How divyaprasad-sahu/radi-ai-buddy-medical works
Radi-AI Buddy is a multilingual educational chest-X-ray screening demo: users upload an image, receive a validated-model screening result, and can discuss it with an API-backed assistant or review reports and aggregate research metrics. The React workspace calls a FastAPI service; serving validates checksum-verified ONNX artifacts, while NIH data cleaning and model training are offline research workflows. Edges are limited to relationships shown in excerpts, the source index, or README; unsampled frontend and offline wiring is not inferred.
Main entry points: Dashboard.tsx Workspace dashboard · main.py FastAPI endpoints · api.ts API client · model.py ONNX validation
Components 20
Screening workspace
- React app
AOML Deploy/radi-ai-buddy-main/radi-ai-buddy-main/src/App.tsx - Workspace dashboard
AOML Deploy/radi-ai-buddy-main/radi-ai-buddy-main/src/components/Dashboard.tsx - Image upload
AOML Deploy/radi-ai-buddy-main/radi-ai-buddy-main/src/components/UploadComponent.tsx - Screening result
AOML Deploy/radi-ai-buddy-main/radi-ai-buddy-main/src/components/ResultCard.tsx - Assistant chat
AOML Deploy/radi-ai-buddy-main/radi-ai-buddy-main/src/components/Chatbot.tsx - Session history
AOML Deploy/radi-ai-buddy-main/radi-ai-buddy-main/src/components/HistoryPanel.tsx - Report preview
AOML Deploy/radi-ai-buddy-main/radi-ai-buddy-main/src/components/ReportBuilder.tsx - Research panel
AOML Deploy/radi-ai-buddy-main/radi-ai-buddy-main/src/components/ResearchPanel.tsx - Translations
AOML Deploy/radi-ai-buddy-main/radi-ai-buddy-main/src/lib/i18n.ts - API client
AOML Deploy/radi-ai-buddy-main/radi-ai-buddy-main/src/lib/api.ts
API service
- FastAPI endpoints
AOML Deploy/backend/app/main.py - Localized summaries
AOML Deploy/backend/app/chatbot.py - Research insights
AOML Deploy/backend/app/main.py
Model inference
- Image inference
AOML Deploy/backend/app/predict.py - ONNX validation
AOML Deploy/backend/app/model.py
Offline research
- Dataset cleaning
AOML Deploy/backend/clean_nih.py - Model training
AOML Deploy/backend/train.py
External services
- Groq assistant
- NIH dataset
Other components
- User
Connections 22
- User → React app: uses
- React app → Workspace dashboard: routes to
- Workspace dashboard → Image upload: renders
AOML Deploy/radi-ai-buddy-main/radi-ai-buddy-main/src/components/Dashboard.tsx - Workspace dashboard → Screening result: renders
AOML Deploy/radi-ai-buddy-main/radi-ai-buddy-main/src/components/Dashboard.tsx - Workspace dashboard → Assistant chat: renders
AOML Deploy/radi-ai-buddy-main/radi-ai-buddy-main/src/components/Dashboard.tsx - Workspace dashboard → Session history: renders
AOML Deploy/radi-ai-buddy-main/radi-ai-buddy-main/src/components/Dashboard.tsx - Workspace dashboard → Report preview: renders
AOML Deploy/radi-ai-buddy-main/radi-ai-buddy-main/src/components/Dashboard.tsx - Workspace dashboard → Research panel: renders
AOML Deploy/radi-ai-buddy-main/radi-ai-buddy-main/src/components/Dashboard.tsx - Image upload → API client: submits image
- Assistant chat → API client: sends chat
AOML Deploy/radi-ai-buddy-main/radi-ai-buddy-main/src/components/Chatbot.tsx - Research panel → API client: loads insights
AOML Deploy/radi-ai-buddy-main/radi-ai-buddy-main/src/components/ResearchPanel.tsx - API client → FastAPI endpoints: calls service
- FastAPI endpoints → Image inference: runs inference
AOML Deploy/backend/app/main.py - Image inference → ONNX validation: loads model
AOML Deploy/backend/app/predict.py - FastAPI endpoints → ONNX validation: checks metadata
AOML Deploy/backend/app/main.py - FastAPI endpoints → Localized summaries: formats result
AOML Deploy/backend/app/main.py - FastAPI endpoints → Groq assistant: dispatches chat
AOML Deploy/backend/app/main.py - FastAPI endpoints → Research insights: serves insights
AOML Deploy/backend/app/main.py - Dataset cleaning → NIH dataset: cleans cohort
README.md - Model training → Dataset cleaning: uses cohort
- Model training → ONNX validation: exports ONNX
- FastAPI endpoints → Research insights: reads summaries
AOML Deploy/backend/app/main.py