterprise_id': 'admin',
    }
)

triggers = IDPTriggers(blocks=[box_folder_trigger])

Cron Listener (v39.1)

Identifier: CRON_TRIGGER

Inputs:

Name

Type

Default Value

Title

Description

cron_spec

string

"*/5 * * * *"

Cron specification

N/A

time_zone

string

“US/Eastern”

Time Zone

N/A

api_params

object

{}

API Parameters

N/A

cron_trigger = IOBlock(
    identifier='CRON_TRIGGER',
    reference_name='cron_trigger',
    title='Cron Trigger',
    enabled=True,
    input={
        'cron_spec': '0 10 * * *',
        'time_zone': 'Europe/Sofia',
    }
)

triggers = IDPTriggers(blocks=[cron_trigger])

Salesforce Listener (v39.1)

Identifier: SALESFORCE_TRIGGER

Parameters:

Name

Type

Default Value

Title

Description

channel_name

string

N/A

Channel Name

Configured in the Hyperscience app in Salesforce.

username

string

N/A

Associated Username

Username associated with the Salesforce Connected App.

consumer_key

Password

N/A

Consumer Key

Consumer key used for authentication with Salesforce. Paired with the Salesforce instance from which it was issued.

private_key

Password

N/A

Private Key

Private key used for authentication with Salesforce.

sandbox_environment

boolean

false

Private Key

Enable if the Salesforce environment is a sandbox.

api_params

object

{}

API Parameters

N/A

Message Queue Listener (v39.1)

Identifier: MQ_LISTENER

Inputs:

Name

Type

Default Value

Title

Description

MQ_TYPE

string

“ACTIVE_MQ”

Message Queue type

Type of the message queue to connect with. Valid values are “ACTIVE_MQ”, “AMAZON_SQS”, “IBM_MQ” and “RABBIT_MQ”

MQ_USERNAME

string

null

Username/Access Key ID

N/A

MQ_PASSWORD

Password

null

Password/Secret Access Key

N/A

MQ_QUEUE_NAME

string

null

Queue Name

Name of the queue (topic) to connect to.

MQ_HOST

string

null

Host Name

N/A

MQ_PORT

integer

5672 when “RABBIT_MQ”, 1414 when “IBM_MQ”, otherwise null

Port Number

N/A

api_params

object

{}

API Parameters

N/A

Additional inputs when "AMAZON_SQS" is the selected MQ_TYPE:

Name

Type

Default Value

Title

Description

MQ_REGION

string

null

AWS Region

N/A

MQ_QUEUE_URL

string

null

Queue URL

N/A

MQ_USE_EC2_INSTANCE_CREDENTIALS

boolean

True

Use AWS EC2 Instance IAM Role Credentials

If selected, credentials are obtained from the EC2 instance directly, and the AWS Access Key ID and Secret Key are not used.

NOTE: MQ_USERNAME can be used to provide Access Key ID and MQ_PASSWORD to provide Secret Access Key.

Additional inputs when "IBM_MQ" is the selected MQ_TYPE:

Name

Type

Default Value

Title

Description

MQ_QUEUE_MANAGER

string

null

Queue Manager (IBM MQ)

N/A

MQ_SSL_CIPHER_SUITE [1]

string

null

SSL cipher suite

SSL cipher suite to use if SSL connection is used.

MQ_CHANNEL

string

null

Channel

N/A

MQ_NO_AUTH_REQUIRED

boolean

false

No auth credentials required

N/A

MQ_MQCSP_AUTHENTICATION_MODE

boolean

false

MQCSP authentication mode

N/A

NOTE: MQ_NO_AUTH_REQUIRED is True, MQ_USERNAME and MQ_PASSWORD are optional.

[1] one of:

"NULL_MD5"
"NULL_SHA"
"RC4_MD5_EXPORT"
"RC4_MD5_US"
"RC4_SHA_US"
"RC2_MD5_EXPORT"
"DES_SHA_EXPORT"
"RC4_56_SHA_EXPORT1024"
"DES_SHA_EXPORT1024"
"TRIPLE_DES_SHA_US"
"TLS_RSA_WITH_NULL_SHA256"
"TLS_RSA_WITH_AES_128_CBC_SHA"
"TLS_RSA_WITH_AES_128_CBC_SHA256"
"TLS_RSA_WITH_AES_256_CBC_SHA"
"TLS_RSA_WITH_AES_256_CBC_SHA256"
"AES_SHA_US"
"TLS_RSA_WITH_DES_CBC_SHA"
"TLS_RSA_WITH_3DES_EDE_CBC_SHA"
"TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256"
"FIPS_WITH_DES_CBC_SHA"
"FIPS_WITH_3DES_EDE_CBC_SHA"

Additional inputs when "RABBIT_MQ" is the selected MQ_TYPE:

Name

Type

Default Value

Title

Description

MQ_VIRTUAL_HOST

string

null

Virtual Host (RABBIT MQ)

N/A

MQ_EXCHANGE

string

“”

Exchange

N/A

MQ_ROUTING_KEY

string

“”

Routing key

N/A

MQ_CONNECTION_TYPE

string

“TCP”

Connection Type

“TCP or SSL”

rabbit_mq_listener = IOBlock(
    identifier='MQ_LISTENER',
    reference_name='mq_listener',
    title='RabbitMQ Listener',
    enabled=True,
    input={
        'MQ_TYPE': 'RABBIT_MQ',
        'MQ_QUEUE_NAME': 'some_queue_name',
        'MQ_HOST': 'somehost.com',
        'MQ_USERNAME': 'foo',
        'MQ_PASSWORD': system_secret('rabbit_mq_1_password'),
    }
)

triggers = IDPTriggers(blocks=[rabbit_mq_listener])

Universal Folder Listener (v39.1)

Identifier: UNIVERSAL_FOLDER_TRIGGER

Inputs:

Name

Type

Default Value

Title

Description

folder

string

N/A

Folder to scan for submissions

Enter the path relative to the base folder. Leave blank to monitor the base folder

file_extensions

array

[]

File extensions

Types of file extensions for which to monitor

other_file_extensions

string

N/A

Other file extensions

Comma separated list of file extensions for which to monitor (e.g. ‘png, jpg, pdf’)

has_meta

boolean

false

Include submission level parameters

Select this to ingest JSON files along with document files and submission folders. These JSON files can contain information such as metadata, cases and external_id. These JSON file names should match the related files or folders (e.g., XXX.jpg.json for XXX.jpg)

poll_interval

integer

10

Polling interval (in seconds)

How often the Folder Listener will check the base folder for submissions

warmup_interval

integer

15

Warm-up interval (in seconds)

How long the Folder Listener will wait to process the document after it was last modified

folder_cleanup_delay

number

24

Empty folder cleanup delay (in hours)

How often the Folder Listener will remove empty folders from the base folder

api_params

object

{}

API Parameters

N/A

universal_folder_listener = IOBlock(
    identifier='UNIVERSAL_FOLDER_TRIGGER',
    reference_name='universal_folder_trigger',
    title='Universal Folder Listener',
    enabled=True,
    input={
        'folder': '/var/www/forms/forms/folder/',
        'file_extensions': 'png, pdf, jpg',
    }
)

triggers = IDPTriggers(blocks=[universal_folder_listener])

Kafka Listener (v39.1)

Identifier: KAFKA_TRIGGER

Inputs:

Name

Type

Default Value

Title

Description

bootstrap_servers

string

N/A

Bootstrap Servers

A Comma separated list of host:port pairs to use for establishing the initial connection to the cluster

group_id

string

N/A

Group Id

A unique string that identifies the consumer group this consumer belongs to. BLOCK_THREADS_KAFKA_TRIGGER should be used to scale up or down the number of consumers (default: number of available cores)

topics

string

N/A

Topics

Comma separated list of topic names that the block will subscribe to

consumer_config_json

json

N/A

Consumer Config

JSON formatted string containing key-value pairs of Kafka consumer configurations described in the official Kafka documentation

api_params

object

{}

API Parameters

N/A

consume_timeout

integer

-1

Consume Timeout

Maximum time, in seconds, that the consumer will wait for new messages. If messages are received within the specified time, they are processed immediately. However, if no messages are received within that specified time, consumption cycle will be restarted internally to continue fetching messages. This ensures continuous message consumption (default: infinite)

consume_num_messages

integer

1

Consume Num Messages

Sets the maximum number of messages the consumer fetches in a single consumption cycle.If the number of available messages is less than the specified, and timeout is reached, messages will be processed. Consumption cycle will be restarted after the current finishes

kafka_listener = IOBlock(
    identifier='KAFKA_TRIGGER',
    reference_name='kafka_trigger',
    title='Kafka Listener',
    enabled=True,
    input={
        'folder': '/var/www/forms/forms/folder/',
        'file_extensions': 'png, pdf, jpg',
         'bootstrap_servers': 'example.com',
         'group_id': 'test_group',
         'topics': 'test_topic1, test_topic2',
         'consumer_config_json': {},
         'number_of_workers': 1,
         'consume_timeout': 5,
         'consume_num_messages': 10
    }
)

triggers = IDPTriggers(blocks=[kafka_listener])

Output Blocks (v39.1)

Warning

Flows pass information to downstream systems using the IDPOutputsBlock. This block is specially built to allow full definition either in code or manually in the Flow Studio. We recommend defining outputs manually in the Flow Studio.

Instructions on how to do so are in our non-technical documentation. Simply include an IDPOutputsBlock in your Python flow definition wherever you wish to send information to downstream systems.

class flows_sdk.implementations.idp_v39_1.idp_blocks.IDPOutputsBlock(inputs, blocks=None)

Bases: Outputs

Output block allows users to send data extracted by an IDP flow to other systems for downstream processing

Mandatory parameters:

Parameters:

inputs (Dict[str, Any]) – Used by the UI to automatically pre-populate the input of newly added blocks. Extended in this class to include 'enabled': True to visualize an enabled/disabled trigger that is controllable via Flow Studio.

Usage when parameters are to be defined in Flow Studio:

outputs = IDPOutputBlock(
    inputs={'submission': submission_bootstrap.output('result.submission')}
)

If instead you wish to define specific outputs in Python, use the outputs definitions below.

HTTP Notifier Output (v39.1)

The HTTP Notifier (REST) output connection will POST results from the system to a specified HTTP endpoint.

Identifier: COB_HTTP_EXPORT

Parameters:

Name

Type

Default Value

Title

Description

submission

Submission

N/A

Submission Object

N/A

enabled

boolean

True

Enabled

Enable or disable this output block

api_version [1]

string

N/A

API Version

API version to be used for rendering payload

export_type

string

N/A

Export Type

Payload rendering choice. If “Flat” is chosen, a simplified version of the payload will be sent, as described in the “flat” parameter for submission retrieval API

export_routing_filter

string

null

Routing Filter

By default, status change notifications are sent for each submission processed by the system. You can filter what notifications are received through this connection by specifying an expression like has_layout_tag(‘some_layout_tag’) or has_source_routing_tag(‘some_source_routing_tag’)

endpoint

string

N/A

Endpoint URL

URL that output notifications will be posted to

endpoint_secrets

string

N/A

Endpoint URL secrets

Endpoint URL Secrets, format: (one per line) secret_key_name=secret_value

timeout

int

600

Timeout (seconds)

How long (in seconds) to keep the connection open if no bytes are received from the endpoint. Set to 0 to wait indefinitely.

authorization_type [2]

string

“none”

Authorization type

Type of authorization

authorization_header_name

string

null

Authorization header name

Authorization header name to be set in the notification request

authorization_header

Password

null

Authorization header value

Authorization header to be set in the notification request

auth_url

string

null

OAuth2 authorization URL

The endpoint for the authorization server

client_id

string

null

Client ID

The client identifier issued to the client during the application registration process

client_secret

Password

null

Client Secret

The client secret issued to the client during the application registration process

scope

string

null

Scope

Scope to request to oauth server

audience

string

null

Audience

Resource service URL where token will be valid

additional_oauth_parameters

json

{}

Additional Oauth Request Parameters

Additional parameters to send when requesting token

ssl_cert_source [3]

string

null

Source

N/A

ssl_cert_path

string

null

Trusted Certificate(s) path

Enter the path relative to the base folder

ssl_cert_value

MultilineText

null

Trusted Certificate(s)

N/A

additional_notification_http_headers

json

null

Additional Notification Request Headers

Key value pairs in json format which will be sent via the request as http headers

[1] one of:

"v5"
"transformed"

[2] one of:

"none"
"http_header"
"oauth_2_client_credentials"

[3] one of:

"env_var"
"ca_bundle_path"
"certificate_value"
cob_http_export = IOBlock(
    identifier='COB_HTTP_EXPORT',
    reference_name='cob_http_export',
    title='HTTP Notifier Output Block',
    enabled=True,
    input={
        'submission': submission_bootstrap.output('result.submission'),
        'enabled': True,
        'api_version': 'v5',
        'export_type': 'SUBMISSION',
        'endpoint': 'example.com',
        'authorization_type': 'none',
    }
)

outputs = IDPOutputsBlock(
    inputs={'submission': submission_bootstrap.output('result.submission')},
    blocks=[cob_http_export],
)

Message Queue Notifier Output (v39.1)

The Message Queue Notifier Output can configure connections to ActiveMQ, Amazon SQS, IBM MQ, and RabbitMQ message queues.

Identifier: COB_MQ_NOTIFIER

Parameters:

Name

Type

Default Value

Title

Description

submission

Submission

N/A

Submission Object

N/A

enabled

boolean

True

Enabled

Enable or disable this output block

api_version [1]

string

N/A

API Version

API version to be used for rendering payload

export_type

string

N/A

Export Type

Payload rendering choice. If “Flat” is chosen, a simplified version of the payload will be sent, as described in the “flat” parameter for submission retrieval API

export_routing_filter

string

null

Routing Filter

By default, status change notifications are sent for each submission processed by the system. You can filter what notifications are received through this connection by specifying an expression like has_layout_tag(‘some_layout_tag’) or has_source_routing_tag(‘some_source_routing_tag’)

MQ_TYPE

string

“ACTIVE_MQ”

Message Queue type

Type of the message queue to connect with. Valid values are “ACTIVE_MQ”, “AMAZON_SQS”, “IBM_MQ” and “RABBIT_MQ”

MQ_USERNAME

string

null

Username/Access Key ID

N/A

MQ_PASSWORD

Password

null

Password/Secret Access Key

N/A

MQ_QUEUE_NAME

string

null

Queue Name

Name of the queue (topic) to connect to. It applies only to “ACTIVE_MQ”, “IBM_MQ” and “RABBIT_MQ” MQ_TYPEs.

MQ_HOST

string

null

Host Name

N/A

MQ_PORT

integer

5672 when “RABBIT_MQ”, 1414 when “IBM_MQ”, otherwise null

Port Number

N/A

[1] one of:

"v5"
"transformed"

Additional inputs when "AMAZON_SQS" is the selected MQ_TYPE:

Name

Type

Default Value

Title

Description

MQ_REGION

string

null

AWS Region

N/A

MQ_QUEUE_URL

string

null

Queue URL

N/A

MQ_USE_EC2_INSTANCE_CREDENTIALS

boolean

True

Use AWS EC2 Instance IAM Role Credentials

If selected, credentials are obtained from the EC2 instance directly, and the AWS Access Key ID and Secret Key are not used.

MQ_MESSAGE_GROUP_ID

string

null

Group ID for FIFO queues

N/A

MQ_MESSAGE_METADATA

string

null

Additional SQS Metadata

Input additional metadata key value pairs in the format of { “key”: {“TYPE”: “String”, “VALUE”: “”}, “key2”: {“TYPE”: “Number”, “VALUE”: “12345”}

NOTE: MQ_USERNAME can be used to provide Access Key ID and MQ_PASSWORD to provide Secret Access Key.

Additional inputs when "IBM_MQ" is the selected MQ_TYPE:

Name

Type

Default Value

Title

Description

MQ_QUEUE_MANAGER

string

null

Queue Manager (IBM MQ)

N/A

MQ_SSL_CIPHER_SUITE [1]

string

null

SSL cipher suite

SSL cipher suite to use if SSL connection is used.

MQ_CHANNEL

string

null

Channel

N/A

MQ_NO_AUTH_REQUIRED

boolean

false

Channel

N/A

MQ_MQCSP_AUTHENTICATION_MODE

boolean

false

Channel

N/A

NOTE: MQ_NO_AUTH_REQUIRED is True, MQ_USERNAME and MQ_PASSWORD are optional.

[1] one of:

"NULL_MD5"
"NULL_SHA"
"RC4_MD5_EXPORT"
"RC4_MD5_US"
"RC4_SHA_US"
"RC2_MD5_EXPORT"
"DES_SHA_EXPORT"
"RC4_56_SHA_EXPORT1024"
"DES_SHA_EXPORT1024"
"TRIPLE_DES_SHA_US"
"TLS_RSA_WITH_NULL_SHA256"
"TLS_RSA_WITH_AES_128_CBC_SHA"
"TLS_RSA_WITH_AES_128_CBC_SHA256"
"TLS_RSA_WITH_AES_256_CBC_SHA"
"TLS_RSA_WITH_AES_256_CBC_SHA256"
"AES_SHA_US"
"TLS_RSA_WITH_DES_CBC_SHA"
"TLS_RSA_WITH_3DES_EDE_CBC_SHA"
"TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256"
"FIPS_WITH_DES_CBC_SHA"
"FIPS_WITH_3DES_EDE_CBC_SHA"

Additional inputs when "RABBIT_MQ" is the selected MQ_TYPE:

Name

Type

Default Value

Title

Description

MQ_VIRTUAL_HOST

string

null

Virtual Host (RABBIT MQ)

N/A

MQ_EXCHANGE

string

“”

Exchange

N/A

MQ_ROUTING_KEY

string

“”

Routing key

N/A

MQ_CONNECTION_TYPE

string

“TCP”

Connection Type

“TCP or SSL”

mq_notifier = IOBlock(
    identifier='COB_MQ_NOTIFIER',
    reference_name='mq_notifier',
    title='MQ Notifier Output Block',
    enabled=True,
    input={
        'submission': submission_bootstrap.output('result.submission'),
        'enabled': True,
        'api_version': 'v5',
        'export_type': 'SUBMISSION',
        'MQ_USERNAME': 'admin',
        'MQ_PASSWORD': 'pass',
        'MQ_QUEUE_NAME': 'queue',
        'MQ_HOST': 'host',
    },
)

outputs = IDPOutputsBlock(
    inputs={'submission': submission_bootstrap.output('result.submission')},
    blocks=[mq_notifier],
)

Box Notifier Output (v39.1)

Provides an out-of-the-box integration into Box systems.

Identifier: COB_BOX_NOTIFIER

Parameters:

Name

Type

Default Value

Title

Description

submission

Submission

N/A

Submission Object

N/A

enabled

boolean

True

Enabled

Enable or disable this output block

api_version [1]

string

N/A

API Version

API version to be used for rendering payload

export_type

string

N/A

Export Type

Payload rendering choice. If “Flat” is chosen, a simplified version of the payload will be sent, as described in the “flat” parameter for submission retrieval API

export_routing_filter

string

null

Routing Filter

By default, status change notifications are sent for each submission processed by the system. You can filter what notifications are received through this connection by specifying an expression like has_layout_tag(‘some_layout_tag’) or has_source_routing_tag(‘some_source_routing_tag’)

public_key_id

string

N/A

Public-Key ID

ID of the Public-key used for authentication with Box

private_key

Password

N/A

Private-Key

Private key used for authentication with Box

passphrase

Password

N/A

Passphrase

Passphrase used for authentication with Box

client_id

string

N/A

Client ID

Client Id used for authentication with Box

client_secret

Password

N/A

Client Secret

Client Secret used for authentication with Box

enterprise_id

string

N/A

Enterprise ID

Enterprise Id used for authentication with Box

template_key

string

“”

Box Metadata Template Key

Enter the key of the Box Metadata template that you would like to map the Hyperscience metadata to.

static_output_fields_to_map

array

[]

Static Metadata Fields

Specify the Hyperscience fields you want to store in Box metadata.

submission_id

string

“”

Key for Mapping Submission ID

Specify a Box metadata field key from your chosen Box template. The Hyperscience field will be stored in this specified field.

document_id

string

“”

Key for Mapping Document ID

Specify a Box metadata field key from your chosen Box template. The Hyperscience field will be stored in this specified field.

submission_state

string

“”

Key for Mapping Submission State

Specify a Box metadata field key from your chosen Box template. The Hyperscience field will be stored in this specified field.

document_state

string

“”

Key for Mapping Document State

Specify a Box metadata field key from your chosen Box template. The Hyperscience field will be stored in this specified field.

submission_exceptions

string

“”

Key for Mapping Submission Exceptions

Specify a Box metadata field key from your chosen Box template. The Hyperscience field will be stored in this specified field.

document_exceptions

string

“”

Key for Mapping Document Exceptions

Specify a Box metadata field key from your chosen Box template. The Hyperscience field will be stored in this specified field.

document_layout_uuid

string

“”

Key for Mapping Document Layout Uuid

Specify a Box metadata field key from your chosen Box template. The Hyperscience field will be stored in this specified field.

document_layout_name

string

“”

Key for Mapping Document Layout Name

Specify a Box metadata field key from your chosen Box template. The Hyperscience field will be stored in this specified field.

start_time

string

“”

Key for Mapping Start Time

Specify a Box metadata field key from your chosen Box template. The Hyperscience field will be stored in this specified field.

complete_time

string

“”

Key for Mapping Complete Time

Specify a Box metadata field key from your chosen Box template. The Hyperscience field will be stored in this specified field.

document_fields_template_mappings

json

{}

JSON for Additional Metadata Fields

Specify additional metadata fields using their key value. Please consult Box Integration setup manual for JSON template and instructions.

[1] one of:

"v5"
"transformed"
box_notifier = IOBlock(
    identifier='COB_BOX_NOTIFIER',
    reference_name='box_notifier',
    title='Box Notifier Output Block',
    enabled=True,
    input={
        'submission': submission_bootstrap.output('result.submission'),
        'enabled': True,
        'private_key': 'password',
        'passphrase': 'password',
        'public_key_id': 'admin',
        'client_id': 'admin',
        'client_secret': 'secret',
        'enterprise_id': 'admin',
        'template_key': 'key',
    },
)

outputs = IDPOutputsBlock(
    inputs={'submission': submission_bootstrap.output('result.submission')},
    blocks=[box_notifier],
)

UiPath Notifier Output (v39.1)

Provides an out-of-the-box integration into UiPath systems.

Identifier: COB_EXPORT_UIPATH

Parameters:

Name

Type

Default Value

Title

Description

submission

Submission

N/A

Submission Object

N/A

enabled

boolean

True

Enabled

Enable or disable this output block

api_version [1]

string

N/A

API Version

API version to be used for rendering payload

export_type

string

N/A

Export Type

Payload rendering choice. If “Flat” is chosen, a simplified version of the payload will be sent, as described in the “flat” parameter for submission retrieval API

export_routing_filter

string

null

Routing Filter

By default, status change notifications are sent for each submission processed by the system. You can filter what notifications are received through this connection by specifying an expression like has_layout_tag(‘some_layout_tag’) or has_source_routing_tag(‘some_source_routing_tag’)

auth_method [2]

string

“basic_auth”

Authentication method

N/A

url

string

N/A

URL

Base URL of the uipath instance

app_id

string

N/A

App ID

App ID provided by UiPath

app_secret

Password

N/A

App Secret

App Secret provided by UiPath

oauth_token_endpoint

string

N/A

OAuth token endpoint

Endpoint to retrieve the token. Using https://cloud.uipath.com/identity_/connect/token by default

organization_unit_id

string

N/A

Organization unit id

Organization unit id

tenant

string

N/A

Tenant

Tenant

username

string

N/A

Username

Username

password

Password

N/A

Password

Password

queue

string

N/A

Queue

Queue

[1] one of:

"v5"
"transformed"

[2] one of:

"basic_auth"
"oauth"
cob_export_uipath = IOBlock(
    identifier='COB_EXPORT_UIPATH',
    reference_name='cob_export_uipath',
    title='CCB Export UiPath',
    enabled=True,
    input={
        'submission': submission_bootstrap.output('result.submission'),
        'enabled': True,
        'api_version': 'v5',
        'export_type': 'FLAT_SUBMISSION',
        'url': 'example.com',
        'tenant': 'tenant',
        'username': 'admin',
        'password': 'pass',
        'queue': 'queue',
    }
)

outputs = IDPOutputsBlock(
    inputs={'submission': submission_bootstrap.output('result.submission')},
    blocks=[cob_export_uipath],
)

Salesforce Notifier Output (v39.1)

With the Salesforce Notifier Block, you can configure your flow to send extracted information to Salesforce. The Salesforce Notifier Block can use extracted information to look up, and then update any number of fields on a Salesforce object. The Salesforce Notifier Block is available in both SaaS and on-premise versions of the Hyperscience application.

