Add backend embedding endpoint via LiteLLM

[Copilot]

Adds server-side embedding generation to support developer-local and hosted deployment architectures through LiteLLM's provider abstraction.

Implementation

Core endpoint (src/codenav/server/vector_api.py)

• POST /api/vector/embedding with OpenAI-compatible request/response format
• Single string or batch input support
• Per-request model override via model parameter
• Provider routing via LiteLLM (Azure OpenAI, OpenAI, Ollama, etc.)

Configuration

• LITELLM_EMBEDDING_MODEL: Default model selection
• AZURE_API_KEY, AZURE_API_BASE, AZURE_API_VERSION: Azure OpenAI
• OPENAI_API_KEY: OpenAI direct

Dependencies

• Added litellm>=1.61.15 to web extras (version vetted for known CVEs)

Example

curl -X POST http://localhost:8000/api/vector/embedding \
  -H "Content-Type: application/json" \
  -d '{"input": ["code snippet", "documentation"], "model": "text-embedding-3-small"}'

Response follows OpenAI embeddings API format:

{
  "object": "list",
  "data": [
    {"object": "embedding", "index": 0, "embedding": [0.1, 0.2, ...]},
    {"object": "embedding", "index": 1, "embedding": [0.3, 0.4, ...]}
  ],
  "model": "text-embedding-3-small",
  "usage": {"prompt_tokens": 10, "total_tokens": 10}
}

Testing

• 11 test cases covering request validation, single/batch processing, model override, environment config, and error handling
• Dual dict/object access pattern for LiteLLM response compatibility

Original prompt

---

This section details on the original issue you should resolve

<issue_title>Add backend embedding endpoint via LiteLLM</issue_title>
<issue_description>## Summary
Add a backend endpoint for embedding text using LiteLLM to support server-side embedding operations.

Background

Per docs/gitnexus-frontend-assessment.md, the architecture should support:

• Developer-local setups (Ollama, in-browser)
• Hosted deployments (Azure OpenAI) routed through LiteLLM
• Embeddings as a configurable provider, not architecture coupling

The frontend now has a configurable embedding provider system (frontend/src/core/embeddings/providers/). The backend needs complementary endpoints.

Requirements

Endpoint: POST /api/vector/embedding

Request:
{
  "input": ["text1", "text2", ...] | "single text",
  "model": "optional-model-override"
}

Response:
{
  "object": "list",
  "data": [
    { "object": "embedding", "index": 0, "embedding": [0.1, 0.2, ...] },
    ...
  ],
  "model": "model-used",
  "usage": { "prompt_tokens": 10, "total_tokens": 10 }
}

Implementation Notes

1. Use LiteLLM for routing to configured embedding provider
2. Support environment-based configuration (LITELLM_EMBEDDING_MODEL, AZURE_OPENAI_*, etc.)
3. Follow OpenAI-compatible response format
4. Add to src/codenav/server/ router

Configuration

• Environment variables for provider selection
• Support Azure OpenAI, OpenAI, Ollama through LiteLLM proxy
• Document in README or .env.example

Related

• Frontend providers: frontend/src/core/embeddings/providers/
• API contract draft: docs/gitnexus-frontend-assessment.md section 11.3
• Session notes: docs/sessions/SESSION_20_SIGMA3_FRONTEND_REWRITE.md</issue_description>

Comments on the Issue (you are @copilot in this section)

• Fixes Add backend embedding endpoint via LiteLLM #52

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.