fix(llama-index): workflow integration (#1095)

[hassiebp]

Important

Fix serialization issues in LlamaIndexSpanHandler by handling BaseModel and removing Context instances from arguments.

• Behavior:
  • In _parse_output_metadata(), handle BaseModel instances by converting them to dictionaries.
  • In _parse_input(), remove Context instances from arguments to avoid serialization issues.
• Imports:
  • Add BaseModel import from pydantic.
  • Add Context import from llama_index.core.workflow.

^{This description was created by for d4c0df1. It will automatically update as commits are pushed.}

Greptile Summary

Disclaimer: Experimental PR review

Enhanced LlamaIndex integration by improving serialization handling for Pydantic BaseModel results and Context objects in the span handler.

• Added proper serialization support for Pydantic BaseModel results in langfuse/llama_index/_span_handler.py
• Added filtering for llama_index Context objects to prevent serialization errors
• Improved type safety by explicitly importing BaseModel from pydantic
• Added handling for raw dictionary and BaseModel usage data in token usage parsing

[greptile-apps]

LGTM

_{1 file(s) reviewed, no comment(s)
Edit PR Review Bot Settings | Greptile}