Identifier: COB_SALESFORCE_NOTIFIER

Parameters:

Name

Type

Default Value

Title

Description

submission

Submission

N/A

Submission Object

N/A

enabled

boolean

True

Enabled

Enable or disable this output block

api_version [1]

string

N/A

API Version

API version to be used for rendering payload

export_type

string

N/A

Export Type

Payload rendering choice. If “Flat” is chosen, a simplified version of the payload will be sent, as described in the “flat” parameter for submission retrieval API

export_routing_filter

string

null

Routing Filter

By default, status change notifications are sent for each submission processed by the system. You can filter what notifications are received through this connection by specifying an expression like has_layout_tag(‘some_layout_tag’) or has_source_routing_tag(‘some_source_routing_tag’)

username

string

N/A

Associated Username

Username associated with the Salesforce Connected App

consumer_key

password

N/A

Consumer Key

Consumer key used for authentication with Salesforce. Paired with the Salesforce instance from which it was issued.

private_key

password

N/A

Private Key

Private key used for authentication with Salesforce.

sandbox_environment

boolean

false

Private Key

Enable if the Salesforce environment is a sandbox.

object_api_name

string

N/A

Object API Name

API name of the object in Salesforce to lookup and update.

new_record_on_lookup_failure

boolean

false

Create new record on lookup failure

When enabled, it will create a new record if the lookup fails to find a record using the lookup parameters. When disabled, lookup will fail if it does not find a record using the lookup parameters.

object_lookup_mapping

json

N/A

Lookup Field Mapping

Specify which Salesforce fields (using their API name) to be queried with which extracted values from the document for the lookup using key-value format. For example: “DOCUMENT_FIELD_NAME”: “OBJECT_FIELD_API_NAME”.

document_fields_mapping

json

N/A

Document Field Mappings

Specify which Salesforce fields (using their API name) you want to update with which extracted data from Hyperscience using key-value format. For example: “DOCUMENT_FIELD_NAME”: “OBJECT_FIELD_API_NAME”.”

overwrite_existing_field_data

boolean

false

Overwrite existing field data

If this is checked and a selected field has data in it, Hyperscience will overwrite that data. If this is unchecked, existing data will not be overwritten and the operation will fail.

[1] one of:

"v5"
"transformed"
cob_salesforce_notifier = IOBlock(
    identifier='COB_SALESFORCE_NOTIFIER',
    reference_name='cob_salesforce_notifier',
    title='Salesforce Notifier Output Block',
    enabled=True,
    input={
        'submission': submission_bootstrap.output('result.submission'),
        'enabled': True,
        'username': 'salesforce_user',
        'consumer_key': 'salesforce_consumer_key',
        'private_key': 'salesforce_private_key',
        'object_api_name': 'salesforce_object_api_name',
        'object_lookup_mapping': {"LAYOUT_VARIATION_XYZ_FIELD_NAME_1": "SF_FIELD_API_NAME_1", "LAYOUT_VARIATION_XYZ_FIELD_NAME_2": "SF_FIELD_API_NAME_2"},
        'document_fields_mapping': {"LAYOUT_VARIATION_XYZ_FIELD_NAME_3": "SF_FIELD_API_NAME_3", "LAYOUT_VARIATION_XYZ_FIELD_NAME_4": "SF_FIELD_API_NAME_4"}
    }
)

outputs = IDPOutputsBlock(
    inputs={'submission': submission_bootstrap.output('result.submission')},
    blocks=[cob_salesforce_notifier],
)

S3 Notifier Output (v39.1)

With the S3 Notifier Block, you can configure your flow to send extracted information to an S3 bucket.

Identifier: COB_S3_NOTIFIER

Parameters:

Name

Type

Default Value

Title

Description

submission

Submission

N/A

Submission Object

N/A

enabled

boolean

True

Enabled

Enable or disable this output block

api_version [1]

string

N/A

API Version

API version to be used for rendering payload

export_type

string

N/A

Export Type

Payload rendering choice. If “Flat” is chosen, a simplified version of the payload will be sent, as described in the “flat” parameter for submission retrieval API

export_routing_filter

string

null

Routing Filter

By default, status change notifications are sent for each submission processed by the system. You can filter what notifications are received through this connection by specifying an expression like has_layout_tag(‘some_layout_tag’) or has_source_routing_tag(‘some_source_routing_tag’)

bucket_uri

string

N/A

S3 URI

S3 URI, where the data should be stored (e.g. s3://hs_bucket/prefix1/prefix2/)

aws_access_key_id

string

N/A

AWS ACCESS KEY ID

Access Key used for authentication

aws_secret_access_key

password

N/A

SECRET ACCESS KEY

Secret Key used for authentication

use_ec2_instance_credentials

boolean

true

Use AWS EC2 Instance IAM Role Credentials

If selected, credentials are obtained from the EC2 instance directly, and the AWS Access Key ID and Secret Key are not used

[1] one of:

"v5"
"transformed"
cob_s3_notifier = IOBlock(
    identifier='COB_S3_NOTIFIER',
    reference_name='cob_s3_notifier',
    title='S3 Notifier Output Block',
    enabled=True,
    input={
        'submission': submission_bootstrap.output('result.submission'),
        'enabled': True,
        'api_version': 'v5',
        'export_type': 'FLAT_SUBMISSION',
        'bucket_uri': 's3://hs_bucket/prefix1/prefix2/',
        'aws_access_key_id': 'aws_access',
        'aws_secret_access_key': system_secret('aws_secret_access_key'),
    }
)

outputs = IDPOutputsBlock(
    inputs={'submission': submission_bootstrap.output('result.submission')},
    blocks=[cob_s3_notifier],
)

Kafka Notifier Output (v39.1)

With the Kafka Notifier Block, you can configure your flow to send extracted information to a kafka topic.

Identifier: COB_KAFKA_NOTIFIER

Parameters:

Name

Type

Default Value

Title

Description

submission

Submission

N/A

Submission Object

N/A

enabled

boolean

True

Enabled

Enable or disable this output block

api_version [1]

string

N/A

API Version

API version to be used for rendering payload

export_type

string

N/A

Export Type

Payload rendering choice. If “Flat” is chosen, a simplified version of the payload will be sent, as described in the “flat” parameter for submission retrieval API

export_routing_filter

string

null

Routing Filter

By default, status change notifications are sent for each submission processed by the system. You can filter what notifications are received through this connection by specifying an expression like has_layout_tag(‘some_layout_tag’) or has_source_routing_tag(‘some_source_routing_tag’)

bootstrap_servers

string

N/A

Bootstrap Servers

A list of Kafka brokers to connect to. For example: “kafka1:9092,kafka2:9092”

topic

string

N/A

Topic

The Kafka topic where your data will be sent

client_id

string

N/A

SECRET ACCESS KEY

An identifier for your application when connecting to Kafka, typically used for tracking and logging purposes

producer_config_json

json

true

Producer Config (JSON)

Configuration settings for the Kafka producer in JSON format. See https://kafka.apache.org/documentation/#producerconfigs for more information

synchronous_publish

boolean

true

Sync Publish

If true, the publish call will block until the message is sent successfully or an error is raised. If false, the publish call will return immediately after enqueueing the message

message_key

string

null

Message Key

The key associated with the message to be sent. This is used to determine which partition the message is sent to. If null, the message will be sent to a random partition

[1] one of:

"v5"
"transformed"
cob_kafka_notifier = IOBlock(
    identifier='COB_KAFKA_NOTIFIER',
    reference_name='cob_kafka_notifier',
    title='Kafka Notifier Output Block',
    enabled=True,
    input={
        'submission': submission_bootstrap.output('result.submission'),
        'enabled': True,
        'api_version': 'v5',
        'export_type': 'FLAT_SUBMISSION',
        'bootstrap_servers': 'kafka1:9092,kafka2:9092',
        'topic': 'topic',
        'client_id': 'client_id',
        'producer_config_json': {},
        'synchronous_publish': True,
        'message_key': 'key'
    }
)

outputs = IDPOutputsBlock(
    inputs={'submission': submission_bootstrap.output('result.submission')},
    blocks=[cob_kafka_notifier],
)

Helper Classes and Functions (v39.1)

The following classes are used to easily instantiate and manage a flow’s settings.

flows_sdk.implementations.idp_v39_1.idp_values.get_idp_wf_inputs(idp_wf_config)

A helper function that maps the values provided by an IdpWorkflowConfig to their corresponding IDP keys.

Parameters:

idp_wf_config (IdpWorkflowConfig) – with static values to be filled.

Return type:

Dict[str, Any]

Returns:

a dictionary with filled static values, compatible with IDP flows.

Example usage with default values:

flow = Flow(
    input=get_idp_wf_inputs(get_idp_wf_config()),
    ...
)

Example usage:

Example usage when instantiating an IDP flow with custom parameters::

flow = Flow(
    input=get_idp_wf_inputs({an instance of IdpWorkflowConfig}),
    ...
)
flows_sdk.implementations.idp_v39_1.idp_values.get_idp_wf_config()

A helper function that instantiates an IdpWorkflowConfig pre-filled with default values. IdpWorkflowConfig is the container class for defining flow-level settings.

Return type:

IdpWorkflowConfig

Returns:

a valid, fully constructed IdpWorkflowConfig instance

Example usage when instantiating an IDP flow:

flow = Flow(
    input=get_idp_wf_inputs(get_idp_wf_config()),
    ...
)
class flows_sdk.implementations.idp_v39_1.idp_values.IdpWorkflowConfig(rotation_correction_enabled, document_grouping_logic, qa_config, manual_document_organization_enabled, transcription_automation_training, transcription_training_legibility_period, transcription_model, manual_transcription_enabled, finetuning_only_trained_layouts, flex_confidence_boosting_enabled, improved_transcription_threshold_accuracy, structured_text, semi_structured_text, semi_structured_table, structured_checkbox, structured_signature, semi_structured_checkbox, field_id_target_accuracy, table_id_target_accuracy, manual_field_id_enabled, machine_classification_config, manual_transcription_config, manual_identification_config, flexible_extraction_config, notification_workflow_initialization=None, notification_workflow_classification=None, notification_workflow_identification=None, notification_workflow_transcription=None, notification_workflow_flexible_extraction=None)

Convenience dataclass outlining all flow-level config values for IDP.

class flows_sdk.implementations.idp_v39_1.idp_values.IDPTriggers(blocks=None)

An adapter class making it easier to instantiate flows_sdk.flows.Triggers for IDP flows. In particular, provides defaults for some fields (e.g. title, description, reference_name) and defines api_params_manifest with IDP specific parameters.

Parameters:

blocks (Optional[Sequence[IOBlock]]) – flows_sdk.blocks.IOBlock s to be included as triggers for an IDP Flow. Optional, defaults to an empty list.

Example usage when instantiating an IDP flow:

folder_listener = IOBlock(...)
flow = Flow(
    triggers=IDPTriggers([folder_listener]),
    ...
)
class flows_sdk.implementations.idp_v39_1.idp_values.IDPWrapperManifest(flow_identifier)

An adapter class making it easier to instantiate flows_sdk.flows.Manifest for IDP wrapper flows. In particular, provides defaults for some fields (e.g. roles) but more importantly, defines all flows_sdk.flows.Parameter that construct the Flow inputs and their UI representation (left-hand side menu in Flow studio).

Parameters:

flow_identifier (str) – A system-wide unique identifier for flows_sdk.flows.Flow

Example usage when instantiating an IDP flow:

flow = Flow(
    manifest=IDPWrapperManifest(flow_identifier='FLOW_IDENTIFIER'),
    ...
)
class flows_sdk.implementations.idp_v39_1.idp_values.IDPCoreManifest(flow_identifier, roles=None)

An adapter class making it easier to instantiate flows_sdk.flows.Manifest for IDP Core flows. In particular, provides defaults for some fields (e.g. roles, output) but more importantly, defines all flows_sdk.flows.Parameter that construct the Flow inputs and their UI representation (left-hand side menu in Flow studio).

Parameters:
  • flow_identifier (str) – A system-wide unique identifier for flows_sdk.flows.Flow

  • roles (Optional[List[str]]) – ‘read_only’, ‘supporting’ and ‘idp_core’ are the default roles of every subflow that can replace the default IDP_CORE subflow inside the top-level IDP flows. ‘supporting’ and ‘idp_core’ are mandatory roles that are necessary for all IDP Core flows. ‘supporting’ ensures that the subflow will be auto-deployed together with the top-level IDP flow. ‘idp_core’ ensures that this subflow can be selected in the UI in dropdowns which allow picking an alternative subflow to the default IDP_CORE subflow. If you are implementing a subflow that is not supposed to be a drop-in replacement for IDP_CORE, you will have to subclass or implement your own version of IDPCoreManifest class in order to remove the idp_core role.

Example usage when instantiating an IDP flow:

flow = Flow(
    manifest=IDPCoreManifest(flow_identifier='FLOW_IDENTIFIER'),
    ...
)

V39

Processing Blocks (v39)

Core Blocks are powerful processing blocks necessary to build intelligent document processing solutions on Hyperscience. They include both manual and machine processing blocks.

Note

Classes in the idp_v39 implementation have parameters with ambiguous typing. For example, a field of type integer could also be represented as a string reference (e.g. that integer gets provided at runtime). The types introduced in the example are therefore incomplete and are likely to be revisioned in a future version.

class flows_sdk.implementations.idp_v39.idp_blocks.IDPCoreBlock(idp_wf_config, reference_name='__idp_core__', title='Start Document Processing Subflow', description='Calls the Document Processing Subflow with the specified settings.', compatibility_spec=CompatibilitySpec(filter_roles=['idp_core'], filter_schema=None, label=None), identifier='IDP_CORE_V39', additional_inputs=None)

Bases: Block

IDP core block includes the steps to process the passed documents into machine-readable output.

Mandatory parameters:

Parameters:

idp_wf_config (IdpWorkflowConfig) – IdpWorkflowConfig object with all flow-level configuration settings for IDP

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique reference name within a flow, a unique one will be generated when not provided

  • compatibility_spec (Optional[CompatibilitySpec]) – defines a filter for flows that could be selected in the UI in place of “IDP Core” for the block. Defaults to flows with the role ‘idp_core’. The user will be able to pick any flow that matches the compatibility_spec. This serves to make upgrading to a new compatible version of “IDP Core” (e.g. IDP_CORE_V39) easier in future releases.

  • identifier (str) – The identifier of the flow. Defaults to the IDP_CORE_VXX identifier of the out-of-the-box flow that comes with the system.

  • additional_inputs (Optional[Dict[str, Any]]) – Dictionary containing key-value pairs that will be added to the inputs of the block. If a key is contained in the default inputs, its value will be overwritten by the value provided here. You generally do not need to provide additional_inputs, unless you are creating a custom derivative of the IDP_CODE_Vxx flow and are modifying or extending the list of inputs it accepts.

Example usage:

idp_core = IDPCoreBlock(idp_wf_config=idp_wf_config)
REFERENCE_NAME = '__idp_core__'
class flows_sdk.implementations.idp_v39.idp_blocks.SubmissionBootstrapBlock(reference_name=None, layout_release_uuid='${workflow.input.layout_release_uuid}', s3_config='${workflow.input.submission_bootstrap_s3_config}', s3_endpoint_url='${workflow.input.submission_bootstrap_s3_endpoint_url}', ocs_config='${workflow.input.submission_bootstrap_ocs_config}', http_config='${workflow.input.submission_bootstrap_http_config}', notification_workflow='${workflow.input.notification_workflow_initialization}', workflow_uuid='${workflow.input.workflow_uuid}', workflow_name='${workflow.input.workflow_name}', workflow_version='${workflow.input.workflow_version}', submission='${workflow.input.submission}')

Bases: Block

Submission bootstrap block initializes the submission object and prepares external images or other submission data if needed. This block should be called at the START of every flow that uses blocks included in the library.

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • layout_release_uuid (str) – Layout Release UUID at submission creation time, defaults to workflow_input(Settings.LayoutReleaseUuid)

  • s3_config (str) – Specify Amazon S3 credentials for submission retrieval store. Expects a json expression: {"aws_access_key_id": "X", "aws_secret_access_key": "Y"}, defaults to workflow_input(Settings.SubmissionBootstrapS3Config)

  • s3_endpoint_url (str) – Endpoint URL for S3 submission retrieval store, defaults to workflow_input(Settings.SubmissionBootstrapS3EndpointUrl)

  • ocs_config (str) – OCS Configuration for downloading submission data, defaults to workflow_input(Settings.SubmissionBootstrapOCSConfig)

  • http_config (Optional[str]) – An optional system secret field expressing the HTTP Configuration for downloading submission data. The content is expected to be a json formatted as: {"username": "X", "password": "Y", "ssl_cert": "CA bundle filename"}, defaults to workflow_input(Settings.SubmissionBootstrapHTTPConfig)

  • notification_workflow (Optional[str]) – Notification flow name to run when the submission is initialized, defaults to workflow_input(Settings.SubmissionBootstrapNotificationWorkflow)

  • workflow_uuid (str) – UUID of the triggered workflow version, defaults to workflow_input(Inputs.WorkflowUuid)

  • workflow_name (str) – Name of the triggered workflow, defaults to workflow_input(Inputs.WorkflowName)

  • workflow_version (str) – Version of the triggered workflow, defaults to workflow_input(Inputs.WorkflowVersion)

  • submission (Any) – Submission object

Example usage:

submission_bootstrap = SubmissionBootstrapBlock(reference_name='submission_bootstrap')
IDENTIFIER = 'SUBMISSION_BOOTSTRAP_2'
output(key=None)

Submission bootstrap ends with a CodeBlock, which nests the actual outputs under the ‘result’ key. This method overrides the default output in the same manner that flows_sdk.blocks.Routing.CodeBlock does for convenience.

Parameters:

key (Optional[str]) –

Optionally provide a key to directly get a nested property. When not provided, the entire ‘result’ will be returned. For example, we have:

{
    "result": {
        "a": {
            "b": 42
        }
    }
}
  • skipping the key output()` will return ``{"a": {"b": 42}}

  • calling with output("a") will result in {"b": 42}

  • calling with output("a.b") will return 42

Return type:

str

Returns:

the value under the provided key

class flows_sdk.implementations.idp_v39.idp_blocks.MachineCollationBlock(submission, cases, dedupe_files='${workflow.input.machine_collation_dedupe_files}', refresh_retention_period='${workflow.input.machine_collation_refresh_retention_period}', remove_from_cases='${workflow.input.machine_collation_remove_from_cases}', retention_period='${workflow.input.machine_collation_retention_period}', reference_name=None)

Bases: Block

Machine collation block groups files, documents and pages (from the submission) into cases.

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • cases (Any) –

    This parameter accepts an array of JSON objects that contain information on how Cases should be created. There are two ways to collate a case: by the filenames specified in the current Submission, or by the ids of the Documents or Pages submitted in previous submissions.

    This parameter follows the given format:

    {
        'cases': [
            'external_case_id': 'HS-1',
            'filename': 'file1.pdf',
            'documents': [42],
            'pages': [43]
        ]
    }
    

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • dedupe_files (Union[bool, str]) – Enabling this setting will replace case data from repeated file names within the same case. Cases will retain data from the most recently submitted version of a particular file. Note that this setting does not delete the old data, it just removes it from the case. Keep in mind this option is only relevant when adding to cases. Defaults to workflow_input(Settings.CollationDedupeFiles)

  • remove_from_cases (Optional[Any]) –

    This parameter accepts an array of JSON objects that contains information on how documents or pages should be removed from a pre-existing case. Currently, the only way to de-collate from a case is by the ids of the Documents or Pages within a given case, defaults to workflow_input(Settings.CollationRemoveFromCases)

    This parameter follows the given format:

    {
        'remove_from_cases': [
            'external_case_id': 'HS-1',
            'documents': [42],
            'pages': [43]
        ]
    }
    

  • retention_period (Union[int, str, None]) – The number of days to retain cases after they are updated by this block. If passed None, no changes to the deletion date will be made. Defaults to workflow_input(Settings.CollationRetentionPeriod)

  • refresh_retention_period (Union[bool, str]) – If True, (re)applies Retention Period parameter to all cases passed into the block. If False, only updates those without a pre-existing deletion date. Defaults to workflow_input(Settings.CollationRefreshRetentionPeriod)

Example usage:

machine_collation = MachineCollationBlock(
    reference_name='machine_collation',
    submission=submission_bootstrap.output('result.submission'),
    cases=submission_bootstrap.output('result.api_params.cases'),
    dedupe_files=True,
    remove_from_cases=custom_supervision.output('remove_from_cases')
)
IDENTIFIER = 'MACHINE_COLLATION_2'
class flows_sdk.implementations.idp_v39.idp_blocks.MachineClassificationBlock(submission, api_params, reference_name=None, rotation_correction_enabled='${workflow.input.machine_classification_rotation_correction_enabled}', mobile_processing_enabled='${workflow.input.machine_classification_mobile_processing_enabled}', file_transcription_enabled='${workflow.input.machine_classification_file_transcription_enabled}', layout_release_uuid='${workflow.input.layout_release_uuid}', vpc_registration_threshold='${workflow.input.structured_layout_match_threshold}', nlc_enabled='${workflow.input.semi_structured_classification}', nlc_target_accuracy='${workflow.input.semi_target_accuracy}', nlc_doc_grouping_logic='${workflow.input.semi_doc_grouping_logic}')

Bases: Block

Machine classification block automatically matches documents to structured, semi-structured or additional layouts

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • api_params (Any) –

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • rotation_correction_enabled (Union[bool, str]) – Identifies and corrects the orientation of semi-structured images, defaults to workflow_input(Settings.RotationCorrectionEnabled)

  • mobile_processing_enabled (Union[bool, str]) – Improves the machine readability of photo captured documents, defaults to workflow_input(Settings.MobileProcessingEnabled)

  • file_transcription_enabled (Union[bool, str]) – Improves the transcription speed of pdf files. Make sure that the “Image Correction” option is disabled and the pdfs you submit do not contain any rotated images defaults to workflow_input(Settings.FileTranscriptionEnabled)

  • layout_release_uuid (str) – Layout Release UUID at submission creation time, defaults to workflow_input(Settings.LayoutReleaseUuid)

  • vpc_registration_threshold (str) – Structured pages above this threshold will be automatically matched to a layout variation, defaults to workflow_input(Settings.StructuredLayoutMatchThreshold)

  • nlc_enabled (str) – Enables workflow to manage a model for automated classification of semi-structured and additional layout variations, defaults to workflow_input(Settings.SemiStructuredClassification)

  • nlc_target_accuracy (str) – defaults to workflow_input(Settings.SemiTargetAccuracy)

  • nlc_doc_grouping_logic (str) – Logic to handle multiple pages matched to the same layout variationin a given submission, defaults to workflow_input(Settings.SemiDocGroupingLogic)

Example usage:

machine_classification = MachineClassificationBlock(
    reference_name='machine_classification',
    submission=case_collation.output('submission'),
    api_params=submission_bootstrap.output('result.api_params'),
    rotation_correction_enabled=idp_wf_config.rotation_correction_enabled,
)
IDENTIFIER = 'MACHINE_CLASSIFICATION_6'
class flows_sdk.implementations.idp_v39.idp_blocks.ManualClassificationBlock(submission, api_params, reference_name=None, layout_release_uuid='${workflow.input.layout_release_uuid}', manual_nlc_enabled='${workflow.input.manual_nlc_enabled}', task_restrictions='${workflow.input.manual_classification_task_restrictions}', notification_workflow='${workflow.input.notification_workflow_classification}', nlc_enabled='${workflow.input.semi_structured_classification}', nlc_target_accuracy='${workflow.input.semi_target_accuracy}', mobile_processing_enabled='${workflow.input.machine_classification_mobile_processing_enabled}', vpc_registration_threshold='${workflow.input.structured_layout_match_threshold}')

