Azure AI Document IntelligenceでPDFからMarkdownを生成すーる

はじめに

Azure AI Document IntelligenceでPDFからMarkdownを生成していきます

開発環境

• Windows 10 PC
• Python 3.11
• Azure

導入

こちらを参考にします

URL指定

PDFファイルをURLで指定して、Document Intelligenceに投げて、結果のMarkdownを保存します

sample_analyze_documents_output_in_markdown.py

import os


def analyze_documents_output_in_markdown():
    # [START analyze_documents_output_in_markdown]
    from azure.core.credentials import AzureKeyCredential
    from azure.ai.documentintelligence import DocumentIntelligenceClient
    from azure.ai.documentintelligence.models import (
        AnalyzeDocumentRequest,
        DocumentContentFormat,
        AnalyzeResult,
    )

    endpoint = os.environ["DOCUMENTINTELLIGENCE_ENDPOINT"]
    key = os.environ["DOCUMENTINTELLIGENCE_API_KEY"]
    url = "https://raw.githubusercontent.com/Azure/azure-sdk-for-python/main/sdk/documentintelligence/azure-ai-documentintelligence/samples/sample_forms/forms/Invoice_1.pdf"

    document_intelligence_client = DocumentIntelligenceClient(
        endpoint=endpoint, credential=AzureKeyCredential(key)
    )
    poller = document_intelligence_client.begin_analyze_document(
        "prebuilt-layout",
        AnalyzeDocumentRequest(url_source=url),
        output_content_format=DocumentContentFormat.MARKDOWN,
    )
    result: AnalyzeResult = poller.result()

    # Save markdown content to file
    output_file = "Invoice_1.md"
    with open(output_file, "w", encoding="utf-8") as f:
        f.write(result.content)

    print(f"Content has been saved to {output_file}")
    print(f"Here's the full content in format {result.content_format}:\n")
    print(result.content)
    # [END analyze_documents_output_in_markdown]


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

    try:
        load_dotenv(find_dotenv())
        analyze_documents_output_in_markdown()
    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

.envを作成してください

.env

DOCUMENTINTELLIGENCE_ENDPOINT=<INSERT-YOUR-ENDPOINT>
DOCUMENTINTELLIGENCE_API_KEY=<INSERT-YOUR-API-KEY>

実行結果は下記のようになります

Contoso

Address:

1 Redmond way Suite

6000 Redmond, WA

99243

Invoice For: Microsoft

1020 Enterprise Way

Sunnayvale, CA 87659


<table>
<tr>
<th>Invoice Number</th>
<th>Invoice Date</th>
<th>Invoice Due Date</th>
<th>Charges</th>
<th>VAT ID</th>
</tr>
<tr>
<td>34278587</td>
<td>6/18/2017</td>
<td>6/24/2017</td>
<td>$56,651.49</td>
<td>PT</td>
</tr>
</table>

ファイル指定

ローカルのPDFファイルをDocument Intelligenceに投げて、結果をMarkdownファイルで保存します

sample_analyze_documents_output_in_markdown_from_bytes.py

import os


def analyze_documents_output_in_markdown():
    # [START analyze_documents_output_in_markdown]
    from azure.core.credentials import AzureKeyCredential
    from azure.ai.documentintelligence import DocumentIntelligenceClient
    from azure.ai.documentintelligence.models import (
        AnalyzeDocumentRequest,
        DocumentContentFormat,
        AnalyzeResult,
    )

    endpoint = os.environ["DOCUMENTINTELLIGENCE_ENDPOINT"]
    key = os.environ["DOCUMENTINTELLIGENCE_API_KEY"]
    filename = "Invoice_1.pdf"

    document_intelligence_client = DocumentIntelligenceClient(
        endpoint=endpoint, credential=AzureKeyCredential(key)
    )
    poller = document_intelligence_client.begin_analyze_document(
        "prebuilt-layout",
        AnalyzeDocumentRequest(bytes_source=open(filename, "rb").read()),
        output_content_format=DocumentContentFormat.MARKDOWN,
    )
    result: AnalyzeResult = poller.result()

    # Save markdown content to file
    output_file = os.path.splitext(filename)[0] + ".md"
    with open(output_file, "w", encoding="utf-8") as f:
        f.write(result.content)

    print(f"Content has been saved to {output_file}")
    print(f"Here's the full content in format {result.content_format}:\n")
    print(result.content)
    # [END analyze_documents_output_in_markdown]


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

    try:
        load_dotenv(find_dotenv())
        analyze_documents_output_in_markdown()
    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

結果はこのようになります

Contoso

Address:

1 Redmond way Suite

6000 Redmond, WA

99243

Invoice For: Microsoft

1020 Enterprise Way

Sunnayvale, CA 87659


<table>
<tr>
<th>Invoice Number</th>
<th>Invoice Date</th>
<th>Invoice Due Date</th>
<th>Charges</th>
<th>VAT ID</th>
</tr>
<tr>
<td>34278587</td>
<td>6/18/2017</td>
<td>6/24/2017</td>
<td>$56,651.49</td>
<td>PT</td>
</tr>
</table>

お疲れ様でした