Linting Refactor: New `ModelResponseChunk` for streaming

[krrishdholakia]

Title

Uses a new pydantic object - ModelResponseChunk for streaming calls.

Relevant issues

Reference: https://github.com/BerriAI/litellm/issues/4206

This fixes linting errors for model response caused because it was used for streaming + non streaming calls

e.g. note the lack of linting errors for .message

Type

🧹 Refactoring

Changes

• new ModelResponseChunk
• CustomStreamWrapper now returns ModelResponseChunk instead of ModelResponse
• Refactors ModelResponse -> moves the 'streamingchoices' logic to ModelResponseChunk

[REQUIRED] Testing - Attach a screenshot of any new tests passing locall

If UI changes, send a screenshot/GIF of working UI fixes

[vercel[bot]]

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

Name	Status	Preview	Comments	Updated (UTC)
litellm	❌ Failed (Inspect)			Jun 16, 2024 8:19am

[krrishdholakia]

This keeps the same response as ModelResponse for streaming.

The only change is moving it to be a response object of it's own (fixing linting errors).

This is not considered a breaking change.

---

We should make sure this is on release notes for the next few releases, to make sure people are aware of the change.