feat: add proto-plus message serialization support in EventSerializer…

[Focadecombate]

User description

Shortcut: sc-55119

Description:

The GalileoCallback (and other handlers using EventSerializer) would produce empty objects {} when serializing proto-plus messages (e.g. google.cloud.aiplatform
types). This happens because proto-plus Message objects store their data internally via protobuf descriptors rather than in __dict__, so the generic __dict__ fallback
in EventSerializer would find nothing to serialize.

This PR adds explicit handling for proto.Message instances in EventSerializer, using proto.Message.to_dict() to correctly convert them to plain dictionaries before JSON
encoding. The proto dependency check is lazy to avoid impacting users who don't use Google Cloud AI Platform or other proto-plus libraries.

Changes:

• Add is_proto_plus_available dependency check in dependencies.py
• Add proto.Message serialization branch in EventSerializer.default(), placed before the __slots__ fallback
• Add tests covering simple, empty, and nested proto-plus message serialization

After change

Tests:

• Unit Tests Added
• E2E Test Added (if it's a user-facing feature, or fixing a bug)

---

Generated description

Below is a concise technical summary of the changes proposed in this PR:
Enhances the GalileoCallback to correctly capture tool calls and token metrics when using ChatVertexAI by adding support for proto-plus message serialization. Updates the EventSerializer and LangChain handlers to handle GCP-specific metadata and ensure proper status code reporting for various node types.

Topic	Details
Proto-Plus Support	Implements proto.Message serialization in EventSerializer to prevent empty objects when processing Google Cloud AI Platform types. Modified files (3) src/galileo/utils/dependencies.py src/galileo/utils/serialization.py tests/utils/test_serialization.py Latest Contributors(2) User Commit Date Focadecombate feat-add-serialization... February 09, 2026 mikebranc fix-Tool-calls-not-app... November 18, 2025
LangChain Handlers	Refactors parse_llm_result to support GCP-style token keys (input_tokens/output_tokens) and adds status_code tracking to LangChain callback handlers. Modified files (4) src/galileo/handlers/langchain/async_handler.py src/galileo/handlers/langchain/handler.py src/galileo/handlers/langchain/utils.py tests/test_langchain.py Latest Contributors(2) User Commit Date AnkushMalaker fix-update-handler-to-... January 30, 2026 Focadecombate feat-Implement-update_... December 18, 2025

This pull request is reviewed by Baz. Review like a pro on (Baz).

[codecov]

Codecov Report

❌ Patch coverage is 73.84615% with 17 lines in your changes missing coverage. Please review.
✅ Project coverage is 81.78%. Comparing base (0e364eb) to head (4463b10).
⚠️ Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
src/galileo/handlers/langchain/handler.py	55.55%	8 Missing ⚠️
src/galileo/handlers/langchain/async_handler.py	70.00%	3 Missing ⚠️
src/galileo/handlers/langchain/utils.py	90.32%	3 Missing ⚠️
src/galileo/utils/serialization.py	40.00%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #485      +/-   ##
==========================================
- Coverage   81.85%   81.78%   -0.07%     
==========================================
  Files          96       96              
  Lines        9164     9197      +33     
==========================================
+ Hits         7501     7522      +21     
- Misses       1663     1675      +12     

☔ View full report in Codecov by Sentry.
📢 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.