fix(ovhcloud): Fix tool calling not working

[eliasto]

Pre-Submission checklist

• I have Added testing in the tests/test_litellm/ directory, Adding at least 1 test is a hard requirement - see details
• My PR passes all unit tests on make test-unit
• My PR's scope is as isolated as possible, it only solves 1 specific problem
• I have requested a Greptile review by commenting @greptileai and received a Confidence Score of at least 4/5 before requesting a maintainer review

Delays in PR merge?

If you're seeing a delay in your PR being merged, ping the LiteLLM Team on Slack (#pr-review).

CI (LiteLLM team)

CI status guideline:

• 50-55 passing tests: main is stable with minor issues.
• 45-49 passing tests: acceptable but needs attention
• <= 40 passing tests: unstable; be careful with your merges and assess the risk.

• Branch creation CI run
  Link:

• CI run for the last commit
  Link:

• Merge / cherry-pick CI run
  Links:

Screenshots / Proof of Fix

Type

🐛 Bug Fix
🧹 Refactoring
✅ Test

Changes

Remove supported openai params, so only the OVHcloud returns an error if the tool calling is not supported.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
litellm	Ready	Preview, Comment	Apr 17, 2026 1:36pm

[eliasto]

@greptileai Please make a review

[greptile-apps]

Greptile Summary

This PR removes the get_supported_openai_params override in OVHCloudChatConfig that was stripping tools, tool_choice, function_call, and response_format from the supported param list based on a static model catalog lookup. After the fix, OVHCloudChatConfig inherits the full OpenAI param list from OpenAIGPTConfig, so tool-calling params are always forwarded to the OVHCloud server, which is responsible for returning an error if a specific model doesn't support them. The new test_tools_not_filtered_by_static_model_map parametrized test correctly verifies this behavior without making network calls.

Confidence Score: 5/5

Safe to merge — targeted removal of a stale static filter with a clear unit test verifying the new behavior.

All findings are P2 or lower. The core change is a clean deletion of over-eager param filtering; the inherited behavior from OpenAIGPTConfig is well-understood and correct. The new test exercises the exact code path that was broken. No security, data-integrity, or reliability concerns.

No files require special attention.

Important Files Changed

Filename	Overview
litellm/llms/ovhcloud/chat/transformation.py	Removes the get_supported_openai_params override that conditionally stripped tool-calling params based on a stale static catalog; now inherits the full OpenAI param list from OpenAIGPTConfig.
tests/test_litellm/llms/ovhcloud/test_ovhcloud_chat_transformation.py	Adds test_tools_not_filtered_by_static_model_map (pure unit test, no network calls) covering 5 model names; remaining changes are Black-style formatting fixes.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A[LiteLLM call with tools param] --> B{OVHCloudChatConfig\nget_supported_openai_params}
    
    subgraph BEFORE["BEFORE (removed)"]
        B1[Call _get_model_info_helper] --> B2{supports_function_calling?}
        B2 -->|True| B3[Include tools, tool_choice,\nfunction_call, response_format]
        B2 -->|False / error| B4[Strip tools, tool_choice,\nfunction_call, response_format]
    end

    subgraph AFTER["AFTER (this PR)"]
        C[Inherit OpenAIGPTConfig list:\nalways includes tools, tool_choice,\nfunction_call, response_format]
    end

    B --> BEFORE
    B --> AFTER

    B3 --> D[OVHCloud API]
    C --> D
    B4 --> E[Tools silently dropped —\nrequest fails for supported models]
    D --> F{Server-side validation}
    F -->|Supported| G[Success]
    F -->|Not supported| H[Server returns error]

Loading

_{Reviews (2): Last reviewed commit: "fix import order" | Re-trigger Greptile}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!