fix: exclude cache tokens from context window calculation

[daniel-lxs]

Description

This PR fixes the token counting discrepancy where the UI was displaying inflated context token counts (e.g., ~170k instead of ~44k) because cache tokens were being included in the context window calculation.

Problem

The function was calculating as:

contextTokens = tokensIn + tokensOut + cacheWrites + cacheReads

However, cache tokens do not consume actual context window space - they are a caching optimization. Including them in the context window display was misleading users about how much of their context window was actually being used.

Solution

Modified the calculation to only include actual context tokens:

contextTokens = tokensIn + tokensOut

Changes

• Modified src/shared/getApiMetrics.ts to exclude cache tokens from the contextTokens calculation
• Updated all related tests in src/shared/__tests__/getApiMetrics.spec.ts to reflect the new behavior
• All tests now pass with the corrected calculation

Testing

• ✅ All unit tests pass
• ✅ Linting passes
• ✅ Type checking passes

Impact

This change only affects the UI display of context tokens. It does not impact:

• Cost calculations (cache tokens are still tracked separately)
• API usage metrics
• Context management logic (sliding window, condensing)

The fix ensures users see an accurate representation of their actual context window usage.

---

Important

Fixes token counting in getApiMetrics.ts by excluding cache tokens for OpenAI protocol and updates related tests.

• Behavior:
  • Fixes token counting in getApiMetrics.ts by excluding cache tokens from contextTokens for OpenAI protocol.
  • Updates getApiMetrics() to calculate contextTokens based on apiProtocol.
• Tests:
  • Updates tests in getApiMetrics.spec.ts to reflect new contextTokens calculation logic.
• Misc:
  • Adds apiProtocol handling in Task.ts to determine protocol based on provider.
  • Adds getApiProtocol() function in provider-settings.ts to determine API protocol.

^{This description was created by for 62c9d5d. You can customize this summary. It will automatically update as commits are pushed.}

[delve-auditor]

✅ No security or compliance issues detected. Reviewed everything up to 62c9d5d.

Security Overview

• 🔎 Scanned files: 6 changed file(s)

Detected Code Changes

Change Type	Relevant files
Enhancement	► provider-settings.ts     Add API protocol helpers and Anthropic style provider constants ► message.ts     Add apiProtocol field to schema ► Task.ts     Implement API protocol handling in messages ► ExtensionMessage.ts     Add apiProtocol type definition
Bug Fix	► getApiMetrics.ts     Update context tokens calculation based on API protocol ► getApiMetrics.spec.ts     Update tests for API protocol logic
Refactor	► runEvals.ts     Simplify log directory handling ► runTask.ts     Streamline logging configuration
Configuration changes	► docker-compose.yml     Update ports and container configuration ► global-settings.ts     Update announcement ID

Reply to this PR with @delve-auditor followed by a description of what change you want and we'll auto-submit a change to this PR to implement it.

[mrubens]

Hmm, I'm confused - I thought that these were additive, at least for some providers. Maybe we're just handling openrouter incorrectly now since we added cached tokens to it?