使用Python调用Azure Cognitive Search向量搜索遇AttributeError问题求助

问题代码 #

query = "What are the advantages of an open-source ai model?"
search_client = SearchClient(AZURE_COGNITIVE_SEARCH_SERVICE_ENDPOINT, AZURE_COGNITIVE_SEARCH_INDEX_NAME, credential=AZURE_COGNITIVE_SEARCH_API_KEY)
vector_query = VectorizableTextQuery(text=query, k=3, fields="content_vector")

results = search_client.search(  
    search_text=None,  
    vector_queries= [vector_query],
    select=["title", "content_vector", "metadata"],
    
)
    
for result in results:  
  print(result)

错误信息 #

AttributeError                            Traceback (most recent call last)
<ipython-input-32-a8dbad346de2> in <cell line: 16>()
     14 print(results)
     15 
---> 16 for result in results:
     17   print(result)

14 frames
/usr/local/lib/python3.10/dist-packages/azure/core/pipeline/policies/_authentication.py in on_request(self, request)
     97                 self._token = self._credential.get_token(*self._scopes, enable_cae=self._enable_cae)
     98             else:
---> 99                 self._token = self._credential.get_token(*self._scopes)
    100         self._update_headers(request.http_request.headers, self._token.token)
    101 

AttributeError: 'str' object has no attribute 'get_token'

执行print(results)返回：

<iterator object azure.core.paging.ItemPaged at 0x7f17cf3af550>

版本信息 #

安装命令及版本：

!pip install azure-search-documents --pre --upgrade

> Requirement already satisfied: azure-search-documents in 
> /usr/local/lib/python3.10/dist-packages (11.4.0b11)

解决方案 #

错误原因是新版azure-search-documents（11.x预览版及以上）的SearchClient不再接受直接传入字符串作为凭证，需要使用AzureKeyCredential类来封装API密钥。

修改步骤：

1. 从azure.core.credentials导入AzureKeyCredential
2. 初始化search_client时，用AzureKeyCredential包裹你的API密钥

修改后的完整代码：

from azure.core.credentials import AzureKeyCredential
from azure.search.documents import SearchClient
from azure.search.documents.models import VectorizableTextQuery

query = "What are the advantages of an open-source ai model?"
# 用AzureKeyCredential包装API密钥
search_client = SearchClient(
    AZURE_COGNITIVE_SEARCH_SERVICE_ENDPOINT, 
    AZURE_COGNITIVE_SEARCH_INDEX_NAME, 
    credential=AzureKeyCredential(AZURE_COGNITIVE_SEARCH_API_KEY)
)
vector_query = VectorizableTextQuery(text=query, k=3, fields="content_vector")

results = search_client.search(  
    search_text=None,  
    vector_queries= [vector_query],
    select=["title", "content_vector", "metadata"],
    
)
    
for result in results:  
  print(result)

这样修改后，凭证会被正确识别为认证对象，不会再抛出get_token相关的属性错误。

内容的提问来源于stack exchange，提问作者STORM