Bases: Block

Manual classification block allows keyers to manually match submissions to their layouts. Keyers may perform manual classification if machine classification cannot automatically match a submission to a layout with high confidence

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • api_params (Any) –

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • layout_release_uuid (str) – Layout Release UUID at submission creation time, defaults to workflow_input(Settings.LayoutReleaseUuid)

  • manual_nlc_enabled (str) – Enables manual classification when applicable, defaults to workflow_input(Settings.ManualNlcEnabled)

  • task_restrictions (Union[List[Any], str, None]) – Defines what users can access Supervision tasks created by this block, defaults to workflow_input(Settings.ManualClassificationTaskRestrictions)

  • notification_workflow (str) – Notification flow name to run when the submission enters Manual Classification, defaults to workflow_input(Settings.ManualClassificationNotificationWorkflow)

  • nlc_enabled (Union[bool, str, None]) – Enables workflow to manage a model for automated classification of semi-structured and additional layout variations, defaults to workflow_input(Settings.SemiStructuredClassification)

  • nlc_target_accuracy (Union[float, str, None]) – Classification Target Accuracy, defaults to workflow_input(Settings.SemiTargetAccuracy)

  • mobile_processing_enabled (Union[bool, str, None]) – Improves the machine readability of photo captured document, defaults to workflow_input(Settings.MobileProcessingEnabled)

  • vpc_registration_threshold (Optional[str]) – Structured pages above this threshold will be automatically matched to a layout variation when performing light vpc registration in the doc-org task, defaults to workflow_input(Settings.StructuredLayoutMatchThreshold)

Example usage:

manual_classification = ManualClassificationBlock(
    reference_name='manual_classification',
    submission=machine_classification.output('submission'),
    api_params=submission_bootstrap.output('result.api_params'),
)
IDENTIFIER = 'MANUAL_CLASSIFICATION_3'
class flows_sdk.implementations.idp_v39.idp_blocks.MachineIdentificationBlock(submission, api_params, layout_release_uuid='${workflow.input.layout_release_uuid}', reference_name=None, manual_field_id_enabled='${workflow.input.manual_field_id_enabled}', field_id_target_accuracy='${workflow.input.field_id_target_accuracy}', table_id_target_accuracy='${workflow.input.table_id_target_accuracy}')

Bases: Block

Machine identification automatically identify fields and tables in the submission

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • api_params (Any) –

Parameters with defaults:

Parameters:
  • layout_release_uuid (Any) – Layout Release UUID at submission creation time, defaults to workflow_input(Settings.LayoutReleaseUuid)

  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • manual_field_id_enabled (str) – Enables manual identification when applicable, defaults to workflow_input(Settings.ManualFieldIdEnabled)

  • field_id_target_accuracy (str) – Field ID Target Accuracy, defaults to workflow_input(Settings.FieldIdTargetAccuracy)

  • table_id_target_accuracy (str) – Table ID Target Accuracy, defaults to workflow_input(Settings.TableIdTargetAccuracy)

Example usage:

machine_identification = MachineIdentificationBlock(
    reference_name='machine_identification',
    submission=manual_classification.output('submission'),
    api_params=submission_bootstrap.output('result.api_params'),
)
IDENTIFIER = 'MACHINE_IDENTIFICATION_7'
class flows_sdk.implementations.idp_v39.idp_blocks.ManualIdentificationBlock(submission, api_params, reference_name=None, task_restrictions='${workflow.input.manual_identification_task_restrictions}', manual_field_id_enabled='${workflow.input.manual_field_id_enabled}', layout_release_uuid='${workflow.input.layout_release_uuid}', notification_workflow='${workflow.input.notification_workflow_identification}')

Bases: Block

Manual identification allows keyers to complete field identification or table identification tasks, where they draw bounding boxes around the contents of certain fields, table columns or table rows. This identification process ensures that the system will be able to transcribe the correct content in the upcoming transcription process

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • api_params (Any) –

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • task_restrictions (Union[List[Any], str, None]) – Defines what users can access Supervision tasks created by this block, defaults to workflow_input(Settings.ManualFieldIdTaskRestrictions)

  • manual_field_id_enabled (str) – Enables manual identification when applicable, defaults to workflow_input(Settings.ManualFieldIdEnabled)

  • layout_release_uuid (str) – Layout Release UUID at submission creation time, defaults to workflow_input(Settings.LayoutReleaseUuid)

  • notification_workflow (str) – Notification flow name to run when the submission enters Manual Identification, defaults to workflow_inputSettings.ManualFieldIdNotificationWorkflow

Example usage:

manual_identification = ManualIdentificationBlock(
    reference_name='manual_identification',
    submission=machine_identification.output('submission'),
    api_params=submission_bootstrap.output('result.api_params'),
    task_restrictions=idp_wf_config.manual_identification_config.task_restrictions,
)
IDENTIFIER = 'MANUAL_IDENTIFICATION_5'
class flows_sdk.implementations.idp_v39.idp_blocks.MachineTranscriptionBlock(submission, api_params, reference_name=None, layout_release_uuid='${workflow.input.layout_release_uuid}', transcription_model='${workflow.input.transcription_model}', finetuning_only_trained_layouts='${workflow.input.finetuning_only_trained_layouts}', transcription_overrides='${workflow.input.transcription_overrides}', structured_text_target_accuracy='${workflow.input.structured_text_target_accuracy}', structured_text_confidence_threshold='${workflow.input.structured_text_threshold}', structured_text_acceptable_confidence='${workflow.input.structured_min_leg_threshold}', semi_structured_text_target_accuracy='${workflow.input.semi_structured_text_target_accuracy}', semi_structured_text_confidence_threshold='${workflow.input.semi_structured_text_threshold}', semi_structured_text_acceptable_confidence='${workflow.input.semi_structured_min_leg_threshold}', semi_structured_table_target_accuracy='${workflow.input.semi_structured_table_target_accuracy}', semi_structured_table_confidence_threshold='${workflow.input.semi_structured_table_threshold}', semi_structured_table_acceptable_confidence='${workflow.input.semi_structured_table_min_leg_threshold}', structured_checkbox_target_accuracy='${workflow.input.structured_checkbox_target_accuracy}', structured_checkbox_confidence_threshold='${workflow.input.structured_checkbox_threshold}', structured_checkbox_acceptable_confidence='${workflow.input.checkbox_min_leg_threshold}', structured_signature_target_accuracy='${workflow.input.structured_signature_target_accuracy}', structured_signature_confidence_threshold='${workflow.input.structured_signature_threshold}', structured_signature_acceptable_confidence='${workflow.input.signature_min_leg_threshold}', semi_structured_checkbox_target_accuracy='${workflow.input.semi_structured_checkbox_target_accuracy}', semi_structured_checkbox_confidence_threshold='${workflow.input.semi_structured_checkbox_threshold}', semi_structured_checkbox_acceptable_confidence='${workflow.input.semi_structured_checkbox_min_leg_threshold}')

Bases: Block

Machine transcription automatically transcribes the content of your submission

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • api_params (Any) –

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • layout_release_uuid (str) – Layout Release UUID at submission creation time, defaults to workflow_input(Settings.LayoutReleaseUuid)

  • transcription_model (str) – The transcription model that will be used for this flow, defaults to workflow_input(Settings.TranscriptionModel)

  • finetuning_only_trained_layouts (str) – defaults to workflow_input( Settings.FinetuningOnlyTrainedLayouts )

  • transcription_overrides (str) – defaults to workflow_input( Settings.TranscriptionOverrides )

  • structured_text_target_accuracy (str) – defaults to workflow_input( Settings.StructuredTextTargetAccuracy )

  • structured_text_confidence_threshold (str) – defaults to workflow_input( Settings.StructuredTextThreshold )

  • structured_text_acceptable_confidence (str) – defaults to workflow_input( Settings.StructuredTextMinLegThreshold )

  • semi_structured_text_target_accuracy (str) – defaults to workflow_input( Settings.SemiStructuredTextTargetAccuracy )

  • semi_structured_text_confidence_threshold (str) – defaults to workflow_input( Settings.SemiStructuredTextThreshold )

  • semi_structured_text_acceptable_confidence (str) – defaults to workflow_input( Settings.SemiStructuredTextMinLegThreshold )

  • structured_checkbox_target_accuracy (str) – defaults to workflow_input( Settings.StructuredCheckboxTargetAccuracy )

  • structured_checkbox_confidence_threshold (str) – defaults to workflow_input( Settings.StructuredCheckboxThreshold )

  • structured_checkbox_acceptable_confidence (str) – defaults to workflow_input( Settings.StructuredCheckboxMinLegThreshold )

  • structured_signature_target_accuracy (str) – defaults to workflow_input( Settings.StructuredSignatureTargetAccuracy )

  • structured_signature_confidence_threshold (str) – defaults to workflow_input( Settings.StructuredSignatureThreshold )

  • structured_signature_acceptable_confidence (str) – defaults to workflow_input( Settings.StructuredSignatureMinLegThreshold )

  • semi_structured_checkbox_target_accuracy (str) – defaults to workflow_input( Settings.SemiStructuredCheckboxTargetAccuracy )

  • semi_structured_checkbox_confidence_threshold (str) – defaults to workflow_input( Settings.SemiStructuredCheckboxThreshold )

  • semi_structured_checkbox_acceptable_confidence (str) – defaults to workflow_input( Settings.SemiStructuredCheckboxMinLegThreshold )

Example usage:

machine_transcription = MachineTranscriptionBlock(
    reference_name='machine_transcription',
    submission=manual_identification.output('submission'),
    api_params=submission_bootstrap.output('result.api_params'),
)
IDENTIFIER = 'MACHINE_TRANSCRIPTION_9'
class flows_sdk.implementations.idp_v39.idp_blocks.ManualTranscriptionBlock(submission, api_params, reference_name=None, supervision_transcription_masking='${workflow.input.manual_transcription_supervision_transcription_masking}', table_output_manual_review='${workflow.input.manual_transcription_table_output_manual_review}', always_supervise_blank_cells='${workflow.input.manual_transcription_always_supervise_blank_cells}', task_restrictions='${workflow.input.manual_transcription_task_restrictions}', manual_transcription_enabled='${workflow.input.manual_transcription_enabled}', notification_workflow='${workflow.input.notification_workflow_transcription}', title='Manual Transcription', description='Manually transcribe fields that could not be automatically transcribed')

Bases: Block

Manual transcription lets your keyers manually enter the text found in fields or tables that could not be automatically transcribed

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • api_params (Any) –

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • supervision_transcription_masking (Union[bool, str]) – Prevents users from inputting invalid characters during Supervision Transcription tasks, defaults to workflow_input(Settings.ManualTranscriptionSupervisionTranscriptionMasking)

  • table_output_manual_review (Union[bool, str]) – Always generates a table transcription task if the layout contains a table. If disabled, a table transcription task will only be generated if one or more cells have transcribed values below the defined thresholds, defaults to workflow_input(Settings.ManualTranscriptionTableOutputManualReview)

  • always_supervise_blank_cells (Union[bool, str]) – If enabled, always send blank cells in transcribed tables to be supervised manually, regardless of confidence defaults to workflow_input(Settings.ManualTranscriptionAlwaysSuperviseBlankCells)

  • task_restrictions (Union[List[Any], str, None]) – Defines what users can access Supervision tasks for submission from a particular source or submissions matching a specific layout, defaults to workflow_input(Settings.ManualTranscriptionTaskRestrictions)

  • manual_transcription_enabled (str) – Enables manual transcription when applicable, defaults to workflow_input(Settings.ManualTranscriptionEnabled)

  • notification_workflow (str) – Notification flow name to run when the submission enters Manual Transcription, defaults to workflow_input(Settings.ManualTranscriptionNotificationWorkflow)

Example usage:

manual_transcription = ManualTranscriptionBlock(
    reference_name='manual_transcription',
    submission=machine_transcription.output('submission'),
    api_params=submission_bootstrap.output('result.api_params'),
    supervision_transcription_masking=(
        idp_wf_config.manual_transcription_config.supervision_transcription_masking
    ),
    table_output_manual_review=(
        idp_wf_config.manual_transcription_config.table_output_manual_review
    ),
    task_restrictions=idp_wf_config.manual_transcription_config.task_restrictions,
)
IDENTIFIER = 'MANUAL_TRANSCRIPTION_5'
class flows_sdk.implementations.idp_v39.idp_blocks.FlexibleExtractionBlock(submission, api_params, reference_name=None, supervision_transcription_masking='${workflow.input.flexible_extraction_supervision_transcription_masking}', task_restrictions='${workflow.input.flexible_extraction_task_restrictions}', layout_release_uuid='${workflow.input.layout_release_uuid}', notification_workflow='${workflow.input.notification_workflow_flexible_extraction}')

Bases: Block

Flexible extraction manually transcribes fields marked for review

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • api_params (Any) –

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • supervision_transcription_masking (Union[bool, str, None]) – Prevents users from inputting invalid characters during the Flexible Extraction task, defaults to workflow_input(Settings.FlexibleExtractionSupervisionTranscriptionMasking)

  • task_restrictions (Union[List[Any], str, None]) – Defines what users can access Supervision tasks created by this block, defaults workflow_input(Settings.FlexibleExtractionTaskRestrictions)

  • layout_release_uuid (str) – Layout Release UUID at submission creation time, defaults to workflow_input(Settings.LayoutReleaseUuid)

  • notification_workflow (str) – Notification flow name to run when the submission enters Flexible Extraction, defaults to workflow_input(Settings.FlexibleExtractionNotificationWorkflow)

Example usage:

flexible_extraction = FlexibleExtractionBlock(
    reference_name='flexible_extraction',
    submission=manual_transcription.output('submission'),
    api_params=submission_bootstrap.output('result.api_params'),
    supervision_transcription_masking=(
        idp_wf_config.flexible_extraction_config.supervision_transcription_masking
    ),
    task_restrictions=idp_wf_config.flexible_extraction_config.task_restrictions,
)
IDENTIFIER = 'FLEXIBLE_EXTRACTION_4'
class flows_sdk.implementations.idp_v39.idp_blocks.ReprocessingBlock(reference_name, submission, parent_flow='${workflow.definition.identifier}', idp_wf_config={'auto_qa_sampling_rate_enabled': '${workflow.input.auto_qa_sampling_rate_enabled}', 'checkbox_min_leg_threshold': '${workflow.input.checkbox_min_leg_threshold}', 'exclude_auto_transcribe_fields_from_qa': '${workflow.input.exclude_auto_transcribe_fields_from_qa}', 'field_id_qa_enabled': '${workflow.input.field_id_qa_enabled}', 'field_id_qa_sample_rate': '${workflow.input.field_id_qa_sample_rate}', 'field_id_target_accuracy': '${workflow.input.field_id_target_accuracy}', 'finetuning_only_trained_layouts': '${workflow.input.finetuning_only_trained_layouts}', 'flexible_extraction_supervision_transcription_masking': '${workflow.input.flexible_extraction_supervision_transcription_masking}', 'flexible_extraction_task_restrictions': '${workflow.input.flexible_extraction_task_restrictions}', 'layout_release_uuid': '${workflow.input.layout_release_uuid}', 'machine_classification_file_transcription_enabled': '${workflow.input.machine_classification_file_transcription_enabled}', 'machine_classification_mobile_processing_enabled': '${workflow.input.machine_classification_mobile_processing_enabled}', 'machine_classification_rotation_correction_enabled': '${workflow.input.machine_classification_rotation_correction_enabled}', 'manual_field_id_enabled': '${workflow.input.manual_field_id_enabled}', 'manual_nlc_enabled': '${workflow.input.manual_nlc_enabled}', 'manual_transcription_always_supervise_blank_cells': '${workflow.input.manual_transcription_always_supervise_blank_cells}', 'manual_transcription_enabled': '${workflow.input.manual_transcription_enabled}', 'manual_transcription_supervision_transcription_masking': '${workflow.input.manual_transcription_supervision_transcription_masking}', 'manual_transcription_table_output_manual_review': '${workflow.input.manual_transcription_table_output_manual_review}', 'manual_transcription_task_restrictions': '${workflow.input.manual_transcription_task_restrictions}', 'notification_workflow_classification': '${workflow.input.notification_workflow_classification}', 'notification_workflow_flexible_extraction': '${workflow.input.notification_workflow_flexible_extraction}', 'notification_workflow_identification': '${workflow.input.notification_workflow_identification}', 'notification_workflow_initialization': '${workflow.input.notification_workflow_initialization}', 'notification_workflow_transcription': '${workflow.input.notification_workflow_transcription}', 'qa': '${workflow.input.qa}', 'semi_doc_grouping_logic': '${workflow.input.semi_doc_grouping_logic}', 'semi_qa_sample_rate': '${workflow.input.semi_qa_sample_rate}', 'semi_structured_checkbox_max_training_records': '${workflow.input.semi_structured_checkbox_max_training_records}', 'semi_structured_checkbox_min_leg_threshold': '${workflow.input.semi_structured_checkbox_min_leg_threshold}', 'semi_structured_checkbox_min_training_records': '${workflow.input.semi_structured_checkbox_min_training_records}', 'semi_structured_checkbox_target_accuracy': '${workflow.input.semi_structured_checkbox_target_accuracy}', 'semi_structured_checkbox_threshold': '${workflow.input.semi_structured_checkbox_threshold}', 'semi_structured_classification': '${workflow.input.semi_structured_classification}', 'semi_structured_improved_transcr_threshold_accuracy': '${workflow.input.semi_structured_improved_transcr_threshold_accuracy}', 'semi_structured_min_leg_threshold': '${workflow.input.semi_structured_min_leg_threshold}', 'semi_structured_qa_sample_rate': '${workflow.input.semi_structured_qa_sample_rate}', 'semi_structured_table_max_training_records': '${workflow.input.semi_structured_table_max_training_records}', 'semi_structured_table_min_leg_threshold': '${workflow.input.semi_structured_table_min_leg_threshold}', 'semi_structured_table_min_training_records': '${workflow.input.semi_structured_table_min_training_records}', 'semi_structured_table_target_accuracy': '${workflow.input.semi_structured_table_target_accuracy}', 'semi_structured_table_threshold': '${workflow.input.semi_structured_table_threshold}', 'semi_structured_text_max_training_records': '${workflow.input.semi_structured_text_max_training_records}', 'semi_structured_text_min_training_records': '${workflow.input.semi_structured_text_min_training_records}', 'semi_structured_text_target_accuracy': '${workflow.input.semi_structured_text_target_accuracy}', 'semi_structured_text_threshold': '${workflow.input.semi_structured_text_threshold}', 'semi_structured_transcription_training': '${workflow.input.semi_structured_transcription_training}', 'semi_structured_transcription_training_period': '${workflow.input.semi_structured_transcription_training_period}', 'semi_target_accuracy': '${workflow.input.semi_target_accuracy}', 'signature_min_leg_threshold': '${workflow.input.signature_min_leg_threshold}', 'structured_checkbox_max_training_records': '${workflow.input.structured_checkbox_max_training_records}', 'structured_checkbox_min_training_records': '${workflow.input.structured_checkbox_min_training_records}', 'structured_checkbox_qa_sample_rate': '${workflow.input.structured_checkbox_qa_sample_rate}', 'structured_checkbox_target_accuracy': '${workflow.input.structured_checkbox_target_accuracy}', 'structured_checkbox_threshold': '${workflow.input.structured_checkbox_threshold}', 'structured_entry_qa_sample_rate': '${workflow.input.structured_entry_qa_sample_rate}', 'structured_improved_transcr_threshold_accuracy': '${workflow.input.structured_improved_transcr_threshold_accuracy}', 'structured_layout_match_threshold': '${workflow.input.structured_layout_match_threshold}', 'structured_min_leg_threshold': '${workflow.input.structured_min_leg_threshold}', 'structured_signature_max_training_records': '${workflow.input.structured_signature_max_training_records}', 'structured_signature_min_training_records': '${workflow.input.structured_signature_min_training_records}', 'structured_signature_qa_sample_rate': '${workflow.input.structured_signature_qa_sample_rate}', 'structured_signature_target_accuracy': '${workflow.input.structured_signature_target_accuracy}', 'structured_signature_threshold': '${workflow.input.structured_signature_threshold}', 'structured_text_max_training_records': '${workflow.input.structured_text_max_training_records}', 'structured_text_min_training_records': '${workflow.input.structured_text_min_training_records}', 'structured_text_target_accuracy': '${workflow.input.structured_text_target_accuracy}', 'structured_text_threshold': '${workflow.input.structured_text_threshold}', 'structured_transcription_training': '${workflow.input.structured_transcription_training}', 'structured_transcription_training_period': '${workflow.input.structured_transcription_training_period}', 'submission_bootstrap_http_config': '${workflow.input.submission_bootstrap_http_config}', 'submission_bootstrap_ocs_config': '${workflow.input.submission_bootstrap_ocs_config}', 'submission_bootstrap_s3_config': '${workflow.input.submission_bootstrap_s3_config}', 'table_cell_transcription_qa_enabled': '${workflow.input.table_cell_transcription_qa_enabled}', 'table_cell_transcription_qa_sample_rate': '${workflow.input.table_cell_transcription_qa_sample_rate}', 'table_id_qa_enabled': '${workflow.input.table_id_qa_enabled}', 'table_id_qa_sample_rate': '${workflow.input.table_id_qa_sample_rate}', 'table_id_target_accuracy': '${workflow.input.table_id_target_accuracy}', 'transcription_model': '${workflow.input.transcription_model}', 'workflow_name': '${workflow.input.workflow_name}', 'workflow_uuid': '${workflow.input.workflow_uuid}', 'workflow_version': '${workflow.input.workflow_version}'}, trigger='${workflow.input.$trigger}')

Bases: Block

Reprocessing block enables a second extraction attempt if documents were initially confidently mis-classified by the machine. When an end-user specifies that a document layout is incorrect during Manual Identification, Manual Transcription, or Flexible Extraction, the flow will reach the Reprocessing block which will mark correctly classified documents as “complete” status, and misclassified documents with “reprocess” status will be processed in a second execution of the flow. This block should be placed before the SubmissionCompleteBlock to enable reprocessing.

Mandatory parameters:

Parameters:

submission (Any) – Submission object

IDENTIFIER = 'REPROCESSING'
class flows_sdk.implementations.idp_v39.idp_blocks.SubmissionCompleteBlock(submission, is_reprocessing=False, reference_name=None, nlc_qa_sampling_ratio='${workflow.input.semi_qa_sample_rate}', field_id_qa_enabled='${workflow.input.field_id_qa_enabled}', field_id_qa_sampling_ratio='${workflow.input.field_id_qa_sample_rate}', table_id_qa_enabled='${workflow.input.table_id_qa_enabled}', table_id_qa_sampling_ratio='${workflow.input.table_id_qa_sample_rate}', transcription_qa_enabled='${workflow.input.qa}', transcription_structured_entry_qa_sample_rate='${workflow.input.structured_entry_qa_sample_rate}', transcription_structured_signature_qa_sample_rate='${workflow.input.structured_signature_qa_sample_rate}', transcription_structured_checkbox_qa_sample_rate='${workflow.input.structured_checkbox_qa_sample_rate}', transcription_semi_structured_qa_sample_rate='${workflow.input.semi_structured_qa_sample_rate}', table_cell_transcription_qa_enabled='${workflow.input.table_cell_transcription_qa_enabled}', table_cell_transcription_qa_sample_rate='${workflow.input.table_cell_transcription_qa_sample_rate}', auto_qa_sampling_rate_enabled='${workflow.input.auto_qa_sampling_rate_enabled}', exclude_auto_transcribe_fields_from_qa='${workflow.input.exclude_auto_transcribe_fields_from_qa}', payload=None)

Bases: Block

Submission complete block finalizes submission processing and updates reporting data. This block should be called at the END of every flow that uses blocks included in the flows_sdk.implementations.idp_v34 library. The block requires both the submission object and a payload to be passed in.

