AWS Comprehend - Classify document action

The Classify document action analyzes a document in real-time using a custom classification model.

Input

Input fields	Description
Endpoint ARN	Select or enter the Amazon Resource Number (ARN) of the endpoint.
File content	Use this parameter to input a text, PDF, Word, or image file. When you classify a document using a custom model, you can use this parameter to input an Amazon Textract, DetectDocumentText, or AnalyzeDocument output file. You can't use this field with the Text field.
Text	Enter the document text to analyze. You can't use this field with the File content field.
Document reader configuration	Expand to configure document reading settings.
Document read action	Select which Textract API to use: Textract detect document text (DetectDocumentText) or Textract analyze document (AnalyzeDocument).
Document read mode	Select Service default to use AWS Comprehend defaults for PDFs or Force document read action to always use the specified Textract API.
Feature types	Select Amazon Textract features to apply. Required when using Textract analyze document.

Output

Output fields	Description
Classes	Includes details about classes including: Name: The name of the class. Page: The page number in the input document. Score: The confidence score that Amazon Comprehend has this class correctly attributed.
Document metadata	Includes document metadata including: Extracted characters: List of pages in the document, with the number of characters extracted from each page. Pages: Number of pages in the document.
Document type	Details about the document type including: Page: The page number in a document. Type: The document type.
Errors	Error code: The code for the cause of the error. Error message: Text message explaining the reason for the error. Page: Page number where the error occurred.
Labels	Name: The name of the label. Page: The page number where the label occurs. This field is present only if you configure the File content input field. Score: The confidence score that AWS Comprehend has the label correctly attributed.
Warnings	Page: Page number in the input document. Warn code: The type of warning. Warn message: Text message associated with the warning.