Upgraded OpenAI SDK Version #65
Conversation
|
@MinuraPunchihewa, @ea-rus What if we make the OpenAI SDK as optional dependency in |
@ZoranPandovski Is that possible? I think that every operation that we are running via the SDK requires the OpenAI SDK, right? Since it interacts with our API? |
|
@ea-rus I think some of the tests were broken, but please check if my changes are correct. |
tests/unit/test_unit.py
Outdated
| @@ -443,3 +444,8 @@ def openai_completion_f(messages, *args, **kwargs): | |||
| if question == chunk.content.lower(): | |||
| success = True | |||
| assert success is True | |||
|
|
|||
|
|
|||
| if __name__ == '__main__': | |||
There was a problem hiding this comment.
this is extra
tests should be run using pytest command
|
|
||
| expected_create_request = { | ||
| 'name': test_knowledge_base_config.name, | ||
| 'description': test_knowledge_base_config.description, | ||
| 'vector_store': { | ||
| 'engine': test_vector_store_config.engine, | ||
| 'connection_data': test_vector_store_config.connection_data | ||
| 'connection_data': test_vector_store_config.connection_data, | ||
| 'table': test_vector_store_config.table |
There was a problem hiding this comment.
ok, this was added here (but test skipped) in https://github.com/mindsdb/minds_python_sdk/pull/63/files#diff-032aaf721a24a943fd1e19218ac0481328467c94e2debfa4e50d4182af0b1984R121
This PR updates the OpenAI SDK version to avoid the issue describe in the issue below.
As per the thread given here, upgrading to the pinned version resolves the issue given above.
Using this new version, I am able to use the SDK without a problem:
Fixes #64