- Move gateway API interaction (health check, text generation) from service.py to gateway_wrapper.py
- service.py now only contains business logic (system prompt and thin delegation)
- app.py imports check_gateway_health from gateway_wrapper instead of service
📌Update linkedin-generator widget to use python 3.12 and add inference-gateway-api-client
- Deleted the entire inference_gateway_client module, including client, API methods, models, and error handling.
- Updated service.py to use the extracted GitHub hosted inference_gateway_api_client package for API interactions.
♻️ (linkedin-generator): Simplify widget API, use inference gateway client
- Replace custom config.py/gateway.py with auto-generated OpenAPI client
- Wire service.py to use inference_gateway_client endpoints
- Update app.py to reference new client pattern
- Remove legacy gateway and config modules
- Update dependencies and README
✨ (linkedin-generator): Add Gradio Server widget for LinkedIn post generation
- Add Gradio Server API with /health and /generate endpoints
- Add service layer with system prompt for LinkedIn-style post transformation
- Include README with API docs, env vars, and deployment instructions
- Configure HF Space deployment via space.yaml
