Spaces:
Running
Running
Commit History
14 days ago
-
fix(ui): sync Today's Briefing banner on import and update BUY score scoring
Ramkumar Shanmugame06f207
-
feat: redesign performance charts, fix nifty index outlier, expand history to 3Y, and refine metrics cards
Ramkumar Shanmugamd1eff7d
-
perf: optimize briefing generation, batch db inserts, and cap history to 100 days
Ramkumar Shanmugam- Streamline Gemini Search Grounding into a targeted candidate funnel (max 3-5 searches) - Eliminate redundant Yahoo Finance downloads in generate/fetch-news endpoints - Batch PostgreSQL operations with execute_values (portfolio holdings, snapshots, trades, signals) - Prevent unawaited competing holdings fetch in frontend during generation - Cap History summaries, dates, and Performance benchmarks to the recent 100 days20822cb -
feat: optimize briefing performance with on-demand news, batch queries, and price caching
Ramkumar Shanmugam- Decouple stock catalyst news from initial briefing generation for faster load times - Add on-demand catalyst news fetching endpoint and UI action - Migrate Gemini API caller to google-genai SDK with search grounding - Implement batch FinBERT sentiment scoring - Add in-memory TTL caching for stock prices and Nifty 50 - Batch database queries for historical briefings and AI suggestions11e84fc
-
feat: add AI-powered Stock Picks tab with Gemini search grounding
Ramkumar Shanmugam- New POST /api/briefings/recommendations endpoint in briefings.py - Resolves current portfolio symbols from DB (owner) or request body (guest) - Calls Gemini with Google Search grounding to find top 5 NSE/BSE stocks - Filters out stocks already in user active portfolio - Added StockRecommendationItem and RecommendationsResponse Pydantic models in schemas.py - Added build_recommendations_prompt() and run_recommendations_analysis() in api_caller.py - Prompt enforces valuation strategy: P/E vs industry, ROE>15%, ROCE>15%, D/E<0.5, RSI<60 - Frontend App.tsx: - New StockRecommendationItem TypeScript interface - Stock Picks state and handleGetRecommendations() with 7-step animated progress - New Stock Picks tab button with AI badge (visible to all roles) - renderRecommendationsTab() with intro card, animated pulse-ring loading state, empty prompt state, responsive card grid with P/E P/B ROE D/E metric badges and expandable Why Now and Risk sections - Added pulse-ring…5c68d47
-
feat: migrate to FastAPI + React stack with PostgreSQL backend
ramkumar-bindrix- Replace Streamlit app with FastAPI backend and React/Vite frontend - Add Docker setup with Dockerfiles and docker-compose.yml - Add PostgreSQL schema and migration scripts - Implement backend routes: auth, portfolio, chat, performance, briefings - Update modules: chat, performance, price_fetcher, db - Add integration test scripts - Remove legacy Streamlit app files (app.py, renderer.py, test_api.py)049861f
-
feat: implement AI performance tracker against Nifty 50 benchmark
ramkumar-bindrixab9a404