How to batch create a searchable pdf using azure document intelligence python API

Andrew Richardson (W) 0

Hi All,

I have stored some PDFs in Azure Blob storage and I am trying to batch OCR these documents while also creating a searchable PDF. Unfortunately at the moment I'm getting a file with the following extension "filename.pdf.ocr.json" which wont open even If I rename it to the correct file extension. I assume I'm using the wrong python setup for the function "document_intelligence_client.begin_analyze_batch_documents()", the variable I have set for this are below:

Any suggestions on how to batch create a searchable pdf using azure document intelligence python API would be greatly appreciated

    poller = document_intelligence_client.begin_analyze_batch_documents(
        model_id="prebuilt-read",
        body=request,
        output=["pdf"],
    )

Vikram Singh 980 Reputation points Microsoft Employee

2025-01-31T12:51:12.24+00:00

Hi Andrew Richardson (W),

Thank you for reaching out to Microsoft Q&A, and apologies for the inconvenience.

It seems you're using the Azure Document Intelligence API to batch OCR PDFs, but the output isn't in the expected searchable PDF format. To resolve this, please ensure you're using the begin_analyze_batch_documents method with the correct input and output settings. Specifically, set the contentType to application/pdf and the output parameter to ["pdf"]. Additionally, ensure the output_content_format is set to "pdf" to generate a searchable PDF.

Reference: Double-check the Azure documentation for the exact syntax to return PDF output: Azure Document Intelligence API.

Let me know if you're still facing issues.
Andrew Richardson (W) 0 Reputation points

2025-01-31T13:52:49.8233333+00:00
Hi Vikram,

Thanks for this info , I tried as you recommended but I seem to be getting an error "azure.core.exceptions.HttpResponseError: (UnsupportedMediaType) Request content type is not supported.

Code: UnsupportedMediaType"

This is the code I used, looked at the documentation but didnt find anything to further resolve this:

poller = document_intelligence_client.begin_analyze_batch_documents( model_id="prebuilt-read", body=request, output=["pdf"], content_type="application/pdf", output_content_format="pdf" )
Vikram Singh 980 Reputation points Microsoft Employee

2025-02-01T05:35:42.1066667+00:00
Hi Andrew Richardson (W),

Thanks for trying the suggestion. The error (UnsupportedMediaType) Request content type is not supported indicates an issue with the content_type parameter. The begin_analyze_batch_documents method does not require content_type="application/pdf". Instead, ensure the request body is correctly formatted and output_content_format is set to "pdf".

Can you try the below code:

# Define the request body request = { "azureBlobSource": { "containerUrl": "https://<your-storage-account>.blob.core.windows.net/<your-container>" }, "outputContentFormat": "pdf" } # Start the batch analysis poller = document_intelligence_client.begin_analyze_batch_documents( model_id="prebuilt-read", body=request, output=["pdf"] )

Let me know if the issue persists!
Vikram Singh 980 Reputation points Microsoft Employee

2025-02-04T04:21:38.4633333+00:00

Hi Andrew Richardson (W),

Greetings.

Just following up to check if my suggestion helped. Please do not forget to "Accept the answer” and “up-vote” wherever the information provided helps you, this can be beneficial to other community members.

Thank you
Andrew Richardson (W) 0 Reputation points

2025-02-04T16:48:01.4466667+00:00
Hi Vikram, I changed the body to suit the format you suggested but it still only provided a json output. I am following this code sample for batching documents https://github.com/Azure/azure-sdk-for-python/blob/azure-ai-documentintelligence_1.0.0/sdk/documentintelligence/azure-ai-documentintelligence/samples/sample_analyze_batch_documents.py

This example uses a different type of request body to the one you suggested utilising "AnalyzeBatchDocumentsRequest"

request = AnalyzeBatchDocumentsRequest( result_container_url=result_container_sas_url, azure_blob_source=AzureBlobContentSource( container_url=batch_training_data_container_sas_url, ), )

1 answer

Vikram Singh 980 Reputation points Microsoft Employee

2025-02-05T07:33:41.6766667+00:00
Hello Andrew Richardson (W),

Thanks for sharing the code sample. Since you're following the official sample and using AnalyzeBatchDocumentsRequest, the JSON output suggests that the API isn't generating a searchable PDF. To resolve this, can you try out the below to ensure that the request includes output_content_format="pdf" so that the output is returned as a searchable PDF instead of JSON.

request = AnalyzeBatchDocumentsRequest( result_container_url=result_container_sas_url, azure_blob_source=AzureBlobContentSource( container_url=batch_training_data_container_sas_url ), output_content_format="pdf" # Ensure this is included )

Do let me know if you are still the facing the issue.

Thanks.
Please sign in to rate this answer.

0 comments No comments
Sign in to comment

Use comments to ask for clarification, additional information, or improvements to the question.

Share via

How to batch create a searchable pdf using azure document intelligence python API

1 answer

Your answer