feat(eval): conversation tracking and --preserve-failed flag

[macekond]

Summary

• Add conversation_id and response_id fields to ChatResult for tracing eval conversations back to the GoodData server
• Capture responseId from SSE event data during stream parsing
• Attach conversation_id to exceptions so error reports include it for debugging
• Add --preserve-failed CLI flag to keep failed conversations on the server for post-mortem inspection
• Add preserve_failed to RunConfig and pass through to ChatClient

Test plan

• Unit tests for response_id capture from SSE events (3 tests: top-level, item-level, first-wins)
• Unit tests for conversation_id attachment to ChatResult on success
• Unit tests for --preserve-failed behavior (keeps failed, deletes on success, deletes without flag)
• Unit test for conversation_id on exceptions
• Unit tests for error report including/omitting conversation_id
• Unit test for --preserve-failed CLI flag parsing and passthrough
• All existing tests still pass (3 pre-existing failures unrelated to this PR)

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 79.21%. Comparing base (04bfbc5) to head (386cef3).

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #1672   +/-   ##
=======================================
  Coverage   79.21%   79.21%           
=======================================
  Files         232      232           
  Lines       15809    15809           
=======================================
  Hits        12523    12523           
  Misses       3286     3286           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.