topic RE: Azure Form Recognizer Analyze Form - PDF in Digital Exchange

Azure Form Recognizer Analyze Form - PDF

TahaSonmez — Fri, 23 Oct 2020 17:24:00 GMT

Hi everyone,

I'm trying to integrate Blue Prism with Microsoft Azure Form Recognizer. I'm using this skill from Digital Exchange.

I have PDFs in machine and I want to analyze them. When I send the request with Base64 string of PDF, it gives an error. I don't know encoding PDF file to base64 is the correct way.

The error:

And how do I convert PDF to Base64;

Any suggestions?

------------------------------
Taha Sonmez
------------------------------

RE: Azure Form Recognizer Analyze Form - PDF

GopalBhaire — Mon, 26 Oct 2020 04:18:00 GMT

Hi,

I think you need to send the pdf data as binary. You can use LoadBinaryFile function in calc stage to do that.

Thanks,

Gopal

------------------------------
Gopal Bhaire
Analyst
Accenture
------------------------------

RE: Azure Form Recognizer Analyze Form - PDF

AminPatel1 — Wed, 28 Oct 2020 11:57:00 GMT

Hi,

There are two ways -

First you have to change Form Recognizer Client Service WebAPI's Analyze Form request to accept single file instead of default template {"source":"[source]"} available. Then you need to convert the pdf data as binary using calculate stage with expression LoadBinaryFile(source-file). OR
In Analyze Form action inputs change ContentType to "application/json" and pass the url for pdf document in source parameter without decorating "".

------------------------------
Amin Patel
Intelligent Automation Developer
Emerson
Pune India
------------------------------

RE: Azure Form Recognizer Analyze Form - PDF

John_L_Lehman — Thu, 28 Jan 2021 11:04:00 GMT

Hi @AminPatel1 and @Taha Sonmez

I am unable to consume this web api with binary data. I changed the body content of request as 'Single File'. And the content-type is application\pdf. I am sending binary data but an error occurs.

Internal : Unexpected error Error during Web API HTTP Request
HTTP Status Code: 400
HTTP Response Content: {"error":{"code":"1002","message":"Analyze request is either invalid or missing required parameters. Refer to the API reference and retry your request."}}

Can you please help me?

------------------------------
John L. Lehman
------------------------------

RE: Azure Form Recognizer Analyze Form - PDF

AminPatel1 — Thu, 28 Jan 2021 12:25:00 GMT

@John_L_Lehman could you please share inputs data screenshot.

------------------------------
Amin Patel
Intelligent Automation Developer
Emerson
Pune India
------------------------------

RE: Azure Form Recognizer Analyze Form - PDF

John_L_Lehman — Thu, 28 Jan 2021 12:28:00 GMT

Hi,

------------------------------
John L. Lehman
------------------------------

RE: Azure Form Recognizer Analyze Form - PDF

shashank.kumar280 — Mon, 01 Feb 2021 15:32:00 GMT

Try setting the content-type to application/octet-stream when sending the Single File in the request. There is also a sample implementation of the same here https://digitalexchange.blueprism.com/dx/entry/3439/solution/microsoft-computer-vision-api-v30-preview . This url is for version 3 of Azure Form Recognizer

------------------------------
Shashank Kumar
DX Integrations Partner Consultant
Blue Prism
Singapore
+6581326707
------------------------------

RE: Azure Form Recognizer Analyze Form - PDF

AminPatel1 — Fri, 05 Feb 2021 11:01:00 GMT

Hello @John_L_Lehman,

PFB parameters datatype required in Analyze Form request -

1.1. Analyze Form

Analyze Form extracts key-value pairs, tables, and semantic values from a given document (resultId) for use with the "Get Analyze Form Result" action. The input document must be of one of the supported content types - 'application/pdf', 'image/jpeg', 'image/png' or 'image/tiff'. Alternatively, use 'application/json' type to specify the Url location of the document to be analyzed.

Note: When configuring source (media type), please note the default value is application/json; you will need to specify the Url location of the document to be analysed. All other media types need to the raw image binary of the file type.

1.1.1. Request

Parameter

Direction

Data Type

Description

modelId

Text

String Format - uuid. Model identifier

includeTextDetails

boolean

Include text lines and element references in the result. Default: false.

ContentType

Text

Media type of the body sent to the API. Default: application/json

source

Text

Request body - .json, .pdf, .jpg, .png or .tiff type file stream.

Change datatype for includeTextDetails from text to boolean in your API request.
Let me know if this works.

------------------------------
Amin Patel
Intelligent Automation Developer
Emerson
Pune India
------------------------------