feat: add custom headers support for OpenAI compatible embeddings

[roomote]

Summary

This PR addresses Issue #8909 by adding support for custom headers when configuring OpenAI compatible gateways for the code base indexing feature. This allows users to add required headers (like API keys or version headers) that some gateways require beyond the standard Authorization header.

Changes

Core Implementation

• ✅ Added codebaseIndexOpenAiCompatibleHeaders field to type definitions in packages/types/src/codebase-index.ts
• ✅ Updated config manager to read and handle custom headers configuration
• ✅ Modified OpenAICompatibleEmbedder to accept custom headers as an optional parameter
• ✅ Updated service factory to pass custom headers when creating the embedder instance
• ✅ Custom headers are properly merged with default headers in both SDK and direct fetch requests

Testing

• ✅ Added comprehensive test coverage for custom headers functionality
• ✅ Tests verify headers are passed correctly to OpenAI client
• ✅ Tests verify headers are included in direct fetch requests
• ✅ Tests verify header override behavior
• ✅ All existing tests continue to pass

Usage

Users can now configure custom headers for OpenAI compatible gateways through the codebase index configuration:

{
  "codebaseIndexOpenAiCompatibleHeaders": {
    "X-Custom-Header": "custom-value",
    "X-API-Version": "v2"
  }
}

These headers will be automatically included in all embedding API requests.

Security Considerations

⚠️ Note: Custom headers are stored as configuration (not as secrets). Users should be aware that sensitive API keys should preferably use the dedicated codebaseIndexOpenAiCompatibleApiKey field which is stored securely as a secret.

Testing

• All tests pass: npx vitest run services/code-index/embedders/__tests__/openai-compatible.spec.ts
• Linting and type checking pass
• 57 tests total, including 3 new tests for custom headers

Fixes #8909

---

Important

Adds support for custom headers in OpenAI-compatible embeddings, allowing configuration of additional headers like API keys.

• Behavior:
  • Adds codebaseIndexOpenAiCompatibleHeaders to configuration schema in codebase-index.ts and config.ts.
  • Updates CodeIndexConfigManager to read and handle custom headers.
  • Modifies OpenAICompatibleEmbedder to accept custom headers and merge them with default headers.
  • Updates service-factory.ts to pass custom headers when creating OpenAICompatibleEmbedder.
• Testing:
  • Adds tests in openai-compatible.spec.ts to verify custom headers are passed correctly and override default headers.
  • Tests include scenarios for direct fetch requests and SDK usage.
• Misc:
  • Updates config-manager.ts to include custom headers in openAiCompatibleOptions.

^{This description was created by for d089fb2. You can customize this summary. It will automatically update as commits are pushed.}

[roomote]

Review Summary

I've reviewed this PR and found 1 issue that needs to be addressed:

Issues to Resolve

• Service restart detection: Custom header changes don't trigger a service restart. The doesConfigChangeRequireRestart() method should check for header changes, and PreviousConfigSnapshot type needs an openAiCompatibleHeaders field to track previous state.

---

Follow Along on Roo Code Cloud

[roomote]

Custom header changes don't trigger a service restart. When users update their headers, the doesConfigChangeRequireRestart() method (lines 259-357) should detect this change and return true, similar to how it checks openAiCompatibleBaseUrl and openAiCompatibleApiKey changes. You'll also need to add an openAiCompatibleHeaders field to the PreviousConfigSnapshot type and update the snapshot capture logic in loadConfiguration() to track previous header state.

[roomote]

Review complete. Found 1 issue that needs to be addressed before approval. Please see the inline comment for details.