Textract / Client / detect_document_text

detect_document_text#

Textract.Client.detect_document_text(**kwargs)#

Detects text in the input document. Amazon Textract can detect lines of text and the words that make up a line of text. The input document must be in one of the following image formats: JPEG, PNG, PDF, or TIFF. DetectDocumentText returns the detected text in an array of Block objects.

Each document page has as an associated Block of type PAGE. Each PAGE Block object is the parent of LINE Block objects that represent the lines of detected text on a page. A LINE Block object is a parent for each word that makes up the line. Words are represented by Block objects of type WORD.

DetectDocumentText is a synchronous operation. To analyze documents asynchronously, use StartDocumentTextDetection.

For more information, see Document Text Detection.

See also: AWS API Documentation

Request Syntax

response = client.detect_document_text(
    Document={
        'Bytes': b'bytes',
        'S3Object': {
            'Bucket': 'string',
            'Name': 'string',
            'Version': 'string'
        }
    }
)
Parameters:

Document (dict) –

[REQUIRED]

The input document as base64-encoded bytes or an Amazon S3 object. If you use the AWS CLI to call Amazon Textract operations, you can’t pass image bytes. The document must be an image in JPEG or PNG format.

If you’re using an AWS SDK to call Amazon Textract, you might not need to base64-encode image bytes that are passed using the Bytes field.

  • Bytes (bytes) –

    A blob of base64-encoded document bytes. The maximum size of a document that’s provided in a blob of bytes is 5 MB. The document bytes must be in PNG or JPEG format.

    If you’re using an AWS SDK to call Amazon Textract, you might not need to base64-encode image bytes passed using the Bytes field.

  • S3Object (dict) –

    Identifies an S3 object as the document source. The maximum size of a document that’s stored in an S3 bucket is 5 MB.

    • Bucket (string) –

      The name of the S3 bucket. Note that the # character is not valid in the file name.

    • Name (string) –

      The file name of the input document. Synchronous operations can use image files that are in JPEG or PNG format. Asynchronous operations also support PDF and TIFF format files.

    • Version (string) –

      If the bucket has versioning enabled, you can specify the object version.

Return type:

dict

Returns:

Response Syntax

{
    'DocumentMetadata': {
        'Pages': 123
    },
    'Blocks': [
        {
            'BlockType': 'KEY_VALUE_SET'|'PAGE'|'LINE'|'WORD'|'TABLE'|'CELL'|'SELECTION_ELEMENT'|'MERGED_CELL'|'TITLE'|'QUERY'|'QUERY_RESULT'|'SIGNATURE'|'TABLE_TITLE'|'TABLE_FOOTER'|'LAYOUT_TEXT'|'LAYOUT_TITLE'|'LAYOUT_HEADER'|'LAYOUT_FOOTER'|'LAYOUT_SECTION_HEADER'|'LAYOUT_PAGE_NUMBER'|'LAYOUT_LIST'|'LAYOUT_FIGURE'|'LAYOUT_TABLE'|'LAYOUT_KEY_VALUE',
            'Confidence': ...,
            'Text': 'string',
            'TextType': 'HANDWRITING'|'PRINTED',
            'RowIndex': 123,
            'ColumnIndex': 123,
            'RowSpan': 123,
            'ColumnSpan': 123,
            'Geometry': {
                'BoundingBox': {
                    'Width': ...,
                    'Height': ...,
                    'Left': ...,
                    'Top': ...
                },
                'Polygon': [
                    {
                        'X': ...,
                        'Y': ...
                    },
                ]
            },
            'Id': 'string',
            'Relationships': [
                {
                    'Type': 'VALUE'|'CHILD'|'COMPLEX_FEATURES'|'MERGED_CELL'|'TITLE'|'ANSWER'|'TABLE'|'TABLE_TITLE'|'TABLE_FOOTER',
                    'Ids': [
                        'string',
                    ]
                },
            ],
            'EntityTypes': [
                'KEY'|'VALUE'|'COLUMN_HEADER'|'TABLE_TITLE'|'TABLE_FOOTER'|'TABLE_SECTION_TITLE'|'TABLE_SUMMARY'|'STRUCTURED_TABLE'|'SEMI_STRUCTURED_TABLE',
            ],
            'SelectionStatus': 'SELECTED'|'NOT_SELECTED',
            'Page': 123,
            'Query': {
                'Text': 'string',
                'Alias': 'string',
                'Pages': [
                    'string',
                ]
            }
        },
    ],
    'DetectDocumentTextModelVersion': 'string'
}

Response Structure

  • (dict) –

    • DocumentMetadata (dict) –

      Metadata about the document. It contains the number of pages that are detected in the document.

      • Pages (integer) –

        The number of pages that are detected in the document.

    • Blocks (list) –

      An array of Block objects that contain the text that’s detected in the document.