Mandatory parameters:

Parameters:
  • submission (str) – Submission object

  • payload (Optional[Any]) – Object to pass to downstream systems

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • nlc_qa_sampling_ratio (Union[str, int]) – Defines the percentage of documents sampled for Classification QA, defaults to workflow_input(Settings.SemiQaSampleRate)

  • field_id_qa_enabled (Union[str, bool]) – defaults to workflow_input(Settings.FieldIdQaEnabled)

  • field_id_qa_sampling_ratio (Union[str, int]) – Defines the percentage of documents sampled for Identification QA, defaults to workflow_input(Settings.FieldIdQaSampleRate)

  • table_id_qa_enabled (Union[str, bool]) – Allows users to verify the location of table cells, defaults to workflow_input(Settings.TableIdQaEnabled)

  • table_id_qa_sampling_ratio (Union[str, int]) – Defines the percentage of tables the system samples for QA, defaults to workflow_input(Settings.TableIdQaSampleRate)

  • transcription_qa_enabled (Union[str, bool]) – defaults to workflow_input(Settings.TranscriptionQaEnabled)

  • transcription_qa_sampling_ratio – Defines the percentage of fields sampled for Transcription QA, defaults to workflow_input(Settings.TranscriptionQaSampleRate)

  • table_cell_transcription_qa_enabled (Union[str, bool]) – defaults to workflow_input(Settings.TableCellTranscriptionQaEnabled)

  • table_cell_transcription_qa_sample_rate (Union[str, int]) – Defines the percentage of cells the system samples for QA. This value is likely to be lower than “Transcription QA Sample Rate” since there are more table cells than fields on any given page, defaults to workflow_input(Settings.TableCellTranscriptionQaSampleRate)

  • exclude_auto_transcribe_fields_from_qa (Union[str, bool]) – Defines fields that are marked as auto-transcribe in layout should be excluded from QA or not. defaults to workflow_input(Settings.ExcludeAutoTranscribeFieldsFromQA)

  • is_reprocessing (Union[str, bool]) – boolean to handle execution when reprocessing defaults to False

Example usage:

submission_complete = SubmissionCompleteBlock(
    reference_name='complete_submission',
    submission=flexible_extraction.output('submission')
)
IDENTIFIER = 'SUBMISSION_COMPLETE_4'
class flows_sdk.implementations.idp_v39.idp_blocks.SetTransformedOutputBlock(submission, transformed_output, reference_name=None)

Bases: Block

Set Transformed Output block sets the transformed output for the provided submission. This can later be used to send the transformed data downstream using output blocks. For more information see https://docs.hyperscience.com/#retrieving-transformed-submissions.

Mandatory parameters:

Parameters:
  • submission (str) – Submission object

  • transformed_output (str) – Object that will be set as the transformed output for this submission

IDENTIFIER = 'SET_TRANSFORMED_OUTPUT'
class flows_sdk.implementations.idp_v39.idp_blocks.IdpCustomSupervisionBlock(submission, task_purpose, supervision_template, reference_name=None, supervision_transcription_masking=True, task_restrictions=None, page_ids=None, notification_workflow='IDP_SUBMISSION_NOTIFY_V39', title='IDP Custom Supervision', description='IDP Custom Supervision')

Bases: Block

An IDP wrapper for the Custom Supervision block. It has the same functionality, but handles reading/writing data from the IDP database.

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • task_purpose (str) – the custom task name that’s given to all supervision tasks that run through a particular instance of a custom supervision block. This can be used to filter tasks in the task queue.

  • supervision_template (Any) – JSON used for configuration of the Custom Supervision Task.

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • supervision_transcription_masking (Union[str, bool]) – Prevents users from inputting invalid characters during the Custom Supervision task, defaults to True

  • task_restrictions (Union[str, List[Any], None]) – Defines what users can access Supervision tasks created by this block, defaults to []

  • page_ids (Union[str, List[int], None]) – This is an optional parameter. If left empty, the IDP wrapper will, by default, pull in all pages in the submission. Adding a list of page IDs will override the default list of pages. defaults to []

  • notification_workflow (str) – Notification flow name to run when the submission enters Custom Supervision (which is called within the IDP Custom Supervision block), defaults to IDP_SUBMISSION_NOTIFY_NAME

  • title (Optional[str]) – UI-visible title of the block. Defaults to ‘IDP Custom Supervision’

  • description (Optional[str]) – Description of the block. Both useful for documentation purpose and visible to users in the Flow studio. Defaults to ‘IDP Custom Supervision’

Example usage:

idp_custom_supervision = IdpCustomSupervisionBlock(
    reference_name='idp_custom_supervision',
    submission=manual_transcription.output('submission'),
    task_purpose='make_custom_decision',
    supervision_template=[
        {
            'name': 'three_column_template',
            'version': '1.0',
            'thumbnail': {
                'group_by_document': True,
                'group_by_case': True,
            },
            'action': [
                {
                    'name': 'my_tab',
                    'display': 'SSN Info',
                    'input': [
                        {
                            'name': 'ssn_field',
                            'type': 'transcription',
                            'data_identifier': '2afd7a67-ddee-484f-ab6b-2d292752e5ee',
                            'title': 'SSN',
                        },
                    ],
                    'ui': {
                        'groups': [
                            {
                                'title': 'My Group',
                                'fields': ['ssn_field'],
                            },
                        ],
                    },
                },
            ],
        },
    ],
    supervision_transcription_masking=False,
)
IDENTIFIER = 'IDP_CUSTOM_SUPERVISION_2'
class flows_sdk.implementations.idp_v39.idp_blocks.IDPFullPageTranscriptionBlock(submission, reference_name=None, title='Full Page Transcription (Submission)', description='Transcribes the documents included in a submission', app_metadata=None)

Bases: Block

IDP Full Page Transcription Block machine-transcribes the documents contained in a submission

Mandatory parameters:

Parameters:

submission (str) – submission object that includes the raw documents that need to be transcribed

Optional Parameters:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • app_metadata (Optional[List[str]]) – used for capturing information about the flow

Example usage:

# This is one way to obtain the submission object (using SubmissionBootstrapBlock)
submission_bootstrap = SubmissionBootstrapBlock(reference_name='submission_bootstrap')
# IDPFullPageTranscriptionBlock transcribes the documents contained in the submission object
idp_fpt_block = IDPFullPageTranscriptionBlock(
    reference_name='idp_fpt',
    submission=submission_bootstrap.output(),
)
# reference to the output of IDPFullPageTranscriptionBlock which can be passed as input to
# the next block. For more information on output() and output references, please take a
# look at class Block and its output() method
output_ref = idp_fpt_block.output()
IDENTIFIER = 'FULL_PAGE_TRANSCRIPTION_SUBMISSION_2'
class flows_sdk.implementations.idp_v39.idp_blocks.IDPImageCorrectionBlock(submission, reference_name=None, rotation_correction_enabled=True, mobile_processing_enabled=False, title='Image Correction (Submission)', description='Rotate and de-skew documents included in a submission', app_metadata=None)

Bases: Block

IDP Image Correction Block rotates and de-skews tilted/skewed images so that they are more transcription-friendly

Mandatory parameters:

Parameters:

submission (str) – submission object that includes the raw documents that need to be image-corrected

Optional Parameters:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • app_metadata (Optional[List[str]]) – used for capturing information about the flow

Example usage:

# This is one way to obtain the submission object (using SubmissionBootstrapBlock)
submission_bootstrap = SubmissionBootstrapBlock(reference_name='submission_bootstrap')
# IDPImageCorrectionBlock takes in the submission object and outputs in the same format with
# its images
idp_image_correct_block = IDPImageCorrectionBlock(
    reference_name='idp_image_correct',
    submission=submission_bootstrap.output(),
)
# reference to the output of IDPImageCorrectionBlock which can be passed as input to the
# next block. For more information on output() and output references, please take a look at
# class Block and its output() method
output_ref = idp_image_correct_block.output()
IDENTIFIER = 'IMAGE_CORRECTION_SUBMISSION'

Additional Blocks (v39)

class flows_sdk.implementations.idp_v39.additional_blocks.JSONOutputsBlock(inputs, blocks=None)

Bases: Outputs

Output block allows users to send JSON data extracted by an IDP flow to other systems for downstream processing Mandatory parameters:

Parameters:

inputs (Dict[str, Any]) – Used by the UI to automatically pre-populate the input of newly added blocks. Extended in this class to include 'enabled': True to visualize an enabled/disabled trigger that is controllable via Flow Studio.

Usage when parameters are to be defined in Flow Studio:

outputs = JSONOutputBlock(
    inputs={'result': }
)
class flows_sdk.implementations.idp_v39.additional_blocks.ImageCorrectionBlock(pages, reference_name=None, rotation_correction_enabled=True, mobile_processing_enabled=False, title='Image Correction (Pages)', description='Rotate and de-skew documents', app_metadata=None)

Bases: Block

Image Correction Block rotates and de-skews tilted/skewed images so that they are more transcription-friendly Mandatory parameters:

Parameters:

pages (str) – list of page object that includes the raw documents that need to be image-corrected

Optional Parameters:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • app_metadata (Optional[List[str]]) – used for capturing information about the flow

Example usage:

# This is one way to obtain the submission object (using SubmissionBootstrapBlock)
submission_bootstrap = SubmissionBootstrapBlock(reference_name='submission_bootstrap')
# ImageCorrectionBlock takes in the submission object and outputs in the same format with
# its images
image_correct_block = ImageCorrectionBlock(
    reference_name='image_correct',
    pages=submission_bootstrap.output(submission.unassigned_pages),
)
# reference to the output of IDPImageCorrectionBlock which can be passed as input to the
# next block. For more information on output() and output references, please take a look
# at class Block and its output() method
output_ref = image_correct_block.output()
IDENTIFIER = 'IMAGE_CORRECTION_PAGES'
class flows_sdk.implementations.idp_v39.additional_blocks.FullPageTranscriptionBlock(pages, reference_name=None, title='Full Page Transcription (Pages)', description='Transcribes documents', app_metadata=None)

Bases: Block

Full Page Transcription Block machine-transcribes the documents contained in a submission

Mandatory parameters:

Parameters:

pages (str) – list of pages that includes the raw documents that need to be transcribed

Optional Parameters:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • app_metadata (Optional[List[str]]) – used for capturing information about the flow

Example usage:

# This is one way to obtain the submission object (using SubmissionBootstrapBlock)
submission_bootstrap = SubmissionBootstrapBlock(reference_name='submission_bootstrap')
# FullPageTranscriptionBlock transcribes the documents contained in the submission object
fpt_block = FullPageTranscriptionBlock(
    reference_name='fpt',
    submission=submission_bootstrap.output(submission.unassigned_pages),
)
# reference to the output of FullPageTranscriptionBlock which can be passed as input to the
# next block. For more information on output() and output references, please take a look at
# class Block and its output() method
output_ref = fpt_block.output()
IDENTIFIER = 'FULL_PAGE_TRANSCRIPTION_PAGES_2'
class flows_sdk.implementations.idp_v39.additional_blocks.CustomSupervisionBlock(submission, task_purpose, data, supervision_template, reference_name=None, supervision_transcription_masking=True, task_restrictions=None, notification_workflow='IDP_SUBMISSION_NOTIFY_V39', title='Custom Supervision', description='Custom Supervision')

Bases: Block

The CustomSupervisionBlock uses a JSON schema as a form builder to create customer-driven UIs for Supervision tasks.

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • task_purpose (str) – the custom task name that’s given to all supervision tasks that run through a particular instance of a custom supervision block. This can be used to filter tasks in the task queue.

  • data (Any) – JSON structure that supplies the data used to populated the supervision_template

  • supervision_template (Any) – JSON used for configuration of the Custom Supervision Task.

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • supervision_transcription_masking (bool) – Prevents users from inputting invalid characters during the Custom Supervision task, defaults to True

  • task_restrictions (Optional[List[Any]]) – Defines what users can access Supervision tasks created by this block, defaults to []

  • notification_workflow (str) – Notification flow name to run when the submission enters Custom Supervision, defaults to IDP_SUBMISSION_NOTIFY_NAME

  • title (Optional[str]) – UI-visible title of the block. Defaults to ‘IDP Custom Supervision’

  • description (Optional[str]) – Description of the block. Both useful for documentation purpose and visible to users in the Flow studio. Defaults to ‘IDP Custom Supervision’

Example usage:

custom_supervision = CustomSupervisionBlock(
    reference_name='custom_supervision',
    submission=manual_transcription.output('submission'),
    task_purpose='make_custom_decision',
    supervision_template=[
        {
            'name': 'three_column_template',
            'version': '1.0',
            'thumbnail': {
                'group_by_document': True,
                'group_by_case': True,
            },
            'action': [
                {
                    'name': 'my_tab',
                    'display': 'SSN Info',
                    'input': [
                        {
                            'name': 'ssn_field',
                            'type': 'transcription',
                            'data_identifier': '2afd7a67-ddee-484f-ab6b-2d292752e5ee',
                            'title': 'SSN',
                        },
                    ],
                    'ui': {
                        'groups': [
                            {
                                'title': 'My Group',
                                'fields': ['ssn_field'],
                            },
                        ],
                    },
                },
            ],
        },
    ],
    data={
        'fields': [
            {
                'id': 967,
                'uuid': 'fe8d4e7a-a82c-448c-85a1-0a3cbbe53b61',
                'value': '111-22-3333',
                'page_id': 15,
                'layout_field_id': '2afd7a67-ddee-484f-ab6b-2d292752e5ee',
                'validation_overridden': False,
                'bounding_box': [
                    0.09360783305186686,
                    0.25792617589433436,
                    0.6358913805295097,
                    0.28862414740388187,
                ],
                'occurrence_index': 1,
            },
        ],
        'template_fields': [
            {
                'uuid': '2afd7a67-ddee-484f-ab6b-2d292752e5ee',
                'notes': '',
                'type': 'entry',
                'data_type_uuid': 'b600a4e4-758f-4c4b-a2f4-cbb5dd870a0c',
                'name': 'SSN',
            },
        ],
        'pages': [
            {
                'id': 15,
                'file_page_number': 1,
                'document_page_number': 1,
                'image_url': '/image/2ed3d989-b64a-44bd-a9fb-6e0dcc5ccdb8',
                'form_id': 2,
                'submission_id': 1,
                'external_case_ids': [],
                'read_only': False,
                'filename': '1 page.pdf',
                'form_page_id': 1
            },
        ],
        'documents': [
            {
                'id': 2,
                'layout_uuid': 'e31c3cbf-dcea-44fd-a052-902463c6040f',
                'layout_name': 'layout_name',
                'page_count': 2,
                'read_only': False,
            },
        ],
        'cases': [],
        'selected_choices': [],
    },
    supervision_transcription_masking=False,
)
IDENTIFIER = 'CUSTOM_SUPERVISION_2'
class flows_sdk.implementations.idp_v39.additional_blocks.SoapRequestBlock(method, reference_name=None, endpoint='', body_namespace=None, headers_namespace=None, headers=None, params=None, use_wsdl=False, wsdl_uri='', wsdl_service_name='', wsdl_port_name='', title='HTTP REST Block', description='HTTP REST block')

Bases: Block

Soap Block allows user to perform SOAP requests Mandatory parameters:

Parameters:
  • method (str) – SOAP method to invoke

  • endpoint (str) – absolute url including schema to HTTP endpoint for this request

Optional Parameters:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • body_namespace (Optional[str]) – SOAP request body namespace url

  • headers_namespace (Optional[str]) – SOAP request headers namespace url

  • headers (Optional[Dict[Any, Any]]) – SOAP headers to use

  • params (Optional[Dict[Any, Any]]) – key-value pair used as query parameters in the request to be inserted as SOAP body

Example usage:

# minimum viable instantiation, the rest of the fields can later be added in the UI
soap_block = SoapRequestBlock(
    reference_name='soap_request',
    title='Soap Request',
    description='Perform soap request',
    method='ExampleMethod',
)
soap_block = SoapRequestBlock(
    reference_name='soap_request',
    title='Soap Request',
    description='Perform soap request',
    method='ExampleMethod',
    endpoint='http://example.org/abc.wso',
    body_namespace='http://www.example.org/abc.countryinfo',
    params={'example': 'stuff'},
    headers={}
)
# reference to the output of SoapRequestBlock which can be passed as input to the next block
# for more information on output() and output references, please take a look at class Block
# and its output() method
output_ref = soap_block.output()
IDENTIFIER = 'SOAP_REQ'
class flows_sdk.implementations.idp_v39.additional_blocks.HttpRestBlock(method, endpoint, reference_name=None, headers=None, params=None, payload=None, json=None, authorization_type=None, handled_error_codes=None, auth_params=None, title='HTTP REST Block', description='HTTP REST block')

Bases: BaseHttpRestBlock

Http Rest Block allows user to perform HTTP requests Mandatory parameters:

Parameters:
  • method (str) – HTTP method to use

  • endpoint (str) – absolute url including schema to HTTP endpoint for this request

Optional Parameters:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • headers (Optional[Dict[str, str]]) – HTTP headers to use

  • params (Optional[Dict[str, Any]]) – key-value pair used as query parameters in the request

  • payload (Optional[Dict[str, str]]) – key-value pair to be used as x-www-form-urlencoded data. Mutually exclusive with json

  • json (Optional[Dict[str, Any]]) – key-value pair to be used as json data. Mutually exclusive with payload

  • authorization_type (Optional[str]) – authorization type (none, http_header, oauth_2_client_credentials)

  • handled_error_codes (Optional[List[Union[int, str]]]) – specify HTTP error codes that will not fail the task, allowing for wildcards with the ‘x’ char. Example: [400, “5xx”]

  • auth_params (Union[HttpHeaderAuthorizationParams, OAuth2AuthorizationParams, None]) – authorization parameters to be passed when the authorization_type is set to “http_header” or “oauth_2_client_credentials”

Example usage:

# minimum viable instantiation, the rest of the fields can later be added in the UI
http_rest_block = HttpRestBlock(
    reference_name='http_request',
    title='Http Request',
    description='Perform http request',
    method='GET',
)
http_rest_block = HttpRestBlock(
    reference_name='http_request',
    title='Http Request',
    description='Perform http request',
    method='GET',
    endpoint='https://sheets.googleapis.com/v4/spreadsheets/example',
)
# reference to the output of HttpRestBlock which can be passed as input to the next block
# for more information on output() and output references, please take a look at class Block
# and its output() method
output_ref = http_rest_block.output()
class flows_sdk.implementations.idp_v39.additional_blocks.HyperscienceRestApiBlock(method, app_endpoint, reference_name=None, headers=None, params=None, payload=None, json=None, handled_error_codes=None, title='Hyperscience HTTP REST Block', description='Hyperscience HTTP REST block')

Bases: BaseHttpRestBlock

Make Http requests to the Hyperscience platform APIs. This block will automatically take care of prepending the base url and will include the appropriate authentication headers with the request. Mandatory parameters:

Parameters:
  • method (str) – HTTP method to use

  • app_endpoint (str) – relative url, beginning with a forward slash, to an HTTP endpoint served by the Hypersciecne platform. For reference on the available endpoints, check out https://docs.hyperscience.com/

Optional Parameters:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • headers (Optional[Dict[str, str]]) – HTTP headers to use

  • params (Optional[Dict[str, Any]]) – key-value pair used as query parameters in the request

  • payload (Optional[Dict[str, str]]) – key-value pair to be used as x-www-form-urlencoded data. Mutually exclusive with json

  • json (Optional[Dict[str, Any]]) – key-value pair to be used as json data. Mutually exclusive with payload

  • handled_error_codes (Optional[List[Union[int, str]]]) – specify HTTP error codes that will not fail the task, allowing for wildcards with the ‘x’ char. Example: [400, “5xx”]

Example usage:

hs_rest_block = HyperscienceRestApiBlock(
    reference_name='http_request',
    app_endpoint='/api/healthcheck',
    method='GET',
)
class flows_sdk.implementations.idp_v39.additional_blocks.DatabaseAccessBlock(reference_name=None, db_type='', database='', host='', username='', password='', query='', port=None, options=None, timeout=None, query_params=None, title='DB Block', description='DB access block')

Bases: Block

Database Block allows user to access and perform queries on the specified database Mandatory parameters:

Parameters:
  • db_type (str) – database type (mssql, oracle, postgres)

  • host (str) – URL/IP address of the server database is hosted on

  • database (str) – database name

  • username (str) – database username

  • password (str) – database password

  • query (str) – parameterized query

Optional parameters:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • port (Optional[int]) – database port number

  • options (Optional[Dict[Any, Any]]) – dictionary of additional connection string options

  • query_params (Optional[Dict[Any, Any]]) – dictionary of values for query placeholders

  • timeout (Optional[int]) – timeout in seconds (mssql and postgres only)

Example usage:

# minimum viable instantiation, the rest of the fields can later be added in the UI
db_lookup = DatabaseAccessBlock(
    reference_name='db_lookup',
    title='Database Lookup',
    description='Perform database lookup',
)
# note that storing password in plain text form is not recommended
# password can later be entered in the UI or defined as a flow secret.
db_lookup = DatabaseAccessBlock(
    reference_name='db_lookup',
    title='Database Lookup',
    description='Perform database lookup',
    db_type='mssql',
    database='cars',
    host='example.com',
    username='user',
    # we recommend skipping the password field or defining it as a system secret
    # e.g. system_secret('{your_secret_identifier}') (both allow for secret input via. UI)
    password='pw'
    port=1433,
    timeout=200,
    query='SELECT * from CAR',
)
# reference to the output of DBAccessBlock which can be passed as input to the next block
# for more information on output() and output references, please take a look at class Block
# and its output() method
output_ref = db_lookup.output()
# content of output_ref:
{
    "result": [
        {
            "id": 1,
            "brand": "Aston Martin",
            "model": "Vanquish"
        },
        {
            "id": 2,
            "brand": "Jaguar",
            "model": "XE 2018"
        }
    ]
}

Database output for the same query (SELECT * FROM CAR)

../_images/db_access_query_example.png
IDENTIFIER = 'DB_ACCESS'
class flows_sdk.implementations.idp_v39.additional_blocks.LlamaBlock(model_id, prompt, reference_name=None, consent=False, tokenize_only=False, predict_extra_kwargs=None, title='Llama Text Generation', description='Text generation using the Llama 2 model')

Bases: Block

Text generation using the Llama 2 model by Meta.

This block utilizes Llama 2 by Meta. Llama 2 is licensed under the LLAMA 2 Community License (https://ai.meta.com/llama/license/). By using the Llama Block, you acknowledge that you have received a copy of the LLAMA 2 Community License.

Note: This block is still experimental and not ready for production use.

Mandatory parameters:

Parameters:
  • model_id (str) – UUID, identifier of the asset with the ML model. Model assets are not part of the Hypersicence bundle and must first be downloaded and imported into the system before they can be used. Please contact your Hyperscience representative to discuss the usecase and the appropriate model and to get instructions on downloading the models. Once downloaded, here are the available model asset UUIDs: b3da05c7-061c-4a58-af6d-4c905d2e618f - Llama 2, 13B, 8-bit quantized

  • prompt (str) – The prompt to generate text from. Use a CodeBlock ahead of the LlamaBlock in order to build the prompt. The context window of the model is limited to 4096 tokens. If the prompt has more tokens, the block would return an error. The total input and output length also cannot exceed 4096 tokens, which means that if the input is too long, the output may be truncated. Also see tokenize_only for a more detailed explanation of how to handle longer texts.

Optional parameters:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • consent (Optional[bool]) – Has the business user read and acknowledged the Llama 2 legal notice. Unless the flow would be used in CI automation, this should be set to False so that a business user can read and acknowledge the notice in the Hyperscience UI. Setting this to True in the Python definition of the flow is also treated as acknowledgement of the notice on behalf of the entity that will be using the flow.

