fix(evaluation): skip invocations without user events to prevent ValidationError

[Oxygen56]

Summary

Changes user_content initialization from Content(parts=[]) to None and skips invocations that have no user-authored events.

Previously, internal/system-driven invocations lacking a user event would produce semantically incorrect eval cases with empty content objects, causing a ValidationError.

Changes

• Initialize user_content = None instead of Content(parts=[])
• Add guard to skip invocations where no user event was found (continue to next invocation)
• Prevents ValidationError from empty Content objects

Fixes #3760

[adk-bot]

I have successfully analyzed GitHub Pull Request #5929 and generated a detailed PR Analysis Report under findings artifact:

👉 analysis_results.md

Summary of Analysis Work

1. Compliance Verified: Confirmed that the Contributor License Agreement (cla/google) check is verified and has successfully passed for PR fix(evaluation): skip invocations without user events to prevent ValidationError #5929.
2. Context Retrieved: Fetched complete PR details & changes, and linked Issue #3760, verifying the underlying Pydantic ValidationError crash during evaluation session addition.
3. Workspace Inspection: Inspected evaluation_generator.py to assess correctness and structural soundness of the fix.
4. Outcome / Recommendation: Approve with Nits. The fix correctly skips empty invocations without crashing, but the contributor has not supplied any unit tests. I have proposed a concrete unit test to add to test_evaluation_generator.py in the report.

Next Steps & Decisions

• Would you like to request the test addition directly on the PR, or should we checkout the branch and add the suggested unit test ourselves?