refactor: Update to StreamingChunk, better index setting and change tool_call to tool_calls

[sjrl]

Related Issues

• fixes Check if the util method _convert_chat_completion_chunk_to_streaming_chunk works as expected for integrations that inherit OpenAIChatGenerator #9502

Proposed Changes:

• Update _convert_chat_completion_chunk_to_streaming_chunk to properly set index=None in the case for streaming chunks that only contain role information or have their content set to None (e.g. when providing finish_reason). This was the original intent of StreamingChunk.index, but was missed in the original PR.
• Update StreamingChunk to have a tool_calls field instead of a singular tool_call. So tool_calls becomes Optional[List[ToolCallDelta]] to better reflect that a OpenAI Chunk can have multiple tool calls in it. I was able to test this behavior for the first time when using the MistralChatGenerator.
  • Updated relevant parts of the code and tests to reflect this change.
  • Updated the original release note to reflect this change as well.
  • Updated the print_streaming_callback function to accommodate this change

How did you test it?

• Updated and added unit tests.

Notes for the reviewer

Example of OpenAI Agent Streaming Formatting

Example of Mistral Agent Streaming Formatting

Checklist

• I have read the contributors guidelines and the code of conduct
• I have updated the related issue with new insights and changes
• I added unit tests and updated the docstrings
• I've used one of the conventional commit types for my PR title: fix:, feat:, build:, chore:, ci:, docs:, style:, refactor:, perf:, test: and added ! in case the PR includes breaking changes.
• I documented my code
• I ran pre-commit hooks and fixed any issue

[sjrl]

@anakin87 I'm setting @julian-risch as the reviewer since he reviewed the original PR that introduced the new fields to StreamingChunk so he has the most context for these changes.

[sjrl]

Sorry @julian-risch and @anakin87 I'm switching this to draft for now. I've run into some challenges with Mistral so I'm considering rethinking how StreamingChunk works with respect to handling multiple tool calls.

[coveralls]

Pull Request Test Coverage Report for Build 15818567920

Warning: This coverage report may be inaccurate.

This pull request's base commit is no longer the HEAD commit of its target branch. This means it includes changes from outside the original pull request, including, potentially, unrelated coverage changes.

• For more information on this, see Tracking coverage changes with pull request builds.
• To avoid this issue with future PRs, see these Recommended CI Configurations.
• For a quick fix, rebase this PR at GitHub. Your next report should be accurate.

Details

• 0 of 0 changed or added relevant lines in 0 files are covered.
• 22 unchanged lines in 5 files lost coverage.
• Overall coverage decreased (-0.02%) to 90.157%

---

Files with Coverage Reduction	New Missed Lines	%
tracing/datadog.py	2	96.0%
components/generators/chat/openai.py	3	96.36%
core/pipeline/async_pipeline.py	3	67.74%
dataclasses/streaming_chunk.py	5	91.23%
components/generators/utils.py	9	75.0%

Totals
Change from base Build 15728526114:	-0.02%
Covered Lines:	11569
Relevant Lines:	12832

---

💛 - Coveralls

[julian-risch]

Looks good to me! 👍 I didn't find anything that needs to be changed. For the three consecutive print lines, I suggest to combine them but I don't want to block the PR because of that simplification.