The Analyze Form operation extracts Key/Value pairs for the given document.
API Reference – https://bit.ly/2NvTaCg
Pre-requisites
- Form Recognizer Cognitive Service provisioned on Azure (Submit for Form Recognizer Access Request)
- Postman App
- Run the Train Model endpoint with the sample documents and get the “Model Id” which is required for using with the “Analyze Form” endpoint.
Image Requirements
- Image needs to be PDF, JPEG, PNG
Analyze Form Operation
- Launch Postman
- Append your endpoint URL from the pre-requisites step with /formrecognizer/v1.0-preview/custom/models/{id}/analyze[?keys]
- Where id – Model Id retrieved from Train Model endpoint.
- Provide the subscription key and content-type in Headers tab as shown below.
- In the Body tab, choose form-data and type the key as form and beside it click the down arrow to choose file as shown below.
- Select the “Invoice_6.pdf” file from the Train folder downloaded from from here.
- Click the Send button and you will get the output of the pdf form in Key/Value pairs as shown below.
{
"status": "success",
"pages": [
{
"number": 1,
"height": 792,
"width": 612,
"clusterId": 0,
"keyValuePairs": [
{
"key": [
{
"text": "Address:",
"boundingBox": [
57.3,
683.0,
100.5,
683.0,
100.5,
673.7,
57.3,
673.7
]
}
],
"value": [
{
"text": "14564 Main St.",
"boundingBox": [
57.3,
672.3,
124.3,
672.3,
124.3,
658.7,
57.3,
658.7
],
"confidence": 0.86
},
{
"text": "Saratoga, CA 94588",
"boundingBox": [
57.3,
657.7,
147.3,
657.7,
147.3,
645.0,
57.3,
645.0
],
"confidence": 0.86
}
]
},
{
"key": [
{
"text": "Invoice For:",
"boundingBox": [
316.1,
683.1,
368.2,
683.1,
368.2,
673.7,
316.1,
673.7
]
}
],
"value": [
{
"text": "First Up Consultants",
"boundingBox": [
374.0,
687.9,
472.0,
687.9,
472.0,
673.7,
374.0,
673.7
],
"confidence": 1.0
},
{
"text": "1234 King St",
"boundingBox": [
373.9,
673.0,
433.7,
673.0,
433.7,
658.7,
373.9,
658.7
],
"confidence": 1.0
},
{
"text": "Redmond, WA 97624",
"boundingBox": [
373.8,
659.2,
472.3,
659.2,
472.3,
645.0,
373.8,
645.0
],
"confidence": 1.0
}
]
},
{
"key": [
{
"text": "Page",
"boundingBox": [
448.6,
96.1,
471.9,
96.1,
471.9,
86.7,
448.6,
86.7
]
}
],
"value": [
{
"text": "1 of",
"boundingBox": [
491.9,
96.1,
508.4,
96.1,
508.4,
86.7,
491.9,
86.7
],
"confidence": 0.86
}
]
},
{
"key": [
{
"text": "__Tokens__",
"boundingBox": [
0.0,
0.0,
0.0,
0.0,
0.0,
0.0,
0.0,
0.0
]
}
],
"value": [
{
"text": "Contoso Suites",
"boundingBox": [
38.0,
716.2,
188.0,
716.2,
188.0,
686.4,
38.0,
686.4
],
"confidence": 0.86
},
{
"text": "1",
"boundingBox": [
532.7,
96.1,
538.2,
96.1,
538.2,
86.7,
532.7,
86.7
],
"confidence": 0.86
}
]
}
],
"tables": [
{
"id": "table_0",
"columns": [
{
"header": [
{
"text": "Invoice Number",
"boundingBox": [
38.5,
585.2,
113.4,
585.2,
113.4,
575.8,
38.5,
575.8
]
}
],
"entries": [
[
{
"text": "7689302",
"boundingBox": [
38.5,
549.8,
77.3,
549.8,
77.3,
536.2,
38.5,
536.2
],
"confidence": 1.0
}
]
]
},
{
"header": [
{
"text": "Invoice Date",
"boundingBox": [
139.7,
585.2,
198.5,
585.2,
198.5,
575.8,
139.7,
575.8
]
}
],
"entries": [
[
{
"text": "3/09/2015",
"boundingBox": [
139.7,
548.1,
184.0,
548.1,
184.0,
536.2,
139.7,
536.2
],
"confidence": 1.0
}
]
]
},
{
"header": [
{
"text": "Invoice Due Date",
"boundingBox": [
240.5,
585.2,
321.0,
585.2,
321.0,
575.8,
240.5,
575.8
]
}
],
"entries": [
[
{
"text": "6/29/2016",
"boundingBox": [
240.5,
549.0,
284.8,
549.0,
284.8,
536.2,
240.5,
536.2
],
"confidence": 1.0
}
]
]
},
{
"header": [
{
"text": "Charges",
"boundingBox": [
341.3,
585.2,
381.2,
585.2,
381.2,
575.8,
341.3,
575.8
]
}
],
"entries": [
[
{
"text": "$22,123.24",
"boundingBox": [
380.6,
548.5,
430.5,
548.5,
430.5,
536.2,
380.6,
536.2
],
"confidence": 1.0
}
]
]
},
{
"header": [
{
"text": "VAT ID",
"boundingBox": [
442.1,
590.0,
474.8,
590.0,
474.8,
575.8,
442.1,
575.8
]
}
],
"entries": [
[
{
"text": "QR",
"boundingBox": [
447.7,
549.8,
462.6,
549.8,
462.6,
536.2,
447.7,
536.2
],
"confidence": 1.0
}
]
]
}
]
}
]
}
],
"errors": []
}
- If you would like to extract only particular set of keys from the form then you can include the additional “keys” parameter as part of the request to get the values related to that key. You will see the output as shown below. Note: the screenshot doesn’t include full output.