      • (dict) –

        A Block represents items that are recognized in a document within a group of pixels close to each other. The information returned in a Block object depends on the type of operation. In text detection for documents (for example DetectDocumentText), you get information about the detected words and lines of text. In text analysis (for example AnalyzeDocument), you can also get information about the fields, tables, and selection elements that are detected in the document.

        An array of Block objects is returned by both synchronous and asynchronous operations. In synchronous operations, such as DetectDocumentText, the array of Block objects is the entire set of results. In asynchronous operations, such as GetDocumentAnalysis, the array is returned over one or more responses.

        For more information, see How Amazon Textract Works.

        • BlockType (string) –

          The type of text item that’s recognized. In operations for text detection, the following types are returned:

          • PAGE - Contains a list of the LINE Block objects that are detected on a document page.

          • WORD - A word detected on a document page. A word is one or more ISO basic Latin script characters that aren’t separated by spaces.

          • LINE - A string of tab-delimited, contiguous words that are detected on a document page.

          In text analysis operations, the following types are returned:

          • PAGE - Contains a list of child Block objects that are detected on a document page.

          • KEY_VALUE_SET - Stores the KEY and VALUE Block objects for linked text that’s detected on a document page. Use the EntityType field to determine if a KEY_VALUE_SET object is a KEY Block object or a VALUE Block object.

          • WORD - A word that’s detected on a document page. A word is one or more ISO basic Latin script characters that aren’t separated by spaces.

          • LINE - A string of tab-delimited, contiguous words that are detected on a document page.

          • TABLE - A table that’s detected on a document page. A table is grid-based information with two or more rows or columns, with a cell span of one row and one column each.

          • TABLE_TITLE - The title of a table. A title is typically a line of text above or below a table, or embedded as the first row of a table.

          • TABLE_FOOTER - The footer associated with a table. A footer is typically a line or lines of text below a table or embedded as the last row of a table.

          • CELL - A cell within a detected table. The cell is the parent of the block that contains the text in the cell.

          • MERGED_CELL - A cell in a table whose content spans more than one row or column. The Relationships array for this cell contain data from individual cells.

          • SELECTION_ELEMENT - A selection element such as an option button (radio button) or a check box that’s detected on a document page. Use the value of SelectionStatus to determine the status of the selection element.

          • SIGNATURE - The location and confidence score of a signature detected on a document page. Can be returned as part of a Key-Value pair or a detected cell.

          • QUERY - A question asked during the call of AnalyzeDocument. Contains an alias and an ID that attaches it to its answer.

          • QUERY_RESULT - A response to a question asked during the call of analyze document. Comes with an alias and ID for ease of locating in a response. Also contains location and confidence score.

          The following BlockTypes are only returned for Amazon Textract Layout.

          • LAYOUT_TITLE - The main title of the document.

          • LAYOUT_HEADER - Text located in the top margin of the document.

          • LAYOUT_FOOTER - Text located in the bottom margin of the document.

          • LAYOUT_SECTION_HEADER - The titles of sections within a document.

          • LAYOUT_PAGE_NUMBER - The page number of the documents.

          • LAYOUT_LIST - Any information grouped together in list form.

          • LAYOUT_FIGURE - Indicates the location of an image in a document.

          • LAYOUT_TABLE - Indicates the location of a table in the document.

          • LAYOUT_KEY_VALUE - Indicates the location of form key-values in a document.

          • LAYOUT_TEXT - Text that is present typically as a part of paragraphs in documents.

        • Confidence (float) –

          The confidence score that Amazon Textract has in the accuracy of the recognized text and the accuracy of the geometry points around the recognized text.

        • Text (string) –

          The word or line of text that’s recognized by Amazon Textract.

        • TextType (string) –

          The kind of text that Amazon Textract has detected. Can check for handwritten text and printed text.

        • RowIndex (integer) –

          The row in which a table cell is located. The first row position is 1. RowIndex isn’t returned by DetectDocumentText and GetDocumentTextDetection.

        • ColumnIndex (integer) –

          The column in which a table cell appears. The first column position is 1. ColumnIndex isn’t returned by DetectDocumentText and GetDocumentTextDetection.

        • RowSpan (integer) –

          The number of rows that a table cell spans. RowSpan isn’t returned by DetectDocumentText and GetDocumentTextDetection.

        • ColumnSpan (integer) –

          The number of columns that a table cell spans. ColumnSpan isn’t returned by DetectDocumentText and GetDocumentTextDetection.

        • Geometry (dict) –

          The location of the recognized text on the image. It includes an axis-aligned, coarse bounding box that surrounds the text, and a finer-grain polygon for more accurate spatial information.

          • BoundingBox (dict) –

            An axis-aligned coarse representation of the location of the recognized item on the document page.

            • Width (float) –

              The width of the bounding box as a ratio of the overall document page width.

