feat(providers): add HoneyHiveProvider

[sanjeed5]

Description

Adds HoneyHiveProvider, a TraceProvider that fetches a Strands agent session from HoneyHive and converts it to the evals trace types, so evaluators can score stored sessions without re-running the agent.

• invoke_agent spans become AgentInvocationSpan, model events become InferenceSpan, and execute_tool spans become ToolExecutionSpan. Event-loop-cycle spans are dropped. Events that fail to convert are logged and skipped.
• Works with sessions from Strands Python (older and current SDK versions) and Strands TypeScript.
• Fetches events from POST /v1/events/search, paginating until the reported count. ReadTimeout is retried.
• Needs a HoneyHive project API key with read access. A 401/403 raises ProviderError with a clear message.
• Configured with api_key / api_url, or the HH_API_KEY / HH_API_URL environment variables.
• Uses httpx directly rather than the HoneyHive SDK, so evals installs don't pull in the HoneyHive tracer and its OpenTelemetry dependencies. The optional honeyhive extra only declares httpx, which strands-agents already depends on.

Related Issues

Closes #418

Documentation PR

N/A. The provider README section is updated in this PR.

Type of Change

New feature

Testing

• Unit tests use three recorded HoneyHive sessions (Strands Python older SDK, Strands Python current SDK, Strands TypeScript) and a mock HTTP transport. No network calls.

• They cover conversion, tool call and result pairing, tool span shapes, parallel tool results, string tool input, skipping events that fail to convert, pagination, retries, rejected API keys, and a missing session (17 tests).

• ruff check, ruff format --check, mypy -p src, and the full unit suite (2233 tests) pass.

• Also checked against live HoneyHive sessions with Experiment(...).run_evaluations(provider.as_task()) and LLM-judge evaluators.

• 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.