AnalyzeDocument Processor

Part of the AWS Textract processor family

The AnalyzeDocument processor will search a document for text, forms, and tables. You can tell the processor to search for specific things by modifying the Feature Types property.

Properties

All of our Textract processors also include these common properties.

Properties whose names are in bold and italic are required.

• Feature Types- a dropdown list that tells controls what the API is looking for. It can be set to one of the following:

  • tables - the API will search for text, as well as tables

  • forms - the API will search for text, as well as forms, which are areas where a user would be expected to input information

  • tables-and-forms - the API will search for text, tables, and forms

Data Output

If the Destination property is set to flowfile-attribute, then the output of this processor will be routed to the FlowFile's ocr.AnalyzedDocument attribute, which will be created if it isn't present.

Output Structure

Field Name	Data Type	Description
blocks	array of Block	The list of blocks returned from the API

Relevant Data Structures

Block

Field Name	Data Type	Description
text	string	The text in this block
confidence	float	How confident the API is in its response
id	string	The UUID pertaining to this block. Can be used to cross-reference relationships between blocks
page	int	The page of the document in which this block resides
columnIndex	int
columnSpan	int
rowIndex	int
rowSpan	int
type	string (BlockType)	The kind of block
geometry	Geometry	The position and size of the block
relationship	array of Relationship	The relationships this block has to others

Geometry

Field Name	Data Type	Description
x	float	The X position of the block on the page
y	float	The Y position of the block on the page
width	float	The width of the block
height	float	The height of the block

Relationship

Field Name	Data Type	Description
type	string (equal to eitherVALUE or CHILD)	The kind of relationship
ids	array of strings	The list of block UUIDs that are connected via this relationship

Example Output

{
	"output": {
		"blocks": [
			{
				"relationships": [],
				"confidence": 99.35694,
				"geometry": {
					"width": 0.2716,
					"height": 0.02702,
					"x": 0.36377,
					"y": 0.07574
				},
				"text": "Spirit Game Script",
				"id": "e63c08cf-0bef-4c6d-ac04-0e250e254229",
				"page": 1,
				"type": "LINE",
			},
			{
				"relationships": [
					{
						"ids": [
							"a5ade6e3-a368-49fe-b338-7a8a4fbe9058",
							"10e1b516-02c3-4eaf-a216-65efc12af6a7",
							"2f44eedd-651b-4ac9-b42a-744bd0dfcbe1",
							"fbb0d176-f9fb-4c48-8a5a-4d09252f6d17"
						],
            "type":"CHILD"
					}
				],
				"confidence": 99.51574,
				"geometry": {
					"width": 0.37119,
					"height": 0.18922,
					"x": 0.321415,
					"y": 0.251234,
				},
				"text": "Courageous, pleading and clear.",
				"id": "7396c015-e0aa-49e5-8d05-3cf9a7430539",
				"page": 1,
				"type": "LINE"
			},
			// ... plus potentially many more entries!
		]
	}
}