pro158127/gastro-ai

How pro158127/gastro-ai works

Gastro-AI is a clinical-image analysis application with JWT login, asynchronous analysis, and live result updates. A doctor signs in, uploads an image with patient-document details, and receives a processing record; a Celery task performs the scaffold’s stub inference, creates a Grad-CAM placeholder, requests an LLM explanation, persists the result, and publishes completion for the dashboard to retrieve and display. The README also documents polling as an alternative to WebSockets. AI integrations are currently stubs; unsampled modules are represented only where documented, so some links rely on README evidence.

Main entry points: tasks.py Analysis task · analisis_service.py Analysis service · main.py