Amazon Textract Python代码处理多页PDF遇UnsupportedDocumentException求助

我有一份多页PDF文档，在Amazon Textract网页界面中可正常处理并提取键值对，但使用Python代码调用AnalyzeDocument操作时，返回以下错误：

UnsupportedDocumentException: An error occurred (UnsupportedDocumentException) when calling the AnalyzeDocument operation: Request has unsupported document format

初始代码如下：

response = textract.analyze_document(
    Document={
        "S3Object": {
            "Bucket": bucketname,
            "Name": filename,
        }
    },
    FeatureTypes=["FORMS"],
    HumanLoopConfig={
        "HumanLoopName": uuid.uuid4().hex,
        "FlowDefinitionArn": FLOW_ARN,
        "DataAttributes": {
            "ContentClassifiers": [
                "FreeOfPersonallyIdentifiableInformation",
                "FreeOfAdultContent",
            ]
        },
    },
)
print(json.dumps(response))

return {
    "statusCode": 200,
    "body": json.dumps("Document processed successfully!"),
}

return {"statusCode": 500, "body": json.dumps("Issue processing file!")}

原以为是多页PDF导致无法读取，于是修改代码尝试逐页处理，修改后的代码如下：

# Start document text detection
response = textract.start_document_text_detection(
    DocumentLocation={
        "S3Object": {
            "Bucket": bucketname,
            "Name": filename,
        }
    },
    ClientRequestToken=str(uuid.uuid4())  # Generate a unique client request token
)

# Retrieve the job ID from the response
job_id = response["JobId"]

# Poll for the completion of the job
while True:
    job_status = textract.get_document_text_detection(JobId=job_id)['JobStatus']
    if job_status in ['SUCCEEDED', 'FAILED']:
        break
    time.sleep(5)  # Wait for 5 seconds before checking again

# Get the results of the detection
response = textract.get_document_text_detection(JobId=job_id)

# Process each page of the document
for page_result in response['Blocks']:
    if page_result['BlockType'] == 'PAGE':
        page_number = page_result['Page']
        response = textract.analyze_document(
            Document={
                "S3Object": {
                    "Bucket": bucketname,
                    "Name": filename,
                }
            },
            FeatureTypes=["FORMS"],
            HumanLoopConfig={
                "HumanLoopName": uuid.uuid4().hex,
                "FlowDefinitionArn": FLOW_ARN,
                "DataAttributes": {
                    "ContentClassifiers": [
                        "FreeOfAdultContent",
                    ]
                },
            },
        )
        print(json.dumps(response))

return {
    "statusCode": 200,
    "body": json.dumps("Document processed successfully!"),
}

return {"statusCode": 500, "body": json.dumps("Issue processing file!")}

但修改后仍出现相同的UnsupportedDocumentException错误，恳请提供帮助或解决思路。

解决思路与修复方案 #

• 核心修复：指定多页PDF的处理页码
  Textract网页端会自动处理多页PDF，但AnalyzeDocument是同步API，默认不支持直接处理多页文档，必须显式指定Pages参数。你修改后的代码虽然获取了页码，但后续调用AnalyzeDocument时未传入该参数，导致报错。

  修改AnalyzeDocument调用部分，添加Pages参数：

  response = textract.analyze_document(
      Document={
          "S3Object": {
              "Bucket": bucketname,
              "Name": filename,
          }
      },
      FeatureTypes=["FORMS"],
      HumanLoopConfig={
          "HumanLoopName": uuid.uuid4().hex,
          "FlowDefinitionArn": FLOW_ARN,
          "DataAttributes": {
              "ContentClassifiers": [
                  "FreeOfAdultContent",
              ]
          },
      },
      Pages=[str(page_number)]  # 新增：指定当前要处理的页码
  )
  

• 其他排查方向

  • 检查文档完整性：确认PDF未损坏、未加密（Textract不支持加密文档）
  • 验证S3权限：确保执行代码的IAM角色拥有目标S3对象的s3:GetObject权限
  • 更新SDK版本：升级boto3到最新版本，避免旧版本兼容性问题：

    pip install --upgrade boto3
    

  • 改用异步API：如果文档页数较多，推荐使用StartDocumentAnalysis异步API，专门处理多页文档，无需逐页调用：

    response = textract.start_document_analysis(
        DocumentLocation={
            "S3Object": {
                "Bucket": bucketname,
                "Name": filename,
            }
        },
        FeatureTypes=["FORMS"],
        HumanLoopConfig={
            "HumanLoopName": uuid.uuid4().hex,
            "FlowDefinitionArn": FLOW_ARN,
            "DataAttributes": {
                "ContentClassifiers": [
                    "FreeOfAdultContent",
                ]
            },
        }
    )
    # 后续轮询获取结果逻辑与start_document_text_detection一致
    

内容的提问来源于stack exchange，提问作者Ritesh Manglani