Azure函数PDF上传+Gen AI分析问题求助：NoneType迭代错误

问题描述 #

我是Azure函数新手，正在开发一个可上传PDF文档、通过生成式AI分析获取贷款风险洞察的函数，目前遇到以下问题：

• 触发NoneType对象不可迭代错误
• 使用Postman测试时，GET方法无法上传文件
• 调用/risk_prediction路由报错
• /ingest路由无响应

错误信息 #

System.Private.CoreLib: Exception while executing function: Functions.risk_prediction. System.Private.CoreLib: Result: Failure
Exception: TypeError: 'NoneType' object is not iterable

代码实现 #

import logging
import pathlib
import tempfile
import PyPDF2
import azure.functions as func
from langchain_groq import ChatGroq


app = func.FunctionApp(http_auth_level=func.AuthLevel.ANONYMOUS)
groq_api_key=" key"
@app.route(route="ingest", auth_level=func.AuthLevel.ANONYMOUS)
def test(req: func.HttpRequest) -> func.HttpResponse:
    logging.info('Python HTTP trigger function processed a request.')

    try:
        # Fetch the file from the form-data
        file = req.files.file('file')
        # Fetch GUID from form-data
        guid = req.form.get('guid') 
        if not guid and req.headers.get('Content-Type') == 'application/json':
            try:
                req_body = req.get_json()
            except ValueError:
                pass
            else:
                guid = req_body.get('guid')
        if not guid:
            return func.HttpResponse(
                "GUID is missing in the request.",
                status_code=400
            )
        
        suffix = pathlib.Path(file.filename).suffix
        with tempfile.NamedTemporaryFile(delete=False, suffix=suffix) as temp:
            temp.write(file.read())
            temp_path = temp.name
        
        documents = []
        with open(temp_path, 'rb') as file:
            reader = PyPDF2.PdfReader(file)
            text = ''
            for page_num in range(len(reader.pages)):
                page = reader.pages[page_num]
                text += page.extract_text()
            documents.append(text)
        message = f"data: {documents}"
        # Join all document texts into a single string
        return func.HttpResponse(message)
    except Exception as e:
        return func.HttpResponse(f"Info: Error Encountered. -- {e}", status_code=500)

 @app.route(route="risk_prediction")
def risk_prediction(req: func.HttpRequest) -> func.HttpResponse:
    logging.info('Python HTTP trigger function processed a request.')
    documents=test(req)
    llm = ChatGroq(groq_api_key=groq_api_key,
                  model_name="llama3-70b-8192")
    result=''
    prompt='''Provide valuable insight on loan risk for applicants, by anlyzing the documents {documents}'''
    for doc in documents:
        messages = [
    (
        "system",prompt,
    ),
    ("human", f'''contract document is {doc} and anlysis of previous document is {result}, include previous findings with current data and update response''')]
    
        result =llm.invoke(messages).content
    
    return result

测试URL #

• http://localhost:7071/api/test
• http://localhost:7071/api/risk_prediction

问题分析与修复方案 #

1. NoneType不可迭代错误根源 #

risk_prediction函数直接调用test(req)获取documents，但test返回的是func.HttpResponse对象（不是PDF文本列表）。若test返回错误响应，documents会成为无法迭代的对象，触发报错。

2. 核心修复点 #

（1）拆分PDF解析逻辑为独立函数

将文件解析与HTTP响应逻辑分离，避免路由函数耦合：

def parse_pdf(file) -> list:
    """解析PDF文件并返回文本列表"""
    documents = []
    suffix = pathlib.Path(file.filename).suffix
    with tempfile.NamedTemporaryFile(delete=False, suffix=suffix) as temp:
        temp.write(file.read())
        temp_path = temp.name
    
    with open(temp_path, 'rb') as f:
        reader = PyPDF2.PdfReader(f)
        text = ''
        for page_num in range(len(reader.pages)):
            page = reader.pages[page_num]
            text += page.extract_text()
        documents.append(text)
    # 清理临时文件
    pathlib.Path(temp_path).unlink()
    return documents

（2）修正risk_prediction函数逻辑

不再调用路由函数获取数据，直接从请求中解析文件并分析：

@app.route(route="risk_prediction", auth_level=func.AuthLevel.ANONYMOUS, methods=['POST'])
def risk_prediction(req: func.HttpRequest) -> func.HttpResponse:
    logging.info('Python HTTP trigger function processed a request.')
    try:
        file = req.files.get('file')
        if not file:
            return func.HttpResponse("PDF文件缺失", status_code=400)
        
        documents = parse_pdf(file)
        llm = ChatGroq(groq_api_key=groq_api_key, model_name="llama3-70b-8192")
        result = ''
        prompt = "基于提供的文档，分析贷款申请人的风险并给出有价值的洞察"
        
        for doc in documents:
            messages = [
                ("system", prompt),
                ("human", f"合同文档内容：{doc}\n之前的分析结果：{result}\n结合已有内容更新风险分析")
            ]
            result = llm.invoke(messages).content
        
        return func.HttpResponse(result, status_code=200)
    except Exception as e:
        return func.HttpResponse(f"分析出错：{str(e)}", status_code=500)

（3）修复ingest路由的文件获取逻辑

原代码req.files.file('file')写法错误，改为req.files.get('file')，并明确支持POST方法：

@app.route(route="ingest", auth_level=func.AuthLevel.ANONYMOUS, methods=['POST'])
def ingest(req: func.HttpRequest) -> func.HttpResponse:
    logging.info('Python HTTP trigger function processed a request.')
    try:
        file = req.files.get('file')
        if not file:
            return func.HttpResponse("PDF文件缺失", status_code=400)
        
        guid = req.form.get('guid') 
        if not guid and req.headers.get('Content-Type') == 'application/json':
            try:
                req_body = req.get_json()
            except ValueError:
                pass
            else:
                guid = req_body.get('guid')
        if not guid:
            return func.HttpResponse("请求中缺少GUID", status_code=400)
        
        documents = parse_pdf(file)
        # 此处可添加文档与GUID关联存储的逻辑（如Azure Blob/Cosmos DB）
        return func.HttpResponse(f"文件已解析，GUID：{guid}，内容长度：{len(documents[0])}", status_code=200)
    except Exception as e:
        return func.HttpResponse(f"处理出错：{str(e)}", status_code=500)

（4）Postman测试规范

• 必须使用POST方法，GET方法无法携带文件
• 请求头Content-Type设置为multipart/form-data
• 在form-data中添加file字段（类型选文件）和guid字段（类型选文本）

3. 其他优化点 #

• 添加临时文件删除逻辑，避免磁盘占用
• 修正拼写错误：anlyzing→analyzing，anlysis→analysis

内容的提问来源于stack exchange，提问作者prateek s