feature(litellm): pandasai litellm wrapper

[scaliseraoul]

• Tests added and passed if fixing a bug or adding a new feature.
• All code checks passed.

---

Important

Introduces LiteLLM wrapper for unified LLM interaction in PandaAI, replacing multiple existing LLM integrations.

• New Feature:
  • Introduces LiteLLM wrapper in litellm.py for unified LLM interaction across multiple providers.
  • Supports models from OpenAI, Anthropic, Google, Azure, AWS, and more.
• Documentation:
  • Updates large-language-models.mdx to include LiteLLM setup instructions.
  • Adds README.md for litellm extension.
• Testing:
  • Adds test_litellm.py for unit testing LiteLLM functionality.
• Removals:
  • Deletes old LLM extensions for Google, Bedrock, HuggingFace, IBM, LangChain, and Local models.
• Miscellaneous:
  • Updates pyproject.toml for litellm extension.
  • Removes LangChain LLM handling from state.py and config.py.

^{This description was created by for b5e87ee. It will automatically update as commits are pushed.}

[ellipsis-dev]

❌ Changes requested. Reviewed everything up to b5e87ee in 2 minutes and 39 seconds

More details

• Looked at 2451 lines of code in 40 files
• Skipped 6 files when reviewing.
• Skipped posting 4 drafted comments based on config settings.

1. extensions/llms/litellm/tests/test_litellm.py:87

• Draft comment:
  The 'test_completion_with_extra_params' expects the extra parameters to appear within an 'optional_params' -> 'extra_body' structure. Verify that LiteLLM’s completion function consistently wraps extra kwargs this way, or adjust the test to match the actual behavior.
• Reason this comment was not posted:
  Comment did not seem useful. Confidence is useful = 30% <= threshold 50%
  The comment is asking the author to verify the behavior of a function, which violates the rule against asking the author to confirm their intention or ensure behavior. However, it does point out a potential inconsistency between the test and the actual behavior, which could be useful if rephrased to suggest a specific action.

2. extensions/llms/litellm/pandasai_litellm/litellm.py:9

• Draft comment:
  Consider updating the docstring to be provider-agnostic. The current text mentions “new OpenAI LLM”, but LiteLLM supports multiple providers.
• Reason this comment was not posted:
  Marked as duplicate.

3. docs/v3/large-language-models.mdx:166

• Draft comment:
  Typo: In the example comment on line 166, 'answer should me (mostly) consistent across devices.' should be corrected to 'answer should be (mostly) consistent across devices.'
• Reason this comment was not posted:
  Comment was not on a location in the diff, so it can't be submitted as a review comment.

4. extensions/llms/litellm/tests/test_litellm.py:34

• Draft comment:
  In the definition of the function test_successful_completion, the parameters are written as llm,prompt without a space after the comma. For better readability and consistency with other functions (e.g., test_invalid_api_key), please add a space (i.e., change to llm, prompt).
• Reason this comment was not posted:
  Decided after close inspection that this draft comment was likely wrong and/or not actionable: usefulness confidence = 10% vs. threshold = 50%
  The rules state not to make comments that are obvious or unimportant. While this is a real formatting inconsistency, it's a very minor style issue that could be automatically fixed by code formatters. It doesn't affect functionality or code quality in any meaningful way. Most teams would handle this via automated formatting rather than PR comments.
  The inconsistent formatting could make the code slightly less readable. Some teams do care about maintaining consistent style manually.
  While consistency is good, this is too minor an issue to warrant a PR comment. It would be better handled by automated tooling like black, autopep8, or similar.
  Delete this comment as it points out a trivial formatting issue that doesn't meaningfully impact code quality.

Workflow ID: wflow_HLZJfzCCT1NWyOub

---

Want Ellipsis to fix these issues? Tag @ellipsis-dev in a comment. You can customize Ellipsis with 👍 / 👎 feedback, review rules, user-specific overrides, quiet mode, and more.