  • tokenize_only (Optional[bool]) – In the context of LLMs, tokens correspond roughly to words, partial words and punctuation. A rough way to compute the number of tokens in a text is number_of_tokens = 1.25 * number_of_words. The Llama block works with a context window of 4096 tokens. If a longer prompt is passed in, it would return an error. Also, if the prompt is shorter, but still close to 4096, the output may be truncated. Since the length of the prompt in tokens cannot be known ahead of time, sometimes the flow may need to make dynamic decisions whether to pass the text to the Llama block, or to split it into multiple chunks, or to use some alternative method for processing it (e.g. a custom supervision manual task). The tokenize_only option enables flows to take such dynamic decisions based on the length of the prompt in tokens. The default is false - it instructs the block to run the ML model inference over the prompt, i.e. to generate a text response. When true is passed, the model will simply tokenize the prompt without running inference and return the list of tokens - this will work regardless of the length of the prompt. A subsequent Decision task can then route longer prompts to an alternative path in the flow, while routing shorter prompts to the LlamaBlock for actual text generation.

  • predict_extra_kwargs (Optional[Dict[Any, Any]]) – Extra keyword arguments used for model inference. (Used to fine-tune the model prediction. Not recommended to change this unless you know what you are doing)

Example usage:

llama = LlamaBlock(
    reference_name='llama',
    model_id='b3da05c7-061c-4a58-af6d-4c905d2e618f',
    prompt='${previous_block.output.llama_prompt}',
)
IDENTIFIER = 'LLAMA'
class flows_sdk.implementations.idp_v39.additional_blocks.EntityRecognitionBlock(pages, regex_map=None, keyword_map=None, reference_name=None, title='Entity Recognition', description='Recognizes entities of different types in documents')

Bases: Block

Entity Recognition Block detects entities of different types within the documents contained in a submission

Mandatory parameters:

Parameters:

pages (str) – list of pages that includes the raw documents that need to be transcribed

Optional Parameters:

Parameters:
  • regex_map (Optional[Dict[str, str]]) – a map that specifies different regex patterns that the block should look for, a default one is present

  • keyword_map (Optional[Dict[str, List[str]]]) – a map that specifies different keyword patterns that the block should look for, a default one is present

  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

Example usage:

# This is one way to obtain the submission object (using SubmissionBootstrapBlock)
submission_bootstrap = SubmissionBootstrapBlock(reference_name='submission_bootstrap')

# EntityRecognitionBlock detects entities within the documents contained in the submission
entity_recognition_block = EntityRecognitionBlock(
    reference_name='entity_recognition',
    pages=submission_bootstrap.output(submission.unassigned_pages),
    regex_map=DEFAULT_PII_REGEX_MAP,
    keyword_map=DEFAULT_PII_KEYWORD_MAP,
)

# reference to the output of EntityRecognitionBlock which can be passed as input to the
# next block. For more information on output() and output references, please take a look at
# class Block and its output() method
output_ref = entity_recognition_block.output()

Example content of output_ref:

{
    "documents": [
        {
            "file_uuid": "da2c6117-43c4-4b30-84c4-cdc59cc3fa4f",
            "entities": [
                {
                    "text": "John Smith",
                    "type": "Person Name",
                    "positions": [
                        [0.5, 0.1, 0.55, 0.13],
                        [0.6, 0.2, 0.65, 0.23],
                        [0.7, 0.3, 0.75, 0.33],
                    ],
                    "page_ids": [0, 0, 1],
                    "origin": "NER",
                },
                {
                    "text": "<SIGNED>",
                    "type": "signature",
                    "positions": [[0.701171875, 0.732421875, 0.92578125, 0.8681640625]],
                    "page_ids": [0],
                    "origin": "Signature",
                },
            ]
        },
        {
            "file_uuid": "4aa38f84-e5af-4cdc-ba5c-edd60f0a6b89",
            "entities": [
                ...
            ]
        }
    ]
}
IDENTIFIER = 'ENTITY_RECOGNITION'

Trigger Blocks (v39)

Note

“Triggers” are referred to as “Input Blocks” in our non-technical documentation.

Warning

Flows are built to allow full definition of triggers either in code or manually in the Flow Studio. We recommend defining triggers manually in the Flow Studio. Instructions on how to do so are in our non-technical documentation.

To enable trigger definition via the Flow Studio UI, you need to instantiate the IDPTriggers() convenience class and pass it to the triggers property of the Flow

return Flow(
   ...
   triggers=IDPTriggers()
)

If instead you wish to define specific triggers in Python, use the trigger definitions below.

Folder Listener (v39)

Identifier: FOLDER_TRIGGER

Parameters:

Name

Type

Default Value

Title

Description

folder

Path

N/A

Folder to scan for submissions

N/A

file_extensions

string

N/A

File extensions

List of file extensions to monitor for (e.g.: ‘png, jpg, pdf’)

has_meta

boolean

False

Enable metadata

Select this if a metadata file is to be expected along with document files (in XXX_index.txt file)

poll_interval

integer

10

Poll interval

Poll interval in seconds

warmup_interval

integer

15

Warm-up interval

Seconds to wait past document last modified time before processing it

folder_cleanup_delay

integer

86400

Folder cleanup delay

Seconds to wait before cleaning up subfolders

api_params

object

{}

API Parameters

N/A

folder_listener = IOBlock(
    identifier='FOLDER_TRIGGER',
    reference_name='folder_trigger',
    title='Folder listener',
    enabled=True,
    input={
        'folder': '/var/www/forms/forms/folder/',
        'file_extensions': 'png, pdf, jpg',
    },
)

triggers = IDPTriggers(blocks=[folder_listener])

S3 Listener (v39)

Identifier: S3_TRIGGER Parameters:

Name

Type

Default Value

Title

Description

bucket_uri

string

N/A

S3 source URI

S3 URI, where the files will be uploaded (e.g. s3://hs_bucket/prefix1/)

archive_bucket_uri

string

N/A

S3 archive URI

S3 URI to move the processed files to (e.g. s3://hs_bucket/prefix1/)

file_extensions

array

[]

File extensions

Types of file extensions for which to monitor

other_file_extensions

string

N/A

Other file extensions

Comma separated list of file extensions for which to monitor (e.g. ‘png, jpg, pdf’)

has_meta

boolean

false

Include submission level parameters

Select this to ingest JSON files along with document files and submission s3 prefixes. These JSON files can contain information such as metadata, cases and external_id. These JSON file names should match the related files or s3 prefixes (e.g., XXX.jpg.json for XXX.jpg)

aws_access_key_id

string

N/A

AWS ACCESS KEY ID

Access Key used for authentication

aws_secret_access_key

password

N/A

SECRET ACCESS KEY

Secret Key used for authentication

use_ec2_instance_credentials

boolean

true

Use AWS EC2 Instance IAM Role Credentials

If selected, credentials are obtained from the EC2 instance directly, and the AWS Access Key ID and Secret Key are not used

poll_interval

integer

10

Polling interval (in seconds)

How often the S3 Listener will check the S3 bucket for submissions

warmup_interval

integer

15

Warm-up interval (in seconds)

How long the S3 Listener will wait to process the document after it was first detected

api_params

object

{}

API Parameters

N/A

s3_listener = IOBlock(
    identifier='S3_TRIGGER',
    reference_name='s3_trigger',
    title='S3 Listener',
    enabled=True,
    input={
        'bucket_uri': 's3://hs_bucket/prefix1/',
        'archive_bucket_uri': 's3://hs_bucket/prefix2/',
        'file_extensions': ['png', 'pdf', 'jpg'],
        'aws_access_key_id': 'admin',
        'aws_secret_access_key': system_secret('s3_secret_access_key'),
        'use_ec2_instance_credentials': False,
    },
)
triggers = IDPTriggers(blocks=[s3_listener])

Email Listener (v39)

Identifier: IMAP_TRIGGER

Parameters:

Name

Type

Default Value

Title

Description

provider [1]

string

imap

Email Provider

N/A

poll_interval

integer

60

Polling interval in seconds

The frequency at which the connection will monitor the email folder for submissions

folder

Path

N/A

Folder to scan for emails

The email folder to retrieve submissions from (e.g., “inbox”)

email_body_treatment

string

“ignore”

Email body treatment

What to do with the body of an email that is ingested. Available options are “process” and “ignore”.

post_process_action

string

“move”

Post process action

What to do with the email after it is processed. Available options are “move” and “delete”.

post_process_move_folder

string

“”

Post process archive folder

Folder to move emails to once they are processed. It will be included only if post_process_action is set to “move”.

api_params

object

{}

API Parameters

N/A

render_headers [2]

array

[]

Headers to include

Headers to render at the top of the email body. Headers will be included only if email_body_treatment is set to “process”.

[1] one of:

imap
graph

[2] any of:

To
From
Subject
Date

Additional inputs when "imap" is the selected provider:

Name

Type

Default Value

Title

Description

host

string

N/A

IMAP server address

N/A

port

integer

993

Port Number

N/A

ssl

boolean

True

Use SSL connection

N/A

username

string

N/A

Username

N/A

password

Password

N/A

Password

N/A

Additional inputs when "graph" is the selected provider:

Name

Type

Default Value

Title

Description

client_id

string

N/A

Client id

Application (client) ID for the corresponding application registered with the Microsoft identity platform.

tenant_id

string

N/A

Tenant id

Also called directory ID. Unique identifier of the tenant in which the corresponding application is registered with.

graph_cloud_endpoint [1]

string

Global

Microsoft Graph Cloud Endpoint

Determines base URL used for accessing cloud services through Microsoft Graph.

azure_ad_endpoint [2]

string

AZURE_PUBLIC_CLOUD

Azure AD Endpoint

Determines base URL for the Azure Active Directory (Azure AD) endpoint to acquire token for each national cloud.

client_secret

Password

N/A

Client secret

Client secret for the corresponding application registered with the Microsoft identity platform.

[1] one of:

Global
US_GOV
US_DoD
Germany
China

[2] one of:

AZURE_PUBLIC_CLOUD
AZURE_CHINA
AZURE_GERMANY
AZURE_GOVERNMENT
imap_trigger = IOBlock(
    identifier='IMAP_TRIGGER',
    reference_name='imap_trigger',
    title='IMAP trigger',
    enabled=True,
    input={
        'host': 'example@mail.com',
        'folder': '/var/www/forms/forms/imap/',
        'username': 'admin',
        'password': 'pass',
    }
)

triggers = IDPTriggers(blocks=[imap_trigger])

Box Folder Listener (v39)

Identifier: BOX_FOLDER_TRIGGER

Parameters:

Name

Type

Default Value

Title

Description

source_folder_id

string

N/A

Folder to scan for submissions

Use the Box Folder ID found in the URL

target_folder_id

string

N/A

Folder to move completed files

Use the Box Folder ID found in the URL

file_extensions

array

[]

File extensions

Types of file extensions for which to monitor

custom_file_extensions

string

N/A

Other file extension types

Comma separated list of file extensions for which to monitor (e.g. ‘png, jpg, pdf’)

poll_interval

integer

10

Polling interval (in seconds)

How often the connector will check the base folder for submissions

warmup_interval

integer

15

Warm-up interval (in seconds)

How long the connector will wait to process the document after it was last modified

public_key_id

string

N/A

Public-Key ID

ID of the Public-key used for authentication with Box

private_key

Password

N/A

Private-Key

Private key used for authentication with Box

passphrase

Password

N/A

Passphrase

Passphrase used for authentication with Box

client_id

string

N/A

Client ID

Client Id used for authentication with Box

client_secret

Password

N/A

Client Secret

Client Secret used for authentication with Box

enterprise_id

string

N/A

Enterprise ID

Enterprise Id used for authentication with Box

api_params

object

{}

API Parameters

N/A

box_folder_trigger = IOBlock(
    identifier='BOX_FOLDER_TRIGGER',
    reference_name='box_folder_trigger',
    title='Box folder',
    enabled=True,
    input={
        'file_extensions': ['png', 'pdf', 'jpg'],
        'source_folder_id': '24',
        'target_folder_id': '42',
        'public_key_id': 'admin',
        'private_key': 'secret',
        'passphrase': 'password',
        'client_id': 'admin',
        'client_secret': 'secret',
        'enterprise_id': 'admin',
    }
)

triggers = IDPTriggers(blocks=[box_folder_trigger])

Cron Listener (v39)

Identifier: CRON_TRIGGER

Inputs:

Name

Type

Default Value

Title

Description

cron_spec

string

"*/5 * * * *"

Cron specification

N/A

time_zone

string

“US/Eastern”

Time Zone

N/A

api_params

object

{}

API Parameters

N/A

cron_trigger = IOBlock(
    identifier='CRON_TRIGGER',
    reference_name='cron_trigger',
    title='Cron Trigger',
    enabled=True,
    input={
        'cron_spec': '0 10 * * *',
        'time_zone': 'Europe/Sofia',
    }
)

triggers = IDPTriggers(blocks=[cron_trigger])

Salesforce Listener (v39)

Identifier: SALESFORCE_TRIGGER

Parameters:

Name

Type

Default Value

Title

Description

channel_name

string

N/A

Channel Name

Configured in the Hyperscience app in Salesforce.

username

string

N/A

Associated Username

Username associated with the Salesforce Connected App.

consumer_key

Password

N/A

Consumer Key

Consumer key used for authentication with Salesforce. Paired with the Salesforce instance from which it was issued.

private_key

Password

N/A

Private Key

Private key used for authentication with Salesforce.

sandbox_environment

boolean

false

Private Key

Enable if the Salesforce environment is a sandbox.

api_params

object

{}

API Parameters

N/A

Message Queue Listener (v39)

Identifier: MQ_LISTENER

Inputs:

Name

Type

Default Value

Title

Description

MQ_TYPE

string

“ACTIVE_MQ”

Message Queue type

Type of the message queue to connect with. Valid values are “ACTIVE_MQ”, “AMAZON_SQS”, “IBM_MQ” and “RABBIT_MQ”

MQ_USERNAME

string

null

Username/Access Key ID

N/A

MQ_PASSWORD

Password

null

Password/Secret Access Key

N/A

MQ_QUEUE_NAME

string

null

Queue Name

Name of the queue (topic) to connect to.

MQ_HOST

string

null

Host Name

N/A

MQ_PORT

integer

5672 when “RABBIT_MQ”, 1414 when “IBM_MQ”, otherwise null

Port Number

N/A

api_params

object

{}

API Parameters

N/A

Additional inputs when "AMAZON_SQS" is the selected MQ_TYPE:

Name

Type

Default Value

Title

Description

MQ_REGION

string

null

AWS Region

N/A

MQ_QUEUE_URL

string

null

Queue URL

N/A

MQ_USE_EC2_INSTANCE_CREDENTIALS

boolean

True

Use AWS EC2 Instance IAM Role Credentials

If selected, credentials are obtained from the EC2 instance directly, and the AWS Access Key ID and Secret Key are not used.

NOTE: MQ_USERNAME can be used to provide Access Key ID and MQ_PASSWORD to provide Secret Access Key.

Additional inputs when "IBM_MQ" is the selected MQ_TYPE:

Name

Type

Default Value

Title

Description

MQ_QUEUE_MANAGER

string

null

Queue Manager (IBM MQ)

N/A

MQ_SSL_CIPHER_SUITE [1]

string

null

SSL cipher suite

SSL cipher suite to use if SSL connection is used.

MQ_CHANNEL

string

null

Channel

N/A

MQ_NO_AUTH_REQUIRED

boolean

false

No auth credentials required

N/A

MQ_MQCSP_AUTHENTICATION_MODE

boolean

false

MQCSP authentication mode

N/A

NOTE: MQ_NO_AUTH_REQUIRED is True, MQ_USERNAME and MQ_PASSWORD are optional.

[1] one of:

"NULL_MD5"
"NULL_SHA"
"RC4_MD5_EXPORT"
"RC4_MD5_US"
"RC4_SHA_US"
"RC2_MD5_EXPORT"
"DES_SHA_EXPORT"
"RC4_56_SHA_EXPORT1024"
"DES_SHA_EXPORT1024"
"TRIPLE_DES_SHA_US"
"TLS_RSA_WITH_NULL_SHA256"
"TLS_RSA_WITH_AES_128_CBC_SHA"
"TLS_RSA_WITH_AES_128_CBC_SHA256"
"TLS_RSA_WITH_AES_256_CBC_SHA"
"TLS_RSA_WITH_AES_256_CBC_SHA256"
"AES_SHA_US"
"TLS_RSA_WITH_DES_CBC_SHA"
"TLS_RSA_WITH_3DES_EDE_CBC_SHA"
"TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256"
"FIPS_WITH_DES_CBC_SHA"
"FIPS_WITH_3DES_EDE_CBC_SHA"

Additional inputs when "RABBIT_MQ" is the selected MQ_TYPE:

Name

Type

Default Value

Title

Description

MQ_VIRTUAL_HOST

string

null

Virtual Host (RABBIT MQ)

N/A

MQ_EXCHANGE

string

“”

Exchange

N/A

MQ_ROUTING_KEY

string

“”

Routing key

N/A

MQ_CONNECTION_TYPE

string

“TCP”

Connection Type

“TCP or SSL”

rabbit_mq_listener = IOBlock(
    identifier='MQ_LISTENER',
    reference_name='mq_listener',
    title='RabbitMQ Listener',
    enabled=True,
    input={
        'MQ_TYPE': 'RABBIT_MQ',
        'MQ_QUEUE_NAME': 'some_queue_name',
        'MQ_HOST': 'somehost.com',
        'MQ_USERNAME': 'foo',
        'MQ_PASSWORD': system_secret('rabbit_mq_1_password'),
    }
)

triggers = IDPTriggers(blocks=[rabbit_mq_listener])

Universal Folder Listener (v39)

Identifier: UNIVERSAL_FOLDER_TRIGGER

Inputs:

Name

Type

Default Value

Title

Description

folder

string

N/A

Folder to scan for submissions

Enter the path relative to the base folder. Leave blank to monitor the base folder

file_extensions

array

[]

File extensions

Types of file extensions for which to monitor

other_file_extensions

string

N/A

Other file extensions

Comma separated list of file extensions for which to monitor (e.g. ‘png, jpg, pdf’)

has_meta

boolean

false

Include submission level parameters

Select this to ingest JSON files along with document files and submission folders. These JSON files can contain information such as metadata, cases and external_id. These JSON file names should match the related files or folders (e.g., XXX.jpg.json for XXX.jpg)

poll_interval

integer

10

Polling interval (in seconds)

How often the Folder Listener will check the base folder for submissions

warmup_interval

integer

15

Warm-up interval (in seconds)

How long the Folder Listener will wait to process the document after it was last modified

folder_cleanup_delay

number

24

Empty folder cleanup delay (in hours)

How often the Folder Listener will remove empty folders from the base folder

api_params

object

{}

API Parameters

N/A

universal_folder_listener = IOBlock(
    identifier='UNIVERSAL_FOLDER_TRIGGER',
    reference_name='universal_folder_trigger',
    title='Universal Folder Listener',
    enabled=True,
    input={
        'folder': '/var/www/forms/forms/folder/',
        'file_extensions': 'png, pdf, jpg',
    }
)

triggers = IDPTriggers(blocks=[universal_folder_listener])

Kafka Listener (v39)

Identifier: KAFKA_TRIGGER

Inputs:

Name

Type

Default Value

Title

Description

bootstrap_servers

string

N/A

Bootstrap Servers

A Comma separated list of host:port pairs to use for establishing the initial connection to the cluster

group_id

string

N/A

Group Id

A unique string that identifies the consumer group this consumer belongs to. BLOCK_THREADS_KAFKA_TRIGGER should be used to scale up or down the number of consumers (default: number of available cores)

topics

string

N/A

Topics

Comma separated list of topic names that the block will subscribe to

consumer_config_json

json

N/A

Consumer Config

JSON formatted string containing key-value pairs of Kafka consumer configurations described in the official Kafka documentation

api_params

object

{}

API Parameters

N/A

consume_timeout

integer

-1

Consume Timeout

Maximum time, in seconds, that the consumer will wait for new messages. If messages are received within the specified time, they are processed immediately. However, if no messages are received within that specified time, consumption cycle will be restarted internally to continue fetching messages. This ensures continuous message consumption (default: infinite)

consume_num_messages

integer

1

Consume Num Messages

Sets the maximum number of messages the consumer fetches in a single consumption cycle.If the number of available messages is less than the specified, and timeout is reached, messages will be processed. Consumption cycle will be restarted after the current finishes

kafka_listener = IOBlock(
    identifier='KAFKA_TRIGGER',
    reference_name='kafka_trigger',
    title='Kafka Listener',
    enabled=True,
    input={
        'folder': '/var/www/forms/forms/folder/',
        'file_extensions': 'png, pdf, jpg',
         'bootstrap_servers': 'example.com',
         'group_id': 'test_group',
         'topics': 'test_topic1, test_topic2',
         'consumer_config_json': {},
         'number_of_workers': 1,
         'consume_timeout': 5,
         'consume_num_messages': 10
    }
)

triggers = IDPTriggers(blocks=[kafka_listener])

Output Blocks (v39)

Warning

Flows pass information to downstream systems using the IDPOutputsBlock. This block is specially built to allow full definition either in code or manually in the Flow Studio. We recommend defining outputs manually in the Flow Studio.

Instructions on how to do so are in our non-technical documentation. Simply include an IDPOutputsBlock in your Python flow definition wherever you wish to send information to downstream systems.

class flows_sdk.implementations.idp_v39.idp_blocks.IDPOutputsBlock(inputs, blocks=None)

Bases: Outputs

Output block allows users to send data extracted by an IDP flow to other systems for downstream processing

Mandatory parameters:

Parameters:

inputs (Dict[str, Any]) – Used by the UI to automatically pre-populate the input of newly added blocks. Extended in this class to include 'enabled': True to visualize an enabled/disabled trigger that is controllable via Flow Studio.

Usage when parameters are to be defined in Flow Studio:

outputs = IDPOutputBlock(
    inputs={'submission': submission_bootstrap.output('result.submission')}
)

If instead you wish to define specific outputs in Python, use the outputs definitions below.

HTTP Notifier Output (v39)

The HTTP Notifier (REST) output connection will POST results from the system to a specified HTTP endpoint.

Identifier: COB_HTTP_EXPORT

Parameters:

Name

Type

Default Value

Title

Description

submission

Submission

N/A

Submission Object

N/A

enabled

boolean

True

Enabled

Enable or disable this output block

api_version [1]

string

N/A

API Version

API version to be used for rendering payload

export_type

string

N/A

Export Type

Payload rendering choice. If “Flat” is chosen, a simplified version of the payload will be sent, as described in the “flat” parameter for submission retrieval API

export_routing_filter

string

null

Routing Filter

By default, status change notifications are sent for each submission processed by the system. You can filter what notifications are received through this connection by specifying an expression like has_layout_tag(‘some_layout_tag’) or has_source_routing_tag(‘some_source_routing_tag’)

endpoint

string

N/A

Endpoint URL

URL that output notifications will be posted to

endpoint_secrets

string

N/A

Endpoint URL secrets

Endpoint URL Secrets, format: (one per line) secret_key_name=secret_value

timeout

int

600

Timeout (seconds)

How long (in seconds) to keep the connection open if no bytes are received from the endpoint. Set to 0 to wait indefinitely.

authorization_type [2]

string

“none”

Authorization type

Type of authorization

authorization_header_name

string

null

Authorization header name

Authorization header name to be set in the notification request

authorization_header

Password

null

Authorization header value

Authorization header to be set in the notification request

auth_url

string

null

OAuth2 authorization URL

The endpoint for the authorization server

client_id

string

null

Client ID

The client identifier issued to the client during the application registration process

client_secret

Password

null

Client Secret

The client secret issued to the client during the application registration process

scope

string

null

Scope

Scope to request to oauth server

audience

string

null

Audience

Resource service URL where token will be valid

additional_oauth_parameters

json

{}

Additional Oauth Request Parameters

Additional parameters to send when requesting token

ssl_cert_source [3]

string

null

Source

N/A

ssl_cert_path

string

null

Trusted Certificate(s) path

Enter the path relative to the base folder

ssl_cert_value

MultilineText

null

Trusted Certificate(s)

N/A

additional_notification_http_headers

json

null

Additional Notification Request Headers

Key value pairs in json format which will be sent via the request as http headers

[1] one of:

"v5"
"transformed"

[2] one of:

"none"
"http_header"
"oauth_2_client_credentials"

[3] one of:

"env_var"
"ca_bundle_path"
"certificate_value"
cob_http_export = IOBlock(
    identifier='COB_HTTP_EXPORT',
    reference_name='cob_http_export',
    title='HTTP Notifier Output Block',
    enabled=True,
    input={
        'submission': submission_bootstrap.output('result.submission'),
        'enabled': True,
        'api_version': 'v5',
        'export_type': 'SUBMISSION',
        'endpoint': 'example.com',
        'authorization_type': 'none',
    }
)

outputs = IDPOutputsBlock(
    inputs={'submission': submission_bootstrap.output('result.submission')},
    blocks=[cob_http_export],
)

Message Queue Notifier Output (v39)

The Message Queue Notifier Output can configure connections to ActiveMQ, Amazon SQS, IBM MQ, and RabbitMQ message queues.

