fix: handle reasoning_effort gracefully across models (#1447)

<!-- .github/pull_request_template.md -->

## Description
The async LLM client fails with non-reasoning models like gpt-4o with
the error: `Completion error: litellm.BadRequestError: OpenAIException -
Unrecognized request argument supplied: reasoning_effort`.

This PR add a sets the `drop_params` config to True to drop all
unsupported model configs instead of catching with a try-except block.

Additionally, the `reasoning_effort` wasn't being set for the sync
client. It adds the parameter for both async & sync for consistency.

## Type of Change
<!-- Please check the relevant option -->
- [x] Bug fix (non-breaking change that fixes an issue)
- [ ] New feature (non-breaking change that adds functionality)
- [ ] Breaking change (fix or feature that would cause existing
functionality to change)
- [ ] Documentation update
- [ ] Code refactoring
- [ ] Performance improvement
- [ ] Other (please specify):

## Changes Made
- Set `litellm.drop_params=True` to auto-drop unsupported parameters
- Changed `reasoning_effort` from `extra_body` to a direct parameter
- Add `reasoning_effort` to the sync client
- Removed redundant retry for `reasoning_effort`

## Testing
<!-- Describe how you tested your changes -->

## Screenshots/Videos (if applicable)
<!-- Add screenshots or videos to help explain your changes -->

## Pre-submission Checklist
<!-- Please check all boxes that apply before submitting your PR -->
- [ ] **I have tested my changes thoroughly before submitting this PR**
- [ ] **This PR contains minimal changes necessary to address the
issue/feature**
- [ ] My code follows the project's coding standards and style
guidelines
- [ ] I have added tests that prove my fix is effective or that my
feature works
- [ ] I have added necessary documentation (if applicable)
- [ ] All new and existing tests pass
- [ ] I have searched existing PRs to ensure this change hasn't been
submitted already
- [ ] I have linked any relevant issues in the description
- [ ] My commits have clear and descriptive messages

## Related Issues
<!-- Link any related issues using "Fixes #issue_number" or "Relates to
#issue_number" -->

## Additional Notes
<!-- Add any additional notes, concerns, or context for reviewers -->

## DCO Affirmation
I affirm that all code in every commit of this pull request conforms to
the terms of the Topoteretes Developer Certificate of Origin.

README.md

@@ -178,7 +178,7 @@ You can also cognify your files and query using cognee UI.
 
 ### Running the UI
 
-Try cognee UI by running ``` cognee-cli -ui ``` command on your terminal.
+Try cognee UI by setting LLM_API_KEY and running ``` cognee-cli -ui ``` command on your terminal.
 
 ## Understand our architecture
 

cognee/infrastructure/llm/structured_output_framework/litellm_instructor/llm/openai/adapter.py

@@ -29,6 +29,9 @@ observe = get_observe()
 
 logger = get_logger()
 
+# litellm to drop unsupported params, e.g., reasoning_effort when not supported by the model.
+litellm.drop_params = True
+
 
 class OpenAIAdapter(LLMInterface):
     """
@@ -132,39 +135,13 @@ class OpenAIAdapter(LLMInterface):
                 api_version=self.api_version,
                 response_model=response_model,
                 max_retries=self.MAX_RETRIES,
-                extra_body={"reasoning_effort": "minimal"},
+                reasoning_effort="minimal",
             )
         except (
             ContentFilterFinishReasonError,
             ContentPolicyViolationError,
             InstructorRetryException,
-        ) as error:
-            if (
-                isinstance(error, InstructorRetryException)
-                and "content management policy" not in str(error).lower()
-            ):
-                logger.debug(
-                    "LLM Model does not support reasoning_effort parameter, trying call without the parameter."
-                )
-                return await self.aclient.chat.completions.create(
-                    model=self.model,
-                    messages=[
-                        {
-                            "role": "user",
-                            "content": f"""{text_input}""",
-                        },
-                        {
-                            "role": "system",
-                            "content": system_prompt,
-                        },
-                    ],
-                    api_key=self.api_key,
-                    api_base=self.endpoint,
-                    api_version=self.api_version,
-                    response_model=response_model,
-                    max_retries=self.MAX_RETRIES,
-                )
-
+        ):
             if not (self.fallback_model and self.fallback_api_key):
                 raise ContentPolicyFilterError(
                     f"The provided input contains content that is not aligned with our content policy: {text_input}"
@@ -246,6 +223,7 @@ class OpenAIAdapter(LLMInterface):
             api_base=self.endpoint,
             api_version=self.api_version,
             response_model=response_model,
+            reasoning_effort="minimal",
             max_retries=self.MAX_RETRIES,
         )