Frontend:
- Add loading guard to AI button disabled state
- Add re-entry guard at top of startMatching and startAIEnhance
Backend:
- Add maxMatchDuration (10 min) with periodic timeout checks
- Check timeout in outer A-loop (every row) and inner B-loop (every 500 iter)
- Reduce HTTP client timeout from 60s to 30s for API calls
Backend:
- Add apiEndpoint / apiModel fields to App struct (configurable, with Deepseek defaults)
- New SetAIConfig(endpoint, model, key) unified config method
- New SetAPIKey(key) for key-only updates
- GetAIStatus() returns { ready, endpoint, model }
- callDeepseekAPI → callAIAPI with configurable endpoint and model
- Remove hardcoded Deepseek URL/model; defaults remain as fallback
Frontend:
- Replace Deepseek-specific state with generic AI config (endpoint / model / key)
- Add endpoint URL and model name input fields
- saveApiKey → saveApiConfig using SetAIConfig
- onMounted restores full AI config via GetAIStatus
- Rename all user-facing 'Deepseek' labels to 'AI / AI API'
Now supports: Deepseek, OpenAI, Ollama, vLLM, or any OpenAI-compatible endpoint.
- Added frontend runtime JavaScript functions for logging, window management, and notifications.
- Created Go module with dependencies for Wails and Excel processing.
- Implemented main application entry point with embedded frontend assets.
- Configured Wails application settings in wails.json.
Co-authored-by: Copilot <copilot@github.com>