Identifier: COB_MQ_NOTIFIER

Parameters:

Name

Type

Default Value

Title

Description

submission

Submission

N/A

Submission Object

N/A

enabled

boolean

True

Enabled

Enable or disable this output block

api_version [1]

string

N/A

API Version

API version to be used for rendering payload

export_type

string

N/A

Export Type

Payload rendering choice. If “Flat” is chosen, a simplified version of the payload will be sent, as described in the “flat” parameter for submission retrieval API

export_routing_filter

string

null

Routing Filter

By default, status change notifications are sent for each submission processed by the system. You can filter what notifications are received through this connection by specifying an expression like has_layout_tag(‘some_layout_tag’) or has_source_routing_tag(‘some_source_routing_tag’)

MQ_TYPE

string

“ACTIVE_MQ”

Message Queue type

Type of the message queue to connect with. Valid values are “ACTIVE_MQ”, “AMAZON_SQS”, “IBM_MQ” and “RABBIT_MQ”

MQ_USERNAME

string

null

Username/Access Key ID

N/A

MQ_PASSWORD

Password

null

Password/Secret Access Key

N/A

MQ_QUEUE_NAME

string

null

Queue Name

Name of the queue (topic) to connect to. It applies only to “ACTIVE_MQ”, “IBM_MQ” and “RABBIT_MQ” MQ_TYPEs.

MQ_HOST

string

null

Host Name

N/A

MQ_PORT

integer

5672 when “RABBIT_MQ”, 1414 when “IBM_MQ”, otherwise null

Port Number

N/A

[1] one of:

"v5"
"transformed"

Additional inputs when "AMAZON_SQS" is the selected MQ_TYPE:

Name

Type

Default Value

Title

Description

MQ_REGION

string

null

AWS Region

N/A

MQ_QUEUE_URL

string

null

Queue URL

N/A

MQ_USE_EC2_INSTANCE_CREDENTIALS

boolean

True

Use AWS EC2 Instance IAM Role Credentials

If selected, credentials are obtained from the EC2 instance directly, and the AWS Access Key ID and Secret Key are not used.

MQ_MESSAGE_GROUP_ID

string

null

Group ID for FIFO queues

N/A

MQ_MESSAGE_METADATA

string

null

Additional SQS Metadata

Input additional metadata key value pairs in the format of { “key”: {“TYPE”: “String”, “VALUE”: “”}, “key2”: {“TYPE”: “Number”, “VALUE”: “12345”}

NOTE: MQ_USERNAME can be used to provide Access Key ID and MQ_PASSWORD to provide Secret Access Key.

Additional inputs when "IBM_MQ" is the selected MQ_TYPE:

Name

Type

Default Value

Title

Description

MQ_QUEUE_MANAGER

string

null

Queue Manager (IBM MQ)

N/A

MQ_SSL_CIPHER_SUITE [1]

string

null

SSL cipher suite

SSL cipher suite to use if SSL connection is used.

MQ_CHANNEL

string

null

Channel

N/A

MQ_NO_AUTH_REQUIRED

boolean

false

Channel

N/A

MQ_MQCSP_AUTHENTICATION_MODE

boolean

false

Channel

N/A

NOTE: MQ_NO_AUTH_REQUIRED is True, MQ_USERNAME and MQ_PASSWORD are optional.

[1] one of:

"NULL_MD5"
"NULL_SHA"
"RC4_MD5_EXPORT"
"RC4_MD5_US"
"RC4_SHA_US"
"RC2_MD5_EXPORT"
"DES_SHA_EXPORT"
"RC4_56_SHA_EXPORT1024"
"DES_SHA_EXPORT1024"
"TRIPLE_DES_SHA_US"
"TLS_RSA_WITH_NULL_SHA256"
"TLS_RSA_WITH_AES_128_CBC_SHA"
"TLS_RSA_WITH_AES_128_CBC_SHA256"
"TLS_RSA_WITH_AES_256_CBC_SHA"
"TLS_RSA_WITH_AES_256_CBC_SHA256"
"AES_SHA_US"
"TLS_RSA_WITH_DES_CBC_SHA"
"TLS_RSA_WITH_3DES_EDE_CBC_SHA"
"TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256"
"FIPS_WITH_DES_CBC_SHA"
"FIPS_WITH_3DES_EDE_CBC_SHA"

Additional inputs when "RABBIT_MQ" is the selected MQ_TYPE:

Name

Type

Default Value

Title

Description

MQ_VIRTUAL_HOST

string

null

Virtual Host (RABBIT MQ)

N/A

MQ_EXCHANGE

string

“”

Exchange

N/A

MQ_ROUTING_KEY

string

“”

Routing key

N/A

MQ_CONNECTION_TYPE

string

“TCP”

Connection Type

“TCP or SSL”

mq_notifier = IOBlock(
    identifier='COB_MQ_NOTIFIER',
    reference_name='mq_notifier',
    title='MQ Notifier Output Block',
    enabled=True,
    input={
        'submission': submission_bootstrap.output('result.submission'),
        'enabled': True,
        'api_version': 'v5',
        'export_type': 'SUBMISSION',
        'MQ_USERNAME': 'admin',
        'MQ_PASSWORD': 'pass',
        'MQ_QUEUE_NAME': 'queue',
        'MQ_HOST': 'host',
    },
)

outputs = IDPOutputsBlock(
    inputs={'submission': submission_bootstrap.output('result.submission')},
    blocks=[mq_notifier],
)

Box Notifier Output (v39)

Provides an out-of-the-box integration into Box systems.

Identifier: COB_BOX_NOTIFIER

Parameters:

Name

Type

Default Value

Title

Description

submission

Submission

N/A

Submission Object

N/A

enabled

boolean

True

Enabled

Enable or disable this output block

api_version [1]

string

N/A

API Version

API version to be used for rendering payload

export_type

string

N/A

Export Type

Payload rendering choice. If “Flat” is chosen, a simplified version of the payload will be sent, as described in the “flat” parameter for submission retrieval API

export_routing_filter

string

null

Routing Filter

By default, status change notifications are sent for each submission processed by the system. You can filter what notifications are received through this connection by specifying an expression like has_layout_tag(‘some_layout_tag’) or has_source_routing_tag(‘some_source_routing_tag’)

public_key_id

string

N/A

Public-Key ID

ID of the Public-key used for authentication with Box

private_key

Password

N/A

Private-Key

Private key used for authentication with Box

passphrase

Password

N/A

Passphrase

Passphrase used for authentication with Box

client_id

string

N/A

Client ID

Client Id used for authentication with Box

client_secret

Password

N/A

Client Secret

Client Secret used for authentication with Box

enterprise_id

string

N/A

Enterprise ID

Enterprise Id used for authentication with Box

template_key

string

“”

Box Metadata Template Key

Enter the key of the Box Metadata template that you would like to map the Hyperscience metadata to.

static_output_fields_to_map

array

[]

Static Metadata Fields

Specify the Hyperscience fields you want to store in Box metadata.

submission_id

string

“”

Key for Mapping Submission ID

Specify a Box metadata field key from your chosen Box template. The Hyperscience field will be stored in this specified field.

document_id

string

“”

Key for Mapping Document ID

Specify a Box metadata field key from your chosen Box template. The Hyperscience field will be stored in this specified field.

submission_state

string

“”

Key for Mapping Submission State

Specify a Box metadata field key from your chosen Box template. The Hyperscience field will be stored in this specified field.

document_state

string

“”

Key for Mapping Document State

Specify a Box metadata field key from your chosen Box template. The Hyperscience field will be stored in this specified field.

submission_exceptions

string

“”

Key for Mapping Submission Exceptions

Specify a Box metadata field key from your chosen Box template. The Hyperscience field will be stored in this specified field.

document_exceptions

string

“”

Key for Mapping Document Exceptions

Specify a Box metadata field key from your chosen Box template. The Hyperscience field will be stored in this specified field.

document_layout_uuid

string

“”

Key for Mapping Document Layout Uuid

Specify a Box metadata field key from your chosen Box template. The Hyperscience field will be stored in this specified field.

document_layout_name

string

“”

Key for Mapping Document Layout Name

Specify a Box metadata field key from your chosen Box template. The Hyperscience field will be stored in this specified field.

start_time

string

“”

Key for Mapping Start Time

Specify a Box metadata field key from your chosen Box template. The Hyperscience field will be stored in this specified field.

complete_time

string

“”

Key for Mapping Complete Time

Specify a Box metadata field key from your chosen Box template. The Hyperscience field will be stored in this specified field.

document_fields_template_mappings

json

{}

JSON for Additional Metadata Fields

Specify additional metadata fields using their key value. Please consult Box Integration setup manual for JSON template and instructions.

[1] one of:

"v5"
"transformed"
box_notifier = IOBlock(
    identifier='COB_BOX_NOTIFIER',
    reference_name='box_notifier',
    title='Box Notifier Output Block',
    enabled=True,
    input={
        'submission': submission_bootstrap.output('result.submission'),
        'enabled': True,
        'private_key': 'password',
        'passphrase': 'password',
        'public_key_id': 'admin',
        'client_id': 'admin',
        'client_secret': 'secret',
        'enterprise_id': 'admin',
        'template_key': 'key',
    },
)

outputs = IDPOutputsBlock(
    inputs={'submission': submission_bootstrap.output('result.submission')},
    blocks=[box_notifier],
)

UiPath Notifier Output (v39)

Provides an out-of-the-box integration into UiPath systems.

Identifier: COB_EXPORT_UIPATH

Parameters:

Name

Type

Default Value

Title

Description

submission

Submission

N/A

Submission Object

N/A

enabled

boolean

True

Enabled

Enable or disable this output block

api_version [1]

string

N/A

API Version

API version to be used for rendering payload

export_type

string

N/A

Export Type

Payload rendering choice. If “Flat” is chosen, a simplified version of the payload will be sent, as described in the “flat” parameter for submission retrieval API

export_routing_filter

string

null

Routing Filter

By default, status change notifications are sent for each submission processed by the system. You can filter what notifications are received through this connection by specifying an expression like has_layout_tag(‘some_layout_tag’) or has_source_routing_tag(‘some_source_routing_tag’)

auth_method [2]

string

“basic_auth”

Authentication method

N/A

url

string

N/A

URL

Base URL of the uipath instance

app_id

string

N/A

App ID

App ID provided by UiPath

app_secret

Password

N/A

App Secret

App Secret provided by UiPath

oauth_token_endpoint

string

N/A

OAuth token endpoint

Endpoint to retrieve the token. Using https://cloud.uipath.com/identity_/connect/token by default

organization_unit_id

string

N/A

Organization unit id

Organization unit id

tenant

string

N/A

Tenant

Tenant

username

string

N/A

Username

Username

password

Password

N/A

Password

Password

queue

string

N/A

Queue

Queue

[1] one of:

"v5"
"transformed"

[2] one of:

"basic_auth"
"oauth"
cob_export_uipath = IOBlock(
    identifier='COB_EXPORT_UIPATH',
    reference_name='cob_export_uipath',
    title='CCB Export UiPath',
    enabled=True,
    input={
        'submission': submission_bootstrap.output('result.submission'),
        'enabled': True,
        'api_version': 'v5',
        'export_type': 'FLAT_SUBMISSION',
        'url': 'example.com',
        'tenant': 'tenant',
        'username': 'admin',
        'password': 'pass',
        'queue': 'queue',
    }
)

outputs = IDPOutputsBlock(
    inputs={'submission': submission_bootstrap.output('result.submission')},
    blocks=[cob_export_uipath],
)

Salesforce Notifier Output (v39)

With the Salesforce Notifier Block, you can configure your flow to send extracted information to Salesforce. The Salesforce Notifier Block can use extracted information to look up, and then update any number of fields on a Salesforce object. The Salesforce Notifier Block is available in both SaaS and on-premise versions of the Hyperscience application.

Identifier: COB_SALESFORCE_NOTIFIER

Parameters:

Name

Type

Default Value

Title

Description

submission

Submission

N/A

Submission Object

N/A

enabled

boolean

True

Enabled

Enable or disable this output block

api_version [1]

string

N/A

API Version

API version to be used for rendering payload

export_type

string

N/A

Export Type

Payload rendering choice. If “Flat” is chosen, a simplified version of the payload will be sent, as described in the “flat” parameter for submission retrieval API

export_routing_filter

string

null

Routing Filter

By default, status change notifications are sent for each submission processed by the system. You can filter what notifications are received through this connection by specifying an expression like has_layout_tag(‘some_layout_tag’) or has_source_routing_tag(‘some_source_routing_tag’)

username

string

N/A

Associated Username

Username associated with the Salesforce Connected App

consumer_key

password

N/A

Consumer Key

Consumer key used for authentication with Salesforce. Paired with the Salesforce instance from which it was issued.

private_key

password

N/A

Private Key

Private key used for authentication with Salesforce.

sandbox_environment

boolean

false

Private Key

Enable if the Salesforce environment is a sandbox.

object_api_name

string

N/A

Object API Name

API name of the object in Salesforce to lookup and update.

new_record_on_lookup_failure

boolean

false

Create new record on lookup failure

When enabled, it will create a new record if the lookup fails to find a record using the lookup parameters. When disabled, lookup will fail if it does not find a record using the lookup parameters.

object_lookup_mapping

json

N/A

Lookup Field Mapping

Specify which Salesforce fields (using their API name) to be queried with which extracted values from the document for the lookup using key-value format. For example: “DOCUMENT_FIELD_NAME”: “OBJECT_FIELD_API_NAME”.

document_fields_mapping

json

N/A

Document Field Mappings

Specify which Salesforce fields (using their API name) you want to update with which extracted data from Hyperscience using key-value format. For example: “DOCUMENT_FIELD_NAME”: “OBJECT_FIELD_API_NAME”.”

overwrite_existing_field_data

boolean

false

Overwrite existing field data

If this is checked and a selected field has data in it, Hyperscience will overwrite that data. If this is unchecked, existing data will not be overwritten and the operation will fail.

[1] one of:

"v5"
"transformed"
cob_salesforce_notifier = IOBlock(
    identifier='COB_SALESFORCE_NOTIFIER',
    reference_name='cob_salesforce_notifier',
    title='Salesforce Notifier Output Block',
    enabled=True,
    input={
        'submission': submission_bootstrap.output('result.submission'),
        'enabled': True,
        'username': 'salesforce_user',
        'consumer_key': 'salesforce_consumer_key',
        'private_key': 'salesforce_private_key',
        'object_api_name': 'salesforce_object_api_name',
        'object_lookup_mapping': {"LAYOUT_VARIATION_XYZ_FIELD_NAME_1": "SF_FIELD_API_NAME_1", "LAYOUT_VARIATION_XYZ_FIELD_NAME_2": "SF_FIELD_API_NAME_2"},
        'document_fields_mapping': {"LAYOUT_VARIATION_XYZ_FIELD_NAME_3": "SF_FIELD_API_NAME_3", "LAYOUT_VARIATION_XYZ_FIELD_NAME_4": "SF_FIELD_API_NAME_4"}
    }
)

outputs = IDPOutputsBlock(
    inputs={'submission': submission_bootstrap.output('result.submission')},
    blocks=[cob_salesforce_notifier],
)

S3 Notifier Output (v39)

With the S3 Notifier Block, you can configure your flow to send extracted information to an S3 bucket.

Identifier: COB_S3_NOTIFIER

Parameters:

Name

Type

Default Value

Title

Description

submission

Submission

N/A

Submission Object

N/A

enabled

boolean

True

Enabled

Enable or disable this output block

api_version [1]

string

N/A

API Version

API version to be used for rendering payload

export_type

string

N/A

Export Type

Payload rendering choice. If “Flat” is chosen, a simplified version of the payload will be sent, as described in the “flat” parameter for submission retrieval API

export_routing_filter

string

null

Routing Filter

By default, status change notifications are sent for each submission processed by the system. You can filter what notifications are received through this connection by specifying an expression like has_layout_tag(‘some_layout_tag’) or has_source_routing_tag(‘some_source_routing_tag’)

bucket_uri

string

N/A

S3 URI

S3 URI, where the data should be stored (e.g. s3://hs_bucket/prefix1/prefix2/)

aws_access_key_id

string

N/A

AWS ACCESS KEY ID

Access Key used for authentication

aws_secret_access_key

password

N/A

SECRET ACCESS KEY

Secret Key used for authentication

use_ec2_instance_credentials

boolean

true

Use AWS EC2 Instance IAM Role Credentials

If selected, credentials are obtained from the EC2 instance directly, and the AWS Access Key ID and Secret Key are not used

[1] one of:

"v5"
"transformed"
cob_s3_notifier = IOBlock(
    identifier='COB_S3_NOTIFIER',
    reference_name='cob_s3_notifier',
    title='S3 Notifier Output Block',
    enabled=True,
    input={
        'submission': submission_bootstrap.output('result.submission'),
        'enabled': True,
        'api_version': 'v5',
        'export_type': 'FLAT_SUBMISSION',
        'bucket_uri': 's3://hs_bucket/prefix1/prefix2/',
        'aws_access_key_id': 'aws_access',
        'aws_secret_access_key': system_secret('aws_secret_access_key'),
    }
)

outputs = IDPOutputsBlock(
    inputs={'submission': submission_bootstrap.output('result.submission')},
    blocks=[cob_s3_notifier],
)

Kafka Notifier Output (v39)

With the Kafka Notifier Block, you can configure your flow to send extracted information to a kafka topic.

Identifier: COB_KAFKA_NOTIFIER

Parameters:

Name

Type

Default Value

Title

Description

submission

Submission

N/A

Submission Object

N/A

enabled

boolean

True

Enabled

Enable or disable this output block

api_version [1]

string

N/A

API Version

API version to be used for rendering payload

export_type

string

N/A

Export Type

Payload rendering choice. If “Flat” is chosen, a simplified version of the payload will be sent, as described in the “flat” parameter for submission retrieval API

export_routing_filter

string

null

Routing Filter

By default, status change notifications are sent for each submission processed by the system. You can filter what notifications are received through this connection by specifying an expression like has_layout_tag(‘some_layout_tag’) or has_source_routing_tag(‘some_source_routing_tag’)

bootstrap_servers

string

N/A

Bootstrap Servers

A list of Kafka brokers to connect to. For example: “kafka1:9092,kafka2:9092”

topic

string

N/A

Topic

The Kafka topic where your data will be sent

client_id

string

N/A

SECRET ACCESS KEY

An identifier for your application when connecting to Kafka, typically used for tracking and logging purposes

producer_config_json

json

true

Producer Config (JSON)

Configuration settings for the Kafka producer in JSON format. See https://kafka.apache.org/documentation/#producerconfigs for more information

synchronous_publish

boolean

true

Sync Publish

If true, the publish call will block until the message is sent successfully or an error is raised. If false, the publish call will return immediately after enqueueing the message

message_key

string

null

Message Key

The key associated with the message to be sent. This is used to determine which partition the message is sent to. If null, the message will be sent to a random partition

[1] one of:

"v5"
"transformed"
cob_kafka_notifier = IOBlock(
    identifier='COB_KAFKA_NOTIFIER',
    reference_name='cob_kafka_notifier',
    title='Kafka Notifier Output Block',
    enabled=True,
    input={
        'submission': submission_bootstrap.output('result.submission'),
        'enabled': True,
        'api_version': 'v5',
        'export_type': 'FLAT_SUBMISSION',
        'bootstrap_servers': 'kafka1:9092,kafka2:9092',
        'topic': 'topic',
        'client_id': 'client_id',
        'producer_config_json': {},
        'synchronous_publish': True,
        'message_key': 'key'
    }
)

outputs = IDPOutputsBlock(
    inputs={'submission': submission_bootstrap.output('result.submission')},
    blocks=[cob_kafka_notifier],
)

Helper Classes and Functions (v39)

The following classes are used to easily instantiate and manage a flow’s settings.

flows_sdk.implementations.idp_v39.idp_values.get_idp_wf_inputs(idp_wf_config)

A helper function that maps the values provided by an IdpWorkflowConfig to their corresponding IDP keys.

Parameters:

idp_wf_config (IdpWorkflowConfig) – with static values to be filled.

Return type:

Dict[str, Any]

Returns:

a dictionary with filled static values, compatible with IDP flows.

Example usage with default values:

flow = Flow(
    input=get_idp_wf_inputs(get_idp_wf_config()),
    ...
)

Example usage:

Example usage when instantiating an IDP flow with custom parameters::

flow = Flow(
    input=get_idp_wf_inputs({an instance of IdpWorkflowConfig}),
    ...
)
flows_sdk.implementations.idp_v39.idp_values.get_idp_wf_config()

A helper function that instantiates an IdpWorkflowConfig pre-filled with default values. IdpWorkflowConfig is the container class for defining flow-level settings.

Return type:

IdpWorkflowConfig

Returns:

a valid, fully constructed IdpWorkflowConfig instance

Example usage when instantiating an IDP flow:

flow = Flow(
    input=get_idp_wf_inputs(get_idp_wf_config()),
    ...
)
class flows_sdk.implementations.idp_v39.idp_values.IdpWorkflowConfig(rotation_correction_enabled, document_grouping_logic, qa_config, manual_document_organization_enabled, transcription_automation_training, transcription_training_legibility_period, transcription_model, manual_transcription_enabled, finetuning_only_trained_layouts, flex_confidence_boosting_enabled, improved_transcription_threshold_accuracy, structured_text, semi_structured_text, semi_structured_table, structured_checkbox, structured_signature, semi_structured_checkbox, field_id_target_accuracy, table_id_target_accuracy, manual_field_id_enabled, machine_classification_config, manual_transcription_config, manual_identification_config, flexible_extraction_config, notification_workflow_initialization=None, notification_workflow_classification=None, notification_workflow_identification=None, notification_workflow_transcription=None, notification_workflow_flexible_extraction=None)

Convenience dataclass outlining all flow-level config values for IDP.

class flows_sdk.implementations.idp_v39.idp_values.IDPTriggers(blocks=None)

An adapter class making it easier to instantiate flows_sdk.flows.Triggers for IDP flows. In particular, provides defaults for some fields (e.g. title, description, reference_name) and defines api_params_manifest with IDP specific parameters.

Parameters:

blocks (Optional[Sequence[IOBlock]]) – flows_sdk.blocks.IOBlock s to be included as triggers for an IDP Flow. Optional, defaults to an empty list.

Example usage when instantiating an IDP flow:

folder_listener = IOBlock(...)
flow = Flow(
    triggers=IDPTriggers([folder_listener]),
    ...
)
class flows_sdk.implementations.idp_v39.idp_values.IDPWrapperManifest(flow_identifier)

An adapter class making it easier to instantiate flows_sdk.flows.Manifest for IDP wrapper flows. In particular, provides defaults for some fields (e.g. roles) but more importantly, defines all flows_sdk.flows.Parameter that construct the Flow inputs and their UI representation (left-hand side menu in Flow studio).

Parameters:

flow_identifier (str) – A system-wide unique identifier for flows_sdk.flows.Flow

Example usage when instantiating an IDP flow:

flow = Flow(
    manifest=IDPWrapperManifest(flow_identifier='FLOW_IDENTIFIER'),
    ...
)
class flows_sdk.implementations.idp_v39.idp_values.IDPCoreManifest(flow_identifier, roles=None)

An adapter class making it easier to instantiate flows_sdk.flows.Manifest for IDP Core flows. In particular, provides defaults for some fields (e.g. roles, output) but more importantly, defines all flows_sdk.flows.Parameter that construct the Flow inputs and their UI representation (left-hand side menu in Flow studio).

