GetDocumentContent Output Schema

When you use the GetDocumentContent API with outputFormat set to EXTRACTED, the response returns extracted text content in JSON format. The output schema is presented in JSON format:

{
   // always V1 for now
   schemaVersionId: string;
   
   // always JSON for now
   outputFormat: string;
   
   // content for plain-text documents
   plainTextDocumentContent: string;
   
   // content for non-plaintext documents such as PDF, DOCX, PPTX, Audio, Video
   nonPlainTextDocumentContent: List<ExtractedDocumentBodyElement>;
}

The schema for non-plaintext documents includes the ExtractedDocumentBodyElement which includes:

{
   text: string;
   
  // Allowed values: TEXT, ARTICLE, SECTION, DIV, IMAGE_DESCRIPTION, CODE,
  // TABLE, LIST, URL, HEADER, FOOTER, FORM, MENU, AUDIO, VIDEO
   elementType: string;
   
   horizontalHeaderIndex: integer;
   verticalHeaderIndex: integer;
   htmlDocumentTitle: string;
   sectionTitle: string;
   sectionBody: string;
   tableCaption: string;
   tableFooter: string;
   tableRowHeaders: List<List<string>>;
   tableColumnHeaders: List<List<string>>;
   tableRows: List<List<string>>;
   tableRowsCount: integer;
   tableColumnsCount: integer;
   tableId: string;
   
   tokens: List<struct>;
   {
        value: string;
        startOffsets: integer;
        endOffsets: integer;
   }
   
   tableType: string;
   tableSummary: string;
   
   columnInfoList: List<struct>;
   {
        columnName: string;
        columnSummary: string;
        columnType: string;
        columnRepresentativeValues: List<string>
   }
   
   // Audio/Video specific fields below
   overallSummary: string;
   
   audioSummaryList: List<struct>;
   {
        summaryText: string;
        startTimeMilliseconds: string;
        endTimeMilliseconds: string;
   }
   
   videoSummaryList: List<struct>;
   {
        summaryText: string;
        startTimeMilliseconds: string;
        endTimeMilliseconds: string;
   }
   
   audioTranscriptList: List<struct>;
   {
        transcriptText: string;
        startTimeMilliseconds: string;
        endTimeMilliseconds: string;
   }
   
   videoTranscriptList: List<struct>;
   {
        transcriptText: string;
        startTimeMilliseconds: string;
        endTimeMilliseconds: string;
   }   
}