fix: add reasoning tokens into token metrics in stream mode for gemini models

[hustxiayang]

Description
The previous implementation on tokens tracking is like

if chunk.UsageMetadata.CandidatesTokenCount >= 0 {
				tokenUsage.SetOutputTokens(uint32(chunk.UsageMetadata.CandidatesTokenCount)) //nolint:gosec
			}

The output tokens should include both the reasoning tokens and the candidates tokens. Thus fix this by reusing the the variable usage, just like non-streaming mode. Also updated the tests to include reasoning tokens

[hustxiayang]

/gemini review

[gemini-code-assist]

Code Review

This pull request refactors how token usage is extracted and recorded during streaming responses in the OpenAI to GCP Vertex AI translator, switching from chunk.UsageMetadata to direct fields on the usage object. However, accessing nested fields like usage.PromptTokensDetails.CachedTokens and usage.CompletionTokensDetails.ReasoningTokens without nil checks introduces a risk of nil pointer dereference panics if those details are not populated in the stream. It is recommended to add defensive nil checks for these fields.

Important

The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.

[gemini-code-assist]

Directly accessing usage.PromptTokensDetails and usage.CompletionTokensDetails without nil checks can lead to a nil pointer dereference panic if these fields are not populated in the streaming chunk. Adding defensive nil checks (similar to the non-streaming implementation) prevents potential runtime panics.

			if usage.PromptTokensDetails != nil {
				tokenUsage.SetCachedInputTokens(uint32(usage.PromptTokensDetails.CachedTokens)) //nolint:gosec
			}
			if usage.CompletionTokensDetails != nil {
				tokenUsage.SetReasoningTokens(uint32(usage.CompletionTokensDetails.ReasoningTokens)) //nolint:gosec
			}

[hustxiayang]

in function geminiUsageToOpenAIUsage, these values are always set

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 84.88%. Comparing base (08ba26e) to head (36e01f1).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2259      +/-   ##
==========================================
- Coverage   84.90%   84.88%   -0.02%     
==========================================
  Files         144      144              
  Lines       21259    21254       -5     
==========================================
- Hits        18049    18042       -7     
- Misses       2131     2132       +1     
- Partials     1079     1080       +1     

☔ 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.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[hustxiayang]

@nacx @aabchoo @yuzisun could you please help to review this fix? Thanks!

[aabchoo]

Thanks for the fix!

[hustxiayang]

/retest