Parameters:
  • flow_identifier (str) – A system-wide unique identifier for flows_sdk.flows.Flow

  • roles (Optional[List[str]]) – ‘read_only’, ‘supporting’ and ‘idp_core’ are the default roles of every subflow that can replace the default IDP_CORE subflow inside the top-level IDP flows. ‘supporting’ and ‘idp_core’ are mandatory roles that are necessary for all IDP Core flows. ‘supporting’ ensures that the subflow will be auto-deployed together with the top-level IDP flow. ‘idp_core’ ensures that this subflow can be selected in the UI in dropdowns which allow picking an alternative subflow to the default IDP_CORE subflow. If you are implementing a subflow that is not supposed to be a drop-in replacement for IDP_CORE, you will have to subclass or implement your own version of IDPCoreManifest class in order to remove the idp_core role.

Example usage when instantiating an IDP flow:

flow = Flow(
    manifest=IDPCoreManifest(flow_identifier='FLOW_IDENTIFIER'),
    ...
)

V38

Processing Blocks (v38)

Core Blocks are powerful processing blocks necessary to build intelligent document processing solutions on Hyperscience. They include both manual and machine processing blocks.

Note

Classes in the idp_v38 implementation have parameters with ambiguous typing. For example, a field of type integer could also be represented as a string reference (e.g. that integer gets provided at runtime). The types introduced in the example are therefore incomplete and are likely to be revisioned in a future version.

class flows_sdk.implementations.idp_v38.idp_blocks.IDPCoreBlock(idp_wf_config, reference_name='__idp_core__', title='Start Document Processing Subflow', description='Calls the Document Processing Subflow with the specified settings.', compatibility_spec=CompatibilitySpec(filter_roles=['idp_core'], filter_schema=None, label=None), identifier='IDP_CORE_V38', additional_inputs=None)

Bases: Block

IDP core block includes the steps to process the passed documents into machine-readable output.

Mandatory parameters:

Parameters:

idp_wf_config (IdpWorkflowConfig) – IdpWorkflowConfig object with all flow-level configuration settings for IDP

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • compatibility_spec (Optional[CompatibilitySpec]) – defines a filter for flows that could be selected in the UI in place of “IDP Core” for the block. Defaults to flows with the role ‘idp_core’. The user will be able to pick any flow that matches the compatibility_spec. This serves to make upgrading to a new compatible version of “IDP Core” (e.g. IDP_CORE_V39) easier in future releases.

  • identifier (str) – The identifier of the flow. Defaults to the IDP_CORE_VXX identifier of the out-of-the-box flow that comes with the system.

  • additional_inputs (Optional[Dict[str, Any]]) – Dictionary containing key-value pairs that will be added to the inputs of the block. If a key is contained in the default inputs, its value will be overwritten by the value provided here. You generally do not need to provide additional_inputs, unless you are creating a custom derivative of the IDP_CODE_Vxx flow and are modifying or extending the list of inputs it accepts.

Example usage:

idp_core = IDPCoreBlock(idp_wf_config=idp_wf_config)
REFERENCE_NAME = '__idp_core__'
class flows_sdk.implementations.idp_v38.idp_blocks.SubmissionBootstrapBlock(reference_name=None, layout_release_uuid='${workflow.input.layout_release_uuid}', s3_config='${workflow.input.submission_bootstrap_s3_config}', s3_endpoint_url='${workflow.input.submission_bootstrap_s3_endpoint_url}', ocs_config='${workflow.input.submission_bootstrap_ocs_config}', http_config='${workflow.input.submission_bootstrap_http_config}', notification_workflow='${workflow.input.notification_workflow_initialization}', workflow_uuid='${workflow.input.workflow_uuid}', workflow_name='${workflow.input.workflow_name}', workflow_version='${workflow.input.workflow_version}', submission='${workflow.input.submission}')

Bases: Block

Submission bootstrap block initializes the submission object and prepares external images or other submission data if needed. This block should be called at the START of every flow that uses blocks included in the library.

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • layout_release_uuid (str) – Layout Release UUID at submission creation time, defaults to workflow_input(Settings.LayoutReleaseUuid)

  • s3_config (str) – Specify Amazon S3 credentials for submission retrieval store. Expects a json expression: {"aws_access_key_id": "X", "aws_secret_access_key": "Y"}, defaults to system_secret(S3_SECRET_KEY)

  • s3_endpoint_url (str) – Endpoint URL for S3 submission retrieval store, defaults to None

  • ocs_config (str) – OCS Configuration for downloading submission data, defaults to system_secret(OCS_SECRET_KEY)

  • http_config (Optional[str]) – An optional system secret field expressing the HTTP Configuration for downloading submission data. The content is expected to be a json formatted as: {"username": "X", "password": "Y", "ssl_cert": "CA bundle filename"}, defaults to None

  • notification_workflow (str) – Notification flow name to run when the submission is initialized, defaults to IDP_SUBMISSION_NOTIFY_NAME

  • workflow_uuid (str) – UUID of the triggered workflow version, defaults to workflow_input(Inputs.WorkflowUuid)

  • workflow_name (str) – Name of the triggered workflow, defaults to workflow_input(Inputs.WorkflowName)

  • workflow_version (str) – Version of the triggered workflow, defaults to workflow_input(Inputs.WorkflowVersion)

Example usage:

submission_bootstrap = SubmissionBootstrapBlock(reference_name='submission_bootstrap')
IDENTIFIER = 'SUBMISSION_BOOTSTRAP_2'
output(key=None)

Submission bootstrap ends with a CodeBlock, which nests the actual outputs under the ‘result’ key. This method overrides the default output in the same manner that flows_sdk.blocks.Routing.CodeBlock does for convenience.

Parameters:

key (Optional[str]) –

Optionally provide a key to directly get a nested property. When not provided, the entire ‘result’ will be returned. For example, we have:

{
    "result": {
        "a": {
            "b": 42
        }
    }
}
  • skipping the key output()` will return ``{"a": {"b": 42}}

  • calling with output("a") will result in {"b": 42}

  • calling with output("a.b") will return 42

Return type:

str

Returns:

the value under the provided key

class flows_sdk.implementations.idp_v38.idp_blocks.MachineCollationBlock(submission, cases, dedupe_files='${workflow.input.machine_collation_dedupe_files}', refresh_retention_period='${workflow.input.machine_collation_refresh_retention_period}', remove_from_cases='${workflow.input.machine_collation_remove_from_cases}', retention_period='${workflow.input.machine_collation_retention_period}', reference_name=None)

Bases: Block

Machine collation block groups files, documents and pages (from the submission) into cases.

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • cases (Any) –

    This parameter accepts an array of JSON objects that contain information on how Cases should be created. There are two ways to collate a case: by the filenames specified in the current Submission, or by the ids of the Documents or Pages submitted in previous submissions.

    This parameter follows the given format:

    {
        'cases': [
            'external_case_id': 'HS-1',
            'filename': 'file1.pdf',
            'documents': [42],
            'pages': [43]
        ]
    }
    

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • dedupe_files (Union[bool, str]) – Enabling this setting will replace case data from repeated file names within the same case. Cases will retain data from the most recently submitted version of a particular file. Note that this setting does not delete the old data, it just removes it from the case. Keep in mind this option is only relevant when adding to cases

  • remove_from_cases (Optional[Any]) –

    This parameter accepts an array of JSON objects that contains information on how documents or pages should be removed from a pre-existing case. Currently, the only way to de-collate from a case is by the ids of the Documents or Pages within a given case

    This parameter follows the given format:

    {
        'remove_from_cases': [
            'external_case_id': 'HS-1',
            'documents': [42],
            'pages': [43]
        ]
    }
    

  • retention_period (Union[int, str, None]) – The number of days to retain cases after they are updated by this block. If passed None, no changes to the deletion date will be made. By default, this value is None

  • refresh_retention_period (Union[bool, str]) – If True, (re)applies Retention Period parameter to all cases passed into the block. If False, only updates those without a pre-existing deletion date. By default, this value is True

Example usage:

machine_collation = MachineCollationBlock(
    reference_name='machine_collation',
    submission=submission_bootstrap.output('result.submission'),
    cases=submission_bootstrap.output('result.api_params.cases'),
    dedupe_files=True,
    remove_from_cases=custom_supervision.output('remove_from_cases')
)
IDENTIFIER = 'MACHINE_COLLATION_2'
class flows_sdk.implementations.idp_v38.idp_blocks.MachineClassificationBlock(submission, api_params, reference_name=None, rotation_correction_enabled='${workflow.input.machine_classification_rotation_correction_enabled}', mobile_processing_enabled='${workflow.input.machine_classification_mobile_processing_enabled}', file_transcription_enabled='${workflow.input.machine_classification_file_transcription_enabled}', layout_release_uuid='${workflow.input.layout_release_uuid}', vpc_registration_threshold='${workflow.input.structured_layout_match_threshold}', nlc_enabled='${workflow.input.semi_structured_classification}', nlc_target_accuracy='${workflow.input.semi_target_accuracy}', nlc_doc_grouping_logic='${workflow.input.semi_doc_grouping_logic}')

Bases: Block

Machine classification block automatically matches documents to structured, semi-structured or additional layouts

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • api_params (Any) –

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • rotation_correction_enabled (Union[bool, str]) – Identifies and corrects the orientation of semi-structured images, defaults to True

  • mobile_processing_enabled (Union[bool, str]) – Improves the machine readability of photo captured documents, defaults to False

  • layout_release_uuid (str) – Layout Release UUID at submission creation time, defaults to workflow_input(Settings.LayoutReleaseUuid)

  • vpc_registration_threshold (str) – Structured pages above this threshold will be automatically matched to a layout variation, defaults to workflow_input(Settings.StructuredLayoutMatchThreshold)

  • nlc_enabled (str) – Enables workflow to manage a model for automated classification of semi-structured and additional layout variations, defaults to workflow_input(Settings.SemiStructuredClassification)

  • nlc_target_accuracy (str) – defaults to workflow_input(Settings.SemiTargetAccuracy)

  • nlc_doc_grouping_logic (str) – Logic to handle multiple pages matched to the same layout variationin a given submission, defaults to workflow_input(Settings.SemiDocGroupingLogic)

Example usage:

machine_classification = MachineClassificationBlock(
    reference_name='machine_classification',
    submission=case_collation.output('submission'),
    api_params=submission_bootstrap.output('result.api_params'),
    rotation_correction_enabled=idp_wf_config.rotation_correction_enabled,
)
IDENTIFIER = 'MACHINE_CLASSIFICATION_5'
class flows_sdk.implementations.idp_v38.idp_blocks.ManualClassificationBlock(submission, api_params, reference_name=None, layout_release_uuid='${workflow.input.layout_release_uuid}', manual_nlc_enabled='${workflow.input.manual_nlc_enabled}', task_restrictions='${workflow.input.manual_classification_task_restrictions}', notification_workflow='${workflow.input.notification_workflow_classification}', nlc_enabled='${workflow.input.semi_structured_classification}', nlc_target_accuracy='${workflow.input.semi_target_accuracy}', mobile_processing_enabled='${workflow.input.machine_classification_mobile_processing_enabled}')

Bases: Block

Manual classification block allows keyers to manually match submissions to their layouts. Keyers may perform manual classification if machine classification cannot automatically match a submission to a layout with high confidence

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • api_params (Any) –

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • layout_release_uuid (str) – Layout Release UUID at submission creation time, defaults to workflow_input(Settings.LayoutReleaseUuid)

  • manual_nlc_enabled (str) – Enables manual classification when applicable, defaults to workflow_input(Settings.ManualNlcEnabled)

  • task_restrictions (Union[List[Any], str, None]) – Defines what users can access Supervision tasks created by this block, defaults to None

  • notification_workflow (str) – Notification flow name to run when the submission enters Manual Classification, defaults to IDP_SUBMISSION_NOTIFY_NAME

Example usage:

manual_classification = ManualClassificationBlock(
    reference_name='manual_classification',
    submission=machine_classification.output('submission'),
    api_params=submission_bootstrap.output('result.api_params'),
)
IDENTIFIER = 'MANUAL_CLASSIFICATION_2'
class flows_sdk.implementations.idp_v38.idp_blocks.MachineIdentificationBlock(submission, api_params, layout_release_uuid='${workflow.input.layout_release_uuid}', reference_name=None, manual_field_id_enabled='${workflow.input.manual_field_id_enabled}', field_id_target_accuracy='${workflow.input.field_id_target_accuracy}', table_id_target_accuracy='${workflow.input.table_id_target_accuracy}')

Bases: Block

Machine identification automatically identify fields and tables in the submission

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • api_params (Any) –

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • manual_field_id_enabled (str) – Enables manual identification when applicable, defaults to workflow_input(Settings.ManualFieldIdEnabled)

  • field_id_target_accuracy (str) – Field ID Target Accuracy, defaults to workflow_input(Settings.FieldIdTargetAccuracy)

  • table_id_target_accuracy (str) – Table ID Target Accuracy, defaults to workflow_input(Settings.TableIdTargetAccuracy)

Example usage:

machine_identification = MachineIdentificationBlock(
    reference_name='machine_identification',
    submission=manual_classification.output('submission'),
    api_params=submission_bootstrap.output('result.api_params'),
)
IDENTIFIER = 'MACHINE_IDENTIFICATION_6'
class flows_sdk.implementations.idp_v38.idp_blocks.ManualIdentificationBlock(submission, api_params, reference_name=None, task_restrictions='${workflow.input.manual_identification_task_restrictions}', manual_field_id_enabled='${workflow.input.manual_field_id_enabled}', layout_release_uuid='${workflow.input.layout_release_uuid}', notification_workflow='${workflow.input.notification_workflow_identification}')

Bases: Block

Manual identification allows keyers to complete field identification or table identification tasks, where they draw bounding boxes around the contents of certain fields, table columns or table rows. This identification process ensures that the system will be able to transcribe the correct content in the upcoming transcription process

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • api_params (Any) –

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • task_restrictions (Union[List[Any], str, None]) – Defines what users can access Supervision tasks created by this block, defaults to []

  • manual_field_id_enabled (str) – Enables manual identification when applicable, defaults to workflow_input(Settings.ManualFieldIdEnabled)

  • layout_release_uuid (str) – Layout Release UUID at submission creation time, defaults to workflow_input(Settings.LayoutReleaseUuid)

  • notification_workflow (str) – Notification flow name to run when the submission enters Manual Identification, defaults to DEFAULT_IDP_SUBMISSION_NOTIFY_IDENTIFIER

Example usage:

manual_identification = ManualIdentificationBlock(
    reference_name='manual_identification',
    submission=machine_identification.output('submission'),
    api_params=submission_bootstrap.output('result.api_params'),
    task_restrictions=idp_wf_config.manual_identification_config.task_restrictions,
)
IDENTIFIER = 'MANUAL_IDENTIFICATION_5'
class flows_sdk.implementations.idp_v38.idp_blocks.MachineTranscriptionBlock(submission, api_params, reference_name=None, layout_release_uuid='${workflow.input.layout_release_uuid}', transcription_model='${workflow.input.transcription_model}', finetuning_only_trained_layouts='${workflow.input.finetuning_only_trained_layouts}', transcription_overrides='${workflow.input.transcription_overrides}', structured_text_target_accuracy='${workflow.input.structured_text_target_accuracy}', structured_text_confidence_threshold='${workflow.input.structured_text_threshold}', structured_text_acceptable_confidence='${workflow.input.structured_min_leg_threshold}', semi_structured_text_target_accuracy='${workflow.input.semi_structured_text_target_accuracy}', semi_structured_text_confidence_threshold='${workflow.input.semi_structured_text_threshold}', semi_structured_text_acceptable_confidence='${workflow.input.semi_structured_min_leg_threshold}', semi_structured_table_target_accuracy='${workflow.input.semi_structured_table_target_accuracy}', semi_structured_table_confidence_threshold='${workflow.input.semi_structured_table_threshold}', semi_structured_table_acceptable_confidence='${workflow.input.semi_structured_table_min_leg_threshold}', structured_checkbox_target_accuracy='${workflow.input.structured_checkbox_target_accuracy}', structured_checkbox_confidence_threshold='${workflow.input.structured_checkbox_threshold}', structured_checkbox_acceptable_confidence='${workflow.input.checkbox_min_leg_threshold}', structured_signature_target_accuracy='${workflow.input.structured_signature_target_accuracy}', structured_signature_confidence_threshold='${workflow.input.structured_signature_threshold}', structured_signature_acceptable_confidence='${workflow.input.signature_min_leg_threshold}', semi_structured_checkbox_target_accuracy='${workflow.input.semi_structured_checkbox_target_accuracy}', semi_structured_checkbox_confidence_threshold='${workflow.input.semi_structured_checkbox_threshold}', semi_structured_checkbox_acceptable_confidence='${workflow.input.semi_structured_checkbox_min_leg_threshold}')

Bases: Block

Machine transcription automatically transcribes the content of your submission

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • api_params (Any) –

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • layout_release_uuid (str) – Layout Release UUID at submission creation time, defaults to workflow_input(Settings.LayoutReleaseUuid)

  • transcription_model (str) – The transcription model that will be used for this flow, defaults to workflow_input(Settings.TranscriptionModel)

  • finetuning_only_trained_layouts (str) – defaults to workflow_input( Settings.FinetuningOnlyTrainedLayouts )

  • transcription_overrides (str) – defaults to workflow_input( Settings.TranscriptionOverrides )

  • structured_text_target_accuracy (str) – defaults to workflow_input( Settings.StructuredTextTargetAccuracy )

  • structured_text_confidence_threshold (str) – defaults to workflow_input( Settings.StructuredTextThreshold )

  • structured_text_acceptable_confidence (str) – defaults to workflow_input( Settings.StructuredTextMinLegThreshold )

  • semi_structured_text_target_accuracy (str) – defaults to workflow_input( Settings.SemiStructuredTextTargetAccuracy )

  • semi_structured_text_confidence_threshold (str) – defaults to workflow_input( Settings.SemiStructuredTextThreshold )

  • semi_structured_text_acceptable_confidence (str) – defaults to workflow_input( Settings.SemiStructuredTextMinLegThreshold )

  • structured_checkbox_target_accuracy (str) – defaults to workflow_input( Settings.StructuredCheckboxTargetAccuracy )

  • structured_checkbox_confidence_threshold (str) – defaults to workflow_input( Settings.StructuredCheckboxThreshold )

  • structured_checkbox_acceptable_confidence (str) – defaults to workflow_input( Settings.StructuredCheckboxMinLegThreshold )

  • structured_signature_target_accuracy (str) – defaults to workflow_input( Settings.StructuredSignatureTargetAccuracy )

  • structured_signature_confidence_threshold (str) – defaults to workflow_input( Settings.StructuredSignatureThreshold )

  • structured_signature_acceptable_confidence (str) – defaults to workflow_input( Settings.StructuredSignatureMinLegThreshold )

  • semi_structured_checkbox_target_accuracy (str) – defaults to workflow_input( Settings.SemiStructuredCheckboxTargetAccuracy )

  • semi_structured_checkbox_confidence_threshold (str) – defaults to workflow_input( Settings.SemiStructuredCheckboxThreshold )

  • semi_structured_checkbox_acceptable_confidence (str) – defaults to workflow_input( Settings.SemiStructuredCheckboxMinLegThreshold )

Example usage:

machine_transcription = MachineTranscriptionBlock(
    reference_name='machine_transcription',
    submission=manual_identification.output('submission'),
    api_params=submission_bootstrap.output('result.api_params'),
)
IDENTIFIER = 'MACHINE_TRANSCRIPTION_9'
class flows_sdk.implementations.idp_v38.idp_blocks.ManualTranscriptionBlock(submission, api_params, reference_name=None, supervision_transcription_masking='${workflow.input.manual_transcription_supervision_transcription_masking}', table_output_manual_review='${workflow.input.manual_transcription_table_output_manual_review}', always_supervise_blank_cells='${workflow.input.manual_transcription_always_supervise_blank_cells}', task_restrictions='${workflow.input.manual_transcription_task_restrictions}', manual_transcription_enabled='${workflow.input.manual_transcription_enabled}', notification_workflow='${workflow.input.notification_workflow_transcription}', title='Manual Transcription', description='Manually transcribe fields that could not be automatically transcribed')

Bases: Block

Manual transcription lets your keyers manually enter the text found in fields or tables that could not be automatically transcribed

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • api_params (Any) –

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • supervision_transcription_masking (Union[bool, str]) – Prevents users from inputting invalid characters during Supervision Transcription tasks, defaults to True

  • table_output_manual_review (Union[bool, str]) – Always generates a table transcription task if the layout contains a table. If disabled, a table transcription task will only be generated if one or more cells have transcribed values below the defined thresholds, defaults to False

  • task_restrictions (Union[List[Any], str, None]) – Defines what users can access Supervision tasks for submission from a particular source or submissions matching a specific layout, defaults to []

  • manual_transcription_enabled (str) – Enables manual transcription when applicable, defaults to workflow_input(Settings.ManualTranscriptionEnabled)

  • notification_workflow (str) – Notification flow name to run when the submission enters Manual Transcription, defaults to DEFAULT_IDP_SUBMISSION_NOTIFY_IDENTIFIER

Example usage:

manual_transcription = ManualTranscriptionBlock(
    reference_name='manual_transcription',
    submission=machine_transcription.output('submission'),
    api_params=submission_bootstrap.output('result.api_params'),
    supervision_transcription_masking=(
        idp_wf_config.manual_transcription_config.supervision_transcription_masking
    ),
    table_output_manual_review=(
        idp_wf_config.manual_transcription_config.table_output_manual_review
    ),
    task_restrictions=idp_wf_config.manual_transcription_config.task_restrictions,
)
IDENTIFIER = 'MANUAL_TRANSCRIPTION_5'
class flows_sdk.implementations.idp_v38.idp_blocks.FlexibleExtractionBlock(submission, api_params, reference_name=None, supervision_transcription_masking='${workflow.input.flexible_extraction_supervision_transcription_masking}', task_restrictions='${workflow.input.flexible_extraction_task_restrictions}', layout_release_uuid='${workflow.input.layout_release_uuid}', notification_workflow='${workflow.input.notification_workflow_flexible_extraction}')

Bases: Block

Flexible extraction manually transcribes fields marked for review

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • api_params (Any) –

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • supervision_transcription_masking (Union[bool, str, None]) – Prevents users from inputting invalid characters during the Flexible Extraction task, defaults to True

  • task_restrictions (Union[List[Any], str, None]) – Defines what users can access Supervision tasks created by this block, defaults to []

  • layout_release_uuid (str) – Layout Release UUID at submission creation time, defaults to workflow_input(Settings.LayoutReleaseUuid)

  • notification_workflow (str) – Notification flow name to run when the submission enters Flexible Extraction, defaults to DEFAULT_IDP_SUBMISSION_NOTIFY_IDENTIFIER

Example usage:

flexible_extraction = FlexibleExtractionBlock(
    reference_name='flexible_extraction',
    submission=manual_transcription.output('submission'),
    api_params=submission_bootstrap.output('result.api_params'),
    supervision_transcription_masking=(
        idp_wf_config.flexible_extraction_config.supervision_transcription_masking
    ),
    task_restrictions=idp_wf_config.flexible_extraction_config.task_restrictions,
)
IDENTIFIER = 'FLEXIBLE_EXTRACTION_4'
class flows_sdk.implementations.idp_v38.idp_blocks.ReprocessingBlock(reference_name, submission, parent_flow='${workflow.definition.identifier}', idp_wf_config=None, trigger='${workflow.input.$trigger}')

Bases: Block

Reprocessing block enables a second extraction attempt if documents were initially confidently mis-classified by the machine. When an end-user specifies that a document layout is incorrect during Manual Identification, Manual Transcription, or Flexible Extraction, the flow will reach the Reprocessing block which will mark correctly classified documents as “complete” status, and misclassified documents with “reprocess” status will be processed in a second execution of the flow. This block should be placed before the SubmissionCompleteBlock to enable reprocessing.

