feat: added llm based context specific for issues in post/debugging

[vedanshkap2692]

Description

Implements optional LLM-powered, context-aware fix suggestions with caching, UI indicators, and graceful fallback to the existing rule-based engine.

Key behavior

• When LLM_ENABLED=true and a valid LLM_API_KEY is available, the engine calls the LLM to generate concise, context-aware fixes for detected issues and replaces the generic rule-based suggestion.
• LLM responses are cached in-process for 1 hour (configurable TTL) to avoid duplicate calls for identical issues.
• If the LLM call fails or is disabled, the engine automatically falls back to the existing rule-based suggestion engine.
• API responses now include ai_generated and ai_cached flags so clients can surface an “AI” badge and tooltip when appropriate.
• Health endpoint and startup visibility added so frontend/Swagger can determine whether LLM integration is enabled.

---

Related Issue

Fixes #

---

Type of change

• Bug fix (duplicate append removal, improved .env loading visibility)
• New feature / enhancement (LLM integration, caching, metadata flags)
• Documentation update (README, health endpoint usage)
• Test addition (LLM tests, cache behavior, fallback validation)
• Refactor (startup/env loading cleanup and health reporting improvements)

---

Files changed (high level)

Backend

• backend/app/services/ai_provider.py

  • Added call_llm_sync()
  • Added get_fix_for_issue()
  • Implemented _LLM_FIX_CACHE with TTL-based caching
  • Improved prompts and fallback handling

• backend/app/services/code_assistant.py

  • run_bug_detection() now integrates with ai_provider.get_fix_for_issue()
  • Added ai_generated and ai_cached metadata
  • Removed duplicate append bug

• backend/app/schemas.py

  • Extended Issue model with:
    • code_context
    • ai_generated
    • ai_cached
  • Extended HealthResponse with:
    • llm_enabled
    • llm_model

• backend/app/main.py

  • Loads .env early during startup/import
  • Health endpoint now reports LLM status
  • Added startup logging for easier debugging

Frontend

• frontend/index.html
• frontend/script.js
  • Added .ai-badge UI support
  • Added rendering logic for AI-generated fixes
  • Updated default backend API port to 8010
  • Added health-driven engine status badge

Tests

• backend/tests/test_endpoints.py
  • Added tests for:
    • AI-generated suggestion replacement
    • Fallback to rule-based suggestions on provider failure
    • Cache reuse behavior

---

Checklist

• I have read CONTRIBUTING.md
• My branch is up to date with main
• I have run pytest -v and all tests pass
• I have not introduced duplicate issues or features
• My PR title follows the format: feat/fix/docs/test: short description
• I have added tests for new features
• No hardcoded secrets or API keys are included
• This PR is linked to a GSSoC 2026 issue

---

How to test locally

1. Configure environment variables

Create a .env file in the project root:

LLM_ENABLED=true
LLM_API_KEY=<your_api_key>

# Example
LLM_BASE_URL=https://openrouter.ai/api/v1
LLM_MODEL=<provider-specific-model>

## How to test

1. Add to `.env`:

```env
LLM_ENABLED=true
LLM_API_KEY=<your_key>
LLM_BASE_URL=https://openrouter.ai/api/v1
LLM_MODEL=<model_name>

2. Start backend:

py -3.11 -m uvicorn app.main:app --reload --host 127.0.0.1 --port 8010

3. Check health:

http://127.0.0.1:8010/health

4. Open Swagger:

http://127.0.0.1:8010/docs

Test POST /debugging/ with:

{
  "code": "eval('2+2')"
}

Expected:

• "ai_generated": true
• "ai_cached": false on first request
• "ai_cached": true on repeated request

5. Run tests:

py -3.11 -m pytest -v

Expected:

63 passed