getmarkdown.py

def get_markdown_format():
    from azure.core.credentials import AzureKeyCredential
    from azure.ai.documentintelligence import DocumentIntelligenceClient
    from azure.ai.documentintelligence.models import AnalyzeResult, ContentFormat
    import os

    # get parameters from environment variables
    endpoint = os.environ["AZURE_DOCUMENT_INTELLIGENCE_ENDPOINT"]
    key = os.environ["AZURE_DOCUMENT_INTELLIGENCE_KEY"]
       
    document_intelligence_client = DocumentIntelligenceClient(endpoint=endpoint, credential=AzureKeyCredential(key))
    path_to_input_document = "input3.pdf"
    path_to_output_document = "output-layout2.md"

    # Check if file exists and delete if it does
    if os.path.exists(path_to_output_document):
        os.remove(path_to_output_document)

    with open(path_to_input_document, "rb") as f:    
        poller = document_intelligence_client.begin_analyze_document(
            "prebuilt-layout",
            analyze_request=f,
            content_type="application/octet-stream",
            output_content_format=ContentFormat.MARKDOWN
    )
    result: AnalyzeResult = poller.result()
    #save result in markdown file
    with open(path_to_output_document, "a") as f:
        f.write(result.content)

        

if __name__ == "__main__":
    from azure.core.exceptions import HttpResponseError
    from dotenv import find_dotenv, load_dotenv

    try:
        load_dotenv(find_dotenv())
        #analyze_layout()
        get_markdown_format()
    except HttpResponseError as error:
        # Examples of how to check an HttpResponseError
        # Check by error code:
        if error.error is not None:
            if error.error.code == "InvalidImage":
                print(f"Received an invalid image error: {error.error}")
            if error.error.code == "InvalidRequest":
                print(f"Received an invalid request error: {error.error}")
            # Raise the error again after printing it
            raise
        # If the inner error is None and then it is possible to check the message to get more information:
        if "Invalid request".casefold() in error.message.casefold():
            print(f"Uh-oh! Seems there was an invalid request: {error}")
        # Raise the error again
        raise

# Next steps:
# Learn more about Layout model: https://aka.ms/di-layout
# Find more sample code: https://aka.ms/doc-intelligence-samples