            • Height (float) –

              The height of the bounding box as a ratio of the overall document page height.

            • Left (float) –

              The left coordinate of the bounding box as a ratio of overall document page width.

            • Top (float) –

              The top coordinate of the bounding box as a ratio of overall document page height.

          • Polygon (list) –

            Within the bounding box, a fine-grained polygon around the recognized item.

            • (dict) –

              The X and Y coordinates of a point on a document page. The X and Y values that are returned are ratios of the overall document page size. For example, if the input document is 700 x 200 and the operation returns X=0.5 and Y=0.25, then the point is at the (350,50) pixel coordinate on the document page.

              An array of Point objects, Polygon, is returned by DetectDocumentText. Polygon represents a fine-grained polygon around detected text. For more information, see Geometry in the Amazon Textract Developer Guide.

              • X (float) –

                The value of the X coordinate for a point on a Polygon.

              • Y (float) –

                The value of the Y coordinate for a point on a Polygon.

        • Id (string) –

          The identifier for the recognized text. The identifier is only unique for a single operation.

        • Relationships (list) –

          A list of relationship objects that describe how blocks are related to each other. For example, a LINE block object contains a CHILD relationship type with the WORD blocks that make up the line of text. There aren’t Relationship objects in the list for relationships that don’t exist, such as when the current block has no child blocks.

          • (dict) –

            Information about how blocks are related to each other. A Block object contains 0 or more Relation objects in a list, Relationships. For more information, see Block.

            The Type element provides the type of the relationship for all blocks in the IDs array.

            • Type (string) –

              The type of relationship between the blocks in the IDs array and the current block. The following list describes the relationship types that can be returned.

              • VALUE - A list that contains the ID of the VALUE block that’s associated with the KEY of a key-value pair.

              • CHILD - A list of IDs that identify blocks found within the current block object. For example, WORD blocks have a CHILD relationship to the LINE block type.

              • MERGED_CELL - A list of IDs that identify each of the MERGED_CELL block types in a table.

              • ANSWER - A list that contains the ID of the QUERY_RESULT block that’s associated with the corresponding QUERY block.

              • TABLE - A list of IDs that identify associated TABLE block types.

              • TABLE_TITLE - A list that contains the ID for the TABLE_TITLE block type in a table.

              • TABLE_FOOTER - A list of IDs that identify the TABLE_FOOTER block types in a table.

            • Ids (list) –

              An array of IDs for related blocks. You can get the type of the relationship from the Type element.

              • (string) –

        • EntityTypes (list) –

          The type of entity.

          The following entity types can be returned by FORMS analysis:

          • KEY - An identifier for a field on the document.

          • VALUE - The field text.

          The following entity types can be returned by TABLES analysis:

          • COLUMN_HEADER - Identifies a cell that is a header of a column.

          • TABLE_TITLE - Identifies a cell that is a title within the table.

          • TABLE_SECTION_TITLE - Identifies a cell that is a title of a section within a table. A section title is a cell that typically spans an entire row above a section.

          • TABLE_FOOTER - Identifies a cell that is a footer of a table.

          • TABLE_SUMMARY - Identifies a summary cell of a table. A summary cell can be a row of a table or an additional, smaller table that contains summary information for another table.

          • STRUCTURED_TABLE - Identifies a table with column headers where the content of each row corresponds to the headers.

          • SEMI_STRUCTURED_TABLE - Identifies a non-structured table.

          EntityTypes isn’t returned by DetectDocumentText and GetDocumentTextDetection.

          • (string) –

        • SelectionStatus (string) –

          The selection status of a selection element, such as an option button or check box.

        • Page (integer) –

          The page on which a block was detected. Page is returned by synchronous and asynchronous operations. Page values greater than 1 are only returned for multipage documents that are in PDF or TIFF format. A scanned image (JPEG/PNG) provided to an asynchronous operation, even if it contains multiple document pages, is considered a single-page document. This means that for scanned images the value of Page is always 1.

        • Query (dict) –

          • Text (string) –

            Question that Amazon Textract will apply to the document. An example would be “What is the customer’s SSN?”

          • Alias (string) –

            Alias attached to the query, for ease of location.

          • Pages (list) –

            Pages is a parameter that the user inputs to specify which pages to apply a query to. The following is a list of rules for using this parameter.

            • If a page is not specified, it is set to ["1"] by default.

            • The following characters are allowed in the parameter’s string: 0 1 2 3 4 5 6 7 8 9 - *. No whitespace is allowed.

            • When using * to indicate all pages, it must be the only element in the list.

            • You can use page intervals, such as [“1-3”, “1-1”, “4-*”]. Where * indicates last page of document.

            • Specified pages must be greater than 0 and less than or equal to the number of pages in the document.

            • (string) –

    • DetectDocumentTextModelVersion (string) –

Exceptions