feat: add integration tests for ADK mapper

[liramon2]

Description

This adds integration tests for the ADK mapper added in #326.

Related Issues

#328

Documentation PR

None needed

Type of Change

New feature

Testing

How have you tested the change? Verify that the changes do not break functionality or introduce new warnings.

• I ran hatch run prepare

Checklist

• I have read the CONTRIBUTING document
• I have reviewed and understand every line of code in this PR, including any generated by AI tools, and I can explain why it works
• My change is focused and reasonably small; I have split unrelated work into separate PRs
• I have added any necessary tests that prove my fix is effective or my feature works
• I have updated the documentation accordingly
• I have added an appropriate example to the documentation to outline the feature, or no new docs are needed
• My changes generate no new warnings
• Any dependent changes have been merged and published

---

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

[liramon2]

@strandly-the-agent Review this PR. Consolidate any findings in a single comment.

[liramon2]

@strandly-the-agent Review this PR. Consolidate any findings in a single comment.

[liramon2]

@strandly-the-agent Review this PR. Consolidate any findings in a single comment.

[strandly-the-agent]

No new findings — and the integ job is green for the first time on this PR. Both items from last round are in, and the model-id change resolves an earlier nit better than what I'd suggested.

✅ Run integration tests on 8b4b4c9: SUCCESS (it was FAILURE on 2883ce0, 0c8fc62, 5db3fd9 and dd12768) · ✅ :343-345 is the replacement assertion verbatim · ✅ pyproject.toml:68 → litellm>=1.84 · ✅ lint + all 11 unit jobs green.

One thing worth calling out, because it tightens the earlier diagnosis: dd12768 changed only the model id and left the bad assertion in place — and stayed red. 8b4b4c9 then fixed the assertion and went green. That isolates the cause to the assertion rather than the dependencies or the model, which is a cleaner confirmation than my local reproduction was.

On the model change (:32-34) — good call, and it's better than what I proposed. Extracting DEFAULT_MODEL_ID = "us.anthropic.claude-sonnet-4-20250514-v1:0" matches both sibling modules exactly (test_langchain_openinference_eval.py:28, test_langchain_traceloop_eval.py:31), which is a more useful kind of consistency than the global.-prefix alignment I'd suggested. It also re-proves on a different model what I'd only verified on sonnet-4-6: CI passing means the tool_call_id is not None and 2-trace-split assertions hold on Sonnet 4 too.

I deliberately didn't re-run anything locally this round — CI's own green on py3.10 under hatch is stronger evidence than my sandbox, so re-running would have burned Bedrock calls to learn less.

From my side this is ready to come out of draft. Two optional leftovers, neither blocking: the ADK case in tests/strands_evals/mappers/test_utils.py::TestDetectOtelMapper (still absent — two lines with the existing make_span_dict helper) and the inert Case.expected_output fields. And the offer stands to file the two reproduced mapper defects (delegating agent's raw-JSON agent_response; delegated agent's plumbing-text user_prompt) as follow-ups against #326 — just say the word and I'll open them.

---

_{Reviewed by strandly-the-agent — an experimental agent; a human should still sign off. No local runs this round; nothing pushed, no AWS resources created.}