Skip to main content

Single-page extraction

UiPath.Ephesoft.Activities.SinglePageExtraction

Submits a single-page file for fast OCR processing, classification, and field extraction.

Properties​

Common​

  • DisplayName (String) - The display name of the activity.

Input​

  • BatchClass (String) - The target bucket, predefined on the Transact server, which contains a group of related documents and their constituent fields.
  • FilePath (String) - The full path of the file to be uploaded. Supports only strings and String variables.

Misc​

  • Private (Boolean) - If selected, the values of variables and arguments are no longer logged at Verbose level.

Output​

  • Confidence (Single) - The Lucene score for confidence.

    note

    Lucene scores are not necessarily on a 0-100 scale. Each extracted value contains a Confidence Score and a Confidence Threshold to which it can be compared.

    You can read more about them here.

  • Fields (EphesoftField[ ]) - A collection of labelled fields found within the uploaded document.

Custom Classes​

EphesoftDocument​

  • Confidence (Single) - A Lucene score for confidence of the classification. This should be over the Confidence Threshold.
  • Confidence Threshold (Single) - The value above which the classification can be trusted.
  • DataTables (DataTable) - Any table expected to be found in the document, if one exists.
  • Fields (EphesoftField[ ]) - A list of fields expected to be found in the document.
  • Identifier (Single) - A unique value identifying this document.
  • isConfident (Boolean) - Whether or not the Confidence exceeds the Confidence Threshold.
  • Type (String) - The type of document identified.

EphesoftField​

  • Confidence (Single) - A Lucene score for confidence of the extracted field.
  • IsConfident (Boolean) - Whether or not the Confidence exceeds a given threshold.
  • Name (String) - The name of the field.
  • OCRConfidence (Single) - Confidence that the OCR engine extracted correctly.
  • OCRConfidenceThreshold (Single) - The value above which the extracted field can be trusted.
  • Type (String) - The type of the field.
  • Value (String) - The value read from the targeted field.