Mandatory parameters:

Parameters:

submission (Any) – Submission object

IDENTIFIER = 'REPROCESSING'
class flows_sdk.implementations.idp_v38.idp_blocks.SubmissionCompleteBlock(submission, is_reprocessing=False, reference_name=None, nlc_qa_sampling_ratio='${workflow.input.semi_qa_sample_rate}', field_id_qa_enabled='${workflow.input.field_id_qa_enabled}', field_id_qa_sampling_ratio='${workflow.input.field_id_qa_sample_rate}', table_id_qa_enabled='${workflow.input.table_id_qa_enabled}', table_id_qa_sampling_ratio='${workflow.input.table_id_qa_sample_rate}', transcription_qa_enabled='${workflow.input.qa}', transcription_structured_entry_qa_sample_rate='${workflow.input.structured_entry_qa_sample_rate}', transcription_structured_signature_qa_sample_rate='${workflow.input.structured_signature_qa_sample_rate}', transcription_structured_checkbox_qa_sample_rate='${workflow.input.structured_checkbox_qa_sample_rate}', transcription_semi_structured_qa_sample_rate='${workflow.input.semi_structured_qa_sample_rate}', table_cell_transcription_qa_enabled='${workflow.input.table_cell_transcription_qa_enabled}', table_cell_transcription_qa_sample_rate='${workflow.input.table_cell_transcription_qa_sample_rate}', auto_qa_sampling_rate_enabled='${workflow.input.auto_qa_sampling_rate_enabled}', exclude_auto_transcribe_fields_from_qa='${workflow.input.exclude_auto_transcribe_fields_from_qa}', payload=None)

Bases: Block

Submission complete block finalizes submission processing and updates reporting data. This block should be called at the END of every flow that uses blocks included in the flows_sdk.implementations.idp_v34 library. The block requires both the submission object and a payload to be passed in.

Mandatory parameters:

Parameters:
  • submission (str) – Submission object

  • payload (Optional[Any]) – Object to pass to downstream systems

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • nlc_qa_sampling_ratio (Union[str, int]) – Defines the percentage of documents sampled for Classification QA, defaults to workflow_input(Settings.SemiQaSampleRate)

  • field_id_qa_enabled (Union[str, bool]) – defaults to workflow_input(Settings.FieldIdQaEnabled)

  • field_id_qa_sampling_ratio (Union[str, int]) – Defines the percentage of documents sampled for Identification QA, defaults to workflow_input(Settings.FieldIdQaSampleRate)

  • table_id_qa_enabled (Union[str, bool]) – Allows users to verify the location of table cells, defaults to workflow_input(Settings.TableIdQaEnabled)

  • table_id_qa_sampling_ratio (Union[str, int]) – Defines the percentage of tables the system samples for QA, defaults to workflow_input(Settings.TableIdQaSampleRate)

  • transcription_qa_enabled (Union[str, bool]) – defaults to workflow_input(Settings.TranscriptionQaEnabled)

  • transcription_qa_sampling_ratio – Defines the percentage of fields sampled for Transcription QA, defaults to workflow_input(Settings.TranscriptionQaSampleRate)

  • table_cell_transcription_qa_enabled (Union[str, bool]) – defaults to workflow_input(Settings.TableCellTranscriptionQaEnabled)

  • table_cell_transcription_qa_sample_rate (Union[str, int]) – Defines the percentage of cells the system samples for QA. This value is likely to be lower than “Transcription QA Sample Rate” since there are more table cells than fields on any given page, defaults to workflow_input(Settings.TableCellTranscriptionQaSampleRate)

  • exclude_auto_transcribe_fields_from_qa (Union[str, bool]) – Defines fields that are marked as auto-transcribe in layout should be excluded from QA or not. defaults to workflow_input(Settings.ExcludeAutoTranscribeFieldsFromQA)

  • is_reprocessing (Union[str, bool]) – boolean to handle execution when reprocessing defaults to False

Example usage:

submission_complete = SubmissionCompleteBlock(
    reference_name='complete_submission',
    submission=flexible_extraction.output('submission')
)
IDENTIFIER = 'SUBMISSION_COMPLETE_4'
class flows_sdk.implementations.idp_v38.idp_blocks.SetTransformedOutputBlock(submission, transformed_output, reference_name=None)

Bases: Block

Set Transformed Output block sets the transformed output for the provided submission. This can later be used to send the transformed data downstream using output blocks. For more information see https://docs.hyperscience.com/#retrieving-transformed-submissions.

Mandatory parameters:

Parameters:
  • submission (str) – Submission object

  • transformed_output (str) – Object that will be set as the transformed output for this submission

IDENTIFIER = 'SET_TRANSFORMED_OUTPUT'
class flows_sdk.implementations.idp_v38.idp_blocks.IdpCustomSupervisionBlock(submission, task_purpose, supervision_template, reference_name=None, supervision_transcription_masking=True, task_restrictions=None, page_ids=None, notification_workflow='IDP_SUBMISSION_NOTIFY_V38', title='IDP Custom Supervision', description='IDP Custom Supervision')

Bases: Block

An IDP wrapper for the Custom Supervision block. It has the same functionality, but handles reading/writing data from the IDP database.

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • task_purpose (str) – the custom task name that’s given to all supervision tasks that run through a particular instance of a custom supervision block. This can be used to filter tasks in the task queue.

  • supervision_template (Any) – JSON used for configuration of the Custom Supervision Task.

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • supervision_transcription_masking (bool) – Prevents users from inputting invalid characters during the Custom Supervision task, defaults to True

  • task_restrictions (Optional[List[Any]]) – Defines what users can access Supervision tasks created by this block, defaults to []

  • page_ids (Optional[List[int]]) – This is an optional parameter. If left empty, the IDP wrapper will, by default, pull in all pages in the submission. Adding a list of page IDs will override the default list of pages. defaults to []

  • notification_workflow (str) – Notification flow name to run when the submission enters Custom Supervision (which is called within the IDP Custom Supervision block), defaults to DEFAULT_IDP_SUBMISSION_NOTIFY_IDENTIFIER

  • title (Optional[str]) – UI-visible title of the block. Defaults to ‘IDP Custom Supervision’

  • description (Optional[str]) – Description of the block. Both useful for documentation purpose and visible to users in the Flow studio. Defaults to ‘IDP Custom Supervision’

Example usage:

idp_custom_supervision = IdpCustomSupervisionBlock(
    reference_name='idp_custom_supervision',
    submission=manual_transcription.output('submission'),
    task_purpose='make_custom_decision',
    supervision_template=[
        {
            'name': 'three_column_template',
            'version': '1.0',
            'thumbnail': {
                'group_by_document': True,
                'group_by_case': True,
            },
            'action': [
                {
                    'name': 'my_tab',
                    'display': 'SSN Info',
                    'input': [
                        {
                            'name': 'ssn_field',
                            'type': 'transcription',
                            'data_identifier': '2afd7a67-ddee-484f-ab6b-2d292752e5ee',
                            'title': 'SSN',
                        },
                    ],
                    'ui': {
                        'groups': [
                            {
                                'title': 'My Group',
                                'fields': ['ssn_field'],
                            },
                        ],
                    },
                },
            ],
        },
    ],
    supervision_transcription_masking=False,
)
IDENTIFIER = 'IDP_CUSTOM_SUPERVISION_2'
class flows_sdk.implementations.idp_v38.idp_blocks.IDPFullPageTranscriptionBlock(submission, reference_name=None, title='Full Page Transcription (Submission)', description='Transcribes the documents included in a submission', app_metadata=None)

Bases: Block

IDP Full Page Transcription Block machine-transcribes the unassigned pages in a submission

Mandatory parameters:

Parameters:

submission (str) – submission object that includes the raw documents that need to be transcribed

Optional Parameters:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • app_metadata (Optional[List[str]]) – used for capturing information about the flow

Example usage:

# This is one way to obtain the submission object (using SubmissionBootstrapBlock)
submission_bootstrap = SubmissionBootstrapBlock(reference_name='submission_bootstrap')
# IDPFullPageTranscriptionBlock transcribes the documents contained in the submission object
idp_fpt_block = IDPFullPageTranscriptionBlock(
    reference_name='idp_fpt',
    submission=submission_bootstrap.output(),
)
# reference to the output of IDPFullPageTranscriptionBlock which can be passed as input to
# the next block. For more information on output() and output references, please take a
# look at class Block and its output() method
output_ref = idp_fpt_block.output()
IDENTIFIER = 'FULL_PAGE_TRANSCRIPTION_SUBMISSION'
class flows_sdk.implementations.idp_v38.idp_blocks.IDPImageCorrectionBlock(submission, reference_name=None, rotation_correction_enabled=True, mobile_processing_enabled=False, title='Image Correction (Submission)', description='Rotate and de-skew documents included in a submission', app_metadata=None)

Bases: Block

IDP Image Correction Block rotates and de-skews tilted/skewed images so that they are more transcription-friendly

Mandatory parameters:

Parameters:

submission (str) – submission object that includes the raw documents that need to be image-corrected

Optional Parameters:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • app_metadata (Optional[List[str]]) – used for capturing information about the flow

Example usage:

# This is one way to obtain the submission object (using SubmissionBootstrapBlock)
submission_bootstrap = SubmissionBootstrapBlock(reference_name='submission_bootstrap')
# IDPImageCorrectionBlock takes in the submission object and outputs in the same format with
# its images
idp_image_correct_block = IDPImageCorrectionBlock(
    reference_name='idp_image_correct',
    submission=submission_bootstrap.output(),
)
# reference to the output of IDPImageCorrectionBlock which can be passed as input to the
# next block. For more information on output() and output references, please take a look at
# class Block and its output() method
output_ref = idp_image_correct_block.output()
IDENTIFIER = 'IMAGE_CORRECTION_SUBMISSION'

Additional Blocks (v38)

class flows_sdk.implementations.idp_v38.additional_blocks.JSONOutputsBlock(inputs, blocks=None)

Bases: Outputs

Output block allows users to send JSON data extracted by an IDP flow to other systems for downstream processing Mandatory parameters:

Parameters:

inputs (Dict[str, Any]) – Used by the UI to automatically pre-populate the input of newly added blocks. Extended in this class to include 'enabled': True to visualize an enabled/disabled trigger that is controllable via Flow Studio.

Usage when parameters are to be defined in Flow Studio:

outputs = JSONOutputBlock(
    inputs={'result': }
)
class flows_sdk.implementations.idp_v38.additional_blocks.ImageCorrectionBlock(pages, reference_name=None, rotation_correction_enabled=True, mobile_processing_enabled=False, title='Image Correction (Pages)', description='Rotate and de-skew documents', app_metadata=None)

Bases: Block

Image Correction Block rotates and de-skews tilted/skewed images so that they are more transcription-friendly Mandatory parameters:

Parameters:

pages (str) – list of page object that includes the raw documents that need to be image-corrected

Optional Parameters:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • app_metadata (Optional[List[str]]) – used for capturing information about the flow

Example usage:

# This is one way to obtain the submission object (using SubmissionBootstrapBlock)
submission_bootstrap = SubmissionBootstrapBlock(reference_name='submission_bootstrap')
# ImageCorrectionBlock takes in the submission object and outputs in the same format with
# its images
image_correct_block = ImageCorrectionBlock(
    reference_name='image_correct',
    pages=submission_bootstrap.output(submission.unassigned_pages),
)
# reference to the output of IDPImageCorrectionBlock which can be passed as input to the
# next block. For more information on output() and output references, please take a look
# at class Block and its output() method
output_ref = image_correct_block.output()
IDENTIFIER = 'IMAGE_CORRECTION_PAGES'
class flows_sdk.implementations.idp_v38.additional_blocks.FullPageTranscriptionBlock(pages, reference_name=None, title='Full Page Transcription (Pages)', description='Transcribes documents', app_metadata=None)

Bases: Block

Full Page Transcription Block machine-transcribes the document pages passed to it Mandatory parameters:

Parameters:

pages (str) – list of pages that includes the raw documents that need to be transcribed

Optional Parameters:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • app_metadata (Optional[List[str]]) – used for capturing information about the flow

Example usage:

# This is one way to obtain the submission object (using SubmissionBootstrapBlock)
submission_bootstrap = SubmissionBootstrapBlock(reference_name='submission_bootstrap')
# FullPageTranscriptionBlock transcribes the documents contained in the submission object
fpt_block = FullPageTranscriptionBlock(
    reference_name='fpt',
    submission=submission_bootstrap.output(submission.unassigned_pages),
)
# reference to the output of FullPageTranscriptionBlock which can be passed as input to the
# next block. For more information on output() and output references, please take a look at
# class Block and its output() method
output_ref = fpt_block.output()
IDENTIFIER = 'FULL_PAGE_TRANSCRIPTION_PAGES'
class flows_sdk.implementations.idp_v38.additional_blocks.CustomSupervisionBlock(submission, task_purpose, data, supervision_template, reference_name=None, supervision_transcription_masking=True, task_restrictions=None, notification_workflow='IDP_SUBMISSION_NOTIFY_V38')

Bases: Block

The CustomSupervisionBlock uses a JSON schema as a form builder to create customer-driven UIs for Supervision tasks.

Mandatory parameters:

Parameters:
  • submission (Any) – Submission object

  • task_purpose (str) – the custom task name that’s given to all supervision tasks that run through a particular instance of a custom supervision block. This can be used to filter tasks in the task queue.

  • data (Any) – JSON structure that supplies the data used to populated the supervision_template

  • supervision_template (Any) – JSON used for configuration of the Custom Supervision Task.

Parameters with defaults:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • supervision_transcription_masking (bool) – Prevents users from inputting invalid characters during the Custom Supervision task, defaults to True

  • task_restrictions (Optional[List[Any]]) – Defines what users can access Supervision tasks created by this block, defaults to []

  • notification_workflow (Optional[str]) – Notification flow name to run when the submission enters Custom Supervision, defaults to IDP_SUBMISSION_NOTIFY_NAME

  • title (Optional[str]) – UI-visible title of the block. Defaults to ‘IDP Custom Supervision’

  • description (Optional[str]) – Description of the block. Both useful for documentation purpose and visible to users in the Flow studio. Defaults to ‘IDP Custom Supervision’

Example usage:

custom_supervision = CustomSupervisionBlock(
    reference_name='custom_supervision',
    submission=manual_transcription.output('submission'),
    task_purpose='make_custom_decision',
    supervision_template=[
        {
            'name': 'three_column_template',
            'version': '1.0',
            'thumbnail': {
                'group_by_document': True,
                'group_by_case': True,
            },
            'action': [
                {
                    'name': 'my_tab',
                    'display': 'SSN Info',
                    'input': [
                        {
                            'name': 'ssn_field',
                            'type': 'transcription',
                            'data_identifier': '2afd7a67-ddee-484f-ab6b-2d292752e5ee',
                            'title': 'SSN',
                        },
                    ],
                    'ui': {
                        'groups': [
                            {
                                'title': 'My Group',
                                'fields': ['ssn_field'],
                            },
                        ],
                    },
                },
            ],
        },
    ],
    data={
        'fields': [
            {
                'id': 967,
                'uuid': 'fe8d4e7a-a82c-448c-85a1-0a3cbbe53b61',
                'value': '111-22-3333',
                'page_id': 15,
                'layout_field_id': '2afd7a67-ddee-484f-ab6b-2d292752e5ee',
                'validation_overridden': False,
                'bounding_box': [
                    0.09360783305186686,
                    0.25792617589433436,
                    0.6358913805295097,
                    0.28862414740388187,
                ],
                'occurrence_index': 1,
            },
        ],
        'template_fields': [
            {
                'uuid': '2afd7a67-ddee-484f-ab6b-2d292752e5ee',
                'notes': '',
                'type': 'entry',
                'data_type_uuid': 'b600a4e4-758f-4c4b-a2f4-cbb5dd870a0c',
                'name': 'SSN',
            },
        ],
        'pages': [
            {
                'id': 15,
                'file_page_number': 1,
                'document_page_number': 1,
                'image_url': '/image/2ed3d989-b64a-44bd-a9fb-6e0dcc5ccdb8',
                'form_id': 2,
                'submission_id': 1,
                'external_case_ids': [],
                'read_only': False,
                'filename': '1 page.pdf',
                'form_page_id': 1
            },
        ],
        'documents': [
            {
                'id': 2,
                'layout_uuid': 'e31c3cbf-dcea-44fd-a052-902463c6040f',
                'layout_name': 'layout_name',
                'page_count': 2,
                'read_only': False,
            },
        ],
        'cases': [],
        'selected_choices': [],
    },
    supervision_transcription_masking=False,
)
IDENTIFIER = 'CUSTOM_SUPERVISION_2'
class flows_sdk.implementations.idp_v38.additional_blocks.SoapRequestBlock(method, reference_name=None, endpoint='', body_namespace=None, headers_namespace=None, headers=None, params=None, title='HTTP REST Block', description='HTTP REST block')

Bases: Block

Soap Block allows user to perform SOAP requests Mandatory parameters:

Parameters:
  • method (str) – SOAP method to invoke

  • endpoint (str) – absolute url including schema to HTTP endpoint for this request

Optional Parameters:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • body_namespace (Optional[str]) – SOAP request body namespace url

  • headers_namespace (Optional[str]) – SOAP request headers namespace url

  • headers (Optional[Dict[Any, Any]]) – SOAP headers to use

  • params (Optional[Dict[Any, Any]]) – key-value pair used as query parameters in the request to be inserted as SOAP body

Example usage:

# minimum viable instantiation, the rest of the fields can later be added in the UI
soap_block = SoapRequestBlock(
    reference_name='soap_request',
    title='Soap Request',
    description='Perform soap request',
    method='ExampleMethod',
)
soap_block = SoapRequestBlock(
    reference_name='soap_request',
    title='Soap Request',
    description='Perform soap request',
    method='ExampleMethod',
    endpoint='http://example.org/abc.wso',
    body_namespace='http://www.example.org/abc.countryinfo',
    params={'example': 'stuff'},
    headers={}
)
# reference to the output of SoapRequestBlock which can be passed as input to the next block
# for more information on output() and output references, please take a look at class Block
# and its output() method
output_ref = soap_block.output()
IDENTIFIER = 'SOAP_REQ'
class flows_sdk.implementations.idp_v38.additional_blocks.HttpRestBlock(method, endpoint, reference_name=None, headers=None, params=None, payload=None, json=None, authorization_type=None, handled_error_codes=None, auth_params=None, title='HTTP REST Block', description='HTTP REST block')

Bases: BaseHttpRestBlock

Http Rest Block allows user to perform HTTP requests Mandatory parameters:

Parameters:
  • method (str) – HTTP method to use

  • endpoint (str) – absolute url including schema to HTTP endpoint for this request

Optional Parameters:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • headers (Optional[Dict[str, str]]) – HTTP headers to use

  • params (Optional[Dict[str, Any]]) – key-value pair used as query parameters in the request

  • payload (Optional[Dict[str, str]]) – key-value pair to be used as x-www-form-urlencoded data. Mutually exclusive with json

  • json (Optional[Dict[str, Any]]) – key-value pair to be used as json data. Mutually exclusive with payload

  • authorization_type (Optional[str]) – authorization type (none, http_header, oauth_2_client_credentials)

  • handled_error_codes (Optional[List[Union[int, str]]]) – specify HTTP error codes that will not fail the task, allowing for wildcards with the ‘x’ char. Example: [400, “5xx”]

  • auth_params (Union[HttpHeaderAuthorizationParams, OAuth2AuthorizationParams, None]) – authorization parameters to be passed when the authorization_type is set to “http_header” or “oauth_2_client_credentials”

Example usage:

# minimum viable instantiation, the rest of the fields can later be added in the UI
http_rest_block = HttpRestBlock(
    reference_name='http_request',
    title='Http Request',
    description='Perform http request',
    method='GET',
)
http_rest_block = HttpRestBlock(
    reference_name='http_request',
    title='Http Request',
    description='Perform http request',
    method='GET',
    endpoint='https://sheets.googleapis.com/v4/spreadsheets/example',
)
# reference to the output of HttpRestBlock which can be passed as input to the next block
# for more information on output() and output references, please take a look at class Block
# and its output() method
output_ref = http_rest_block.output()
class flows_sdk.implementations.idp_v38.additional_blocks.HyperscienceRestApiBlock(method, app_endpoint, reference_name=None, headers=None, params=None, payload=None, json=None, handled_error_codes=None, title='Hyperscience HTTP REST Block', description='Hyperscience HTTP REST block')

Bases: BaseHttpRestBlock

Make Http requests to the Hyperscience platform APIs. This block will automatically take care of prepending the base url and will include the appropriate authentication headers with the request. Mandatory parameters:

Parameters:
  • method (str) – HTTP method to use

  • app_endpoint (str) – relative url, beginning with a forward slash, to an HTTP endpoint served by the Hypersciecne platform. For reference on the available endpoints, check out https://docs.hyperscience.com/

Optional Parameters:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • headers (Optional[Dict[str, str]]) – HTTP headers to use

  • params (Optional[Dict[str, Any]]) – key-value pair used as query parameters in the request

  • payload (Optional[Dict[str, str]]) – key-value pair to be used as x-www-form-urlencoded data. Mutually exclusive with json

  • json (Optional[Dict[str, Any]]) – key-value pair to be used as json data. Mutually exclusive with payload

  • handled_error_codes (Optional[List[Union[int, str]]]) – specify HTTP error codes that will not fail the task, allowing for wildcards with the ‘x’ char. Example: [400, “5xx”]

Example usage:

hs_rest_block = HyperscienceRestApiBlock(
    reference_name='http_request',
    app_endpoint='/api/healthcheck',
    method='GET',
)
class flows_sdk.implementations.idp_v38.additional_blocks.DatabaseAccessBlock(reference_name=None, db_type='', database='', host='', username='', password='', query='', port=None, options=None, timeout=None, query_params=None, title='DB Block', description='DB access block')

Bases: Block

Database Block allows user to access and perform queries on the specified database Mandatory parameters:

Parameters:
  • db_type (str) – database type (mssql, oracle, postgres)

  • host (str) – URL/IP address of the server database is hosted on

  • database (str) – database name

  • username (str) – database username

  • password (str) – database password

  • query (str) – parameterized query

Optional parameters:

Parameters:
  • reference_name (Optional[str]) – unique identifier within a flow, a unique one will be generated when not provided

  • port (Optional[int]) – database port number

  • options (Optional[Dict[Any, Any]]) – dictionary of additional connection string options

  • query_params (Optional[Dict[Any, Any]]) – dictionary of values for query placeholders

  • timeout (Optional[int]) – timeout in seconds (mssql and postgres only)

Example usage:

# minimum viable instantiation, the rest of the fields can later be added in the UI
db_lookup = DatabaseAccessBlock(
    reference_name='db_lookup',
    title='Database Lookup',
    description='Perform database lookup',
)
# note that storing password in plain text form is not recommended
# password can later be entered in the UI or defined as a flow secret.
db_lookup = DatabaseAccessBlock(
    reference_name='db_lookup',
    title='Database Lookup',
    description='Perform database lookup',
    db_type='mssql',
    database='cars',
    host='example.com',
    username='user',
    # we recommend skipping the password field or defining it as a system secret
    # e.g. system_secret('{your_secret_identifier}') (both allow for secret input via. UI)
    password='pw'
    port=1433,
    timeout=200,
    query='SELECT * from CAR',
)
# reference to the output of DBAccessBlock which can be passed as input to the next block
# for more information on output() and output references, please take a look at class Block
# and its output() method
output_ref = db_lookup.output()
# content of output_ref:
{
    "result": [
        {
            "id": 1,
            "brand": "Aston Martin",
            "model": "Vanquish"
        },
        {
            "id": 2,
            "brand": "Jaguar",
            "model": "XE 2018"
        }
    ]
}

Database output for the same query (SELECT * FROM CAR)

../_images/db_access_query_example.png
IDENTIFIER = 'DB_ACCESS'
class flows_sdk.implementations.idp_v38.additional_blocks.LlamaBlock(model_id, prompt, reference_name=None, consent=False, tokenize_only=False, predict_extra_kwargs=None, title='Llama Text Generation', description='Text generation using the Llama 2 model')

Bases: Block

Text generation using the Llama 2 model by Meta.

This block utilizes Llama 2 by Meta. Llama 2 is licensed under the LLAMA 2 Community License (https://ai.meta.com/llama/license/). By using the Llama Block, you acknowledge that you have received a copy of the LLAMA 2 Community License.

Note: This block is still experimental and not ready for production use.

Mandatory parameters:

Parameters: