fix(vertexai): preserve ExecutableCode and CodeExecutionResult in session storage - #1729

Linked issue

Closes: Vertex AI Session Storage drops ExecutableCode and CodeExecutionResult parts #1728

2. Describe the change:

Problem:
When storing and retrieving session history using the Vertex AI backend (session/vertexai), the aiplatformToGenaiContent and createAiplatformpbContent mappers silently drop ExecutableCode, CodeExecutionResult, and FileData parts. There's a TODO in both functions acknowledging this omission. When an agent uses Gemini's Code Execution tool, the session history successfully saves the model's text thought but completely loses the code block and its result, causing the model to hallucinate on subsequent turns because it lacks the execution context it expects.

Solution:
Added mapping logic for Part_ExecutableCode, Part_CodeExecutionResult, and Part_FileData in both aiplatformToGenaiContent and createAiplatformpbContent to ensure these fields are preserved round-trip in session storage.

Behavior change

Agents utilizing Vertex AI for session storage will now correctly preserve code execution history and file data references across conversation turns. Before, these parts were silently dropped during Protobuf conversion.

Testing Plan

Unit Tests:
All unit tests pass locally.
With your source change reverted and your tests kept, which test fails?
go test ./session/vertexai/ -run TestAiplatformToGenaiContent_OtherParts

--- FAIL: TestAiplatformToGenaiContent_OtherParts (0.00s)
    vertexai_test.go:1016: incorrect ExecutableCode: <nil>
FAIL

With the change, go test -race ./... passes, and gofmt / golangci-lint are clean.

Manual End-to-End (E2E) Tests:
Ran an agent with the Code Execution tool enabled using Vertex AI for session storage.
before: The session event stored in Vertex AI lacked the code snippet, causing the model to immediately repeat itself on turn 2.
after: The session event in Vertex AI successfully stored and retrieved the executed code and outcome, allowing a multi-turn code-generation conversation.

Checklist

• I have read the CONTRIBUTING.md document.
• I have performed a self-review of my own code.
• I have commented my code, particularly in hard-to-understand areas.
• I have added tests that prove my fix is effective or that my feature works.
• New and existing unit tests pass locally with my changes.
• I have manually tested my changes end-to-end.
• Any dependent changes have been merged and published in downstream modules.