fixes openai import #65
Conversation
There was a problem hiding this comment.
Pull Request Overview
This PR updates the OpenAI client implementation to use the legacy OpenAI Python library API instead of the newer client-based approach. The changes switch from using the OpenAI class constructor to the older global API key configuration pattern.
- Changes import from
from openai import OpenAItoimport openai - Replaces client instantiation with global API key assignment
- Updates API call from client method to module-level method
| load_dotenv() | ||
|
|
||
| _client = OpenAI(api_key=os.getenv("OPENAI_API_KEY")) | ||
| openai.api_key = os.getenv("OPENAI_API_KEY") |
There was a problem hiding this comment.
This change reverts to the deprecated OpenAI v0.x API pattern. The newer OpenAI v1.x+ client-based approach (using OpenAI() constructor) is the recommended and supported method. Consider keeping the modern client pattern instead.
| def get_chat_completion(instructions, input, model="gpt-4o", temperature=0.2): | ||
| """Return the assistant message from an OpenAI responses API call.""" | ||
| resp = _client.responses.create( | ||
| resp = openai.responses.create( |
There was a problem hiding this comment.
The API endpoint 'responses.create' does not exist in the OpenAI API. This should likely be 'chat.completions.create' or 'completions.create' depending on the intended functionality.
No description provided.