HTTPSConnection object at 0x00000244F06ABA10>: Failed to resolve 'https' ([Errno 11001] getaddrinfo failed)

[cetmca]

I am getting the below error like this, any thoughts

python -m scripts generate --output=example_input/qa.jsonl --numquestions=200 --persource=5

File "C:\ragevaluator.venv\Lib\site-packages\azure\core\pipeline\transport_requests_basic.py", line 401, in send
raise error
azure.core.exceptions.ServiceRequestError: <urllib3.connection.HTTPSConnection object at 0x00000244F06ABA10>: Failed to resolve 'https' ([Errno 11001] getaddrinfo failed)

my .env looks like this

OPENAI_HOST="azure"
OPENAI_GPT_MODEL="gpt-35"

For Azure OpenAI only:

AZURE_OPENAI_EVAL_DEPLOYMENT="gpt-35-turbo"
AZURE_OPENAI_SERVICE="https://openai-xxxxxzammoai-dev-xxxx.openai.azure.com/"
AZURE_OPENAI_KEY="42db2466e"

For generating QA based on search index:

AZURE_SEARCH_SERVICE="https://acxxs-xxxxx-311zammoai-dev.search.windows.net"
AZURE_SEARCH_INDEX="index-xxxx-zmobot-dev"
AZURE_SEARCH_KEY=""

[pamelafox]

Ah, the SERVICE env vars should only be the service name, not the full endpoint. I construct the endpoint in the code:

For example, mine looks like:

AZURE_OPENAI_SERVICE="evaldjgxrh3u7sxlu-openai"
AZURE_SEARCH_SERVICE="gptkb-xw55anu4yrb3k"

[cetmca]

Now getting the forbidden response, can you please suggest File "C:\ragevaluator\.venv\Lib\site-packages\azure\search\documents\_generated\operations\_documents_operations.py", line 756, in search_post raise HttpResponseError(response=response, model=error) azure.core.exceptions.HttpResponseError: Operation returned an invalid status 'Forbidden'

…

On Tue, Jun 11, 2024 at 12:48 PM Pamela Fox ***@***.***> wrote: Ah, the SERVICE env vars should only be the service name, not the full endpoint. I construct the endpoint in the code: For example, mine looks like: AZURE_OPENAI_SERVICE="evaldjgxrh3u7sxlu-openai" AZURE_SEARCH_SERVICE="gptkb-xw55anu4yrb3k" — Reply to this email directly, view it on GitHub <#93 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACHE7VPLEQWXMSNMP6ST4LTZG4S7RAVCNFSM6AAAAABJEWXCW2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCNRRGIYDMOBXGI> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[pamelafox]

Hm, is your search service setup with role-based access control for the currently logged in user?

[cetmca]

Its not role based,
I just enabled more logs and see the search service itself is not retrieving the response(empty response)

how can we test that ?

I tested with a standalone script and it is working fine,i am getting the response from our search

r = search_client.search("", top=1000)     ## this response is empty. 
qa: list[dict] = []
for doc in r:
    if len(qa) > num_questions_total:
        break
   ## logger.info("Processing search document %s", doc["sourcepage"])
    logger.info("test - line 3666666666",doc)
    
    
    
    Error - 
    
    r = search_client.search("", top=1000)     ## this response is empty. 

qa: list[dict] = []
for doc in r:
if len(qa) > num_questions_total:
break

logger.info("Processing search document %s", doc["sourcepage"])

logger.info("test - line 3666666666",doc)

[cetmca]

any thought @pamelafox ??
hardcoded the end points in return SearchClient(), and still it is not working.
But Like i said, standalone code works fine with the same end points and below is returning the value.

##Perform the search
results = search_client.search("", top=10)