使用LlamaIndex+MongoDB向量搜索遇API连接及协议错误求助

我参考LlamaIndex官方的MongoDBAtlasVectorSearch示例实现向量搜索，运行代码如下：

# mongo_uri = os.environ["MONGO_URI"]
mongo_uri = (
    "mongodb+srv://<username>:<password>@ai-test.ugfvqq8.mongodb.net?retryWrites=true&w=majority"
)
mongodb_client = pymongo.MongoClient(mongo_uri)
store = MongoDBAtlasVectorSearch(mongodb_client)
storage_context = StorageContext.from_defaults(vector_store=store)
uber_docs = SimpleDirectoryReader(
    input_files=["./data/10k/uber_2021.pdf"]
).load_data()
index = VectorStoreIndex.from_documents(
    uber_docs, storage_context=storage_context
)

运行后出现如下错误：

WARNING:llama_index.llms.openai_utils:Retrying llama_index.embeddings.openai.get_embeddings in 0.02318287802497976 seconds as it raised APIConnectionError: Connection error..
---------------------------------------------------------------------------
LocalProtocolError                        Traceback (most recent call last)
/usr/local/lib/python3.10/dist-packages/httpcore/_exceptions.py in map_exceptions(map)
      9     try:
---> 10         yield
     11     except Exception as exc:  # noqa: PIE786

294 frames
/usr/local/lib/python3.10/dist-packages/httpcore/_sync/http11.py in _send_request_headers(self, request)
    141         with map_exceptions({h11.LocalProtocolError: LocalProtocolError}):
---> 142             event = h11.Request(
    143                 method=request.method,

/usr/local/lib/python3.10/dist-packages/h11/_events.py in __init__(self, method, headers, target, http_version, _parsed)
     95             object.__setattr__(
---> 96                 self, "headers",     normalize_and_validate(headers, _parsed=_parsed)
     97             )

/usr/local/lib/python3.10/dist-packages/h11/_headers.py in   normalize_and_validate(headers, _parsed)
    163             validate(_field_name_re, name, "Illegal header name {!r}", name)
---> 164             validate(_field_value_re, value, "Illegal header value {!r}", value)
    165         assert isinstance(name, bytes)

/usr/local/lib/python3.10/dist-packages/h11/_util.py in validate(regex, data, msg, *format_args)
     90             msg = msg.format(*format_args)
---> 91         raise LocalProtocolError(msg)
     92     return match.groupdict()

LocalProtocolError: Illegal header value b'Bearer '

已将MongoDB的网络访问权限设置为0.0.0.0/0，但问题仍存在，求解决方案。

解决方案 #

• 检查OpenAI API密钥配置
  错误提示Illegal header value b'Bearer '说明请求OpenAI Embedding接口时，Authorization头的Bearer值为空，核心原因是未正确配置OpenAI API密钥：

  1. 确认环境变量OPENAI_API_KEY已设置，值为你的有效OpenAI API密钥，无多余空格或换行符。
  2. 也可在代码中显式指定API密钥：

     from llama_index.embeddings.openai import OpenAIEmbedding
     
     # 显式设置嵌入模型及API密钥
     embed_model = OpenAIEmbedding(api_key="你的实际OpenAI API密钥")
     index = VectorStoreIndex.from_documents(
         uber_docs, storage_context=storage_context, embed_model=embed_model
     )
     

• 验证MongoDB连接有效性
  虽已放开网络权限，仍需确认连接配置正确性：

  1. 替换MongoDB URI中的<username>和<password>为实际账号密码，若密码含特殊字符（如@、$）需做URL编码（例如@替换为%40）。
  2. 添加代码测试MongoDB连接：

     try:
         mongodb_client.admin.command('ping')
         print("MongoDB连接正常")
     except Exception as e:
         print(f"MongoDB连接失败：{str(e)}")
     

• 更新依赖包至稳定版本
  部分旧版本的httpcore、h11或llama-index可能存在兼容性问题，执行以下命令更新：

  pip install --upgrade llama-index pymongo httpcore h11 openai
  

内容的提问来源于stack exchange，提问作者Ping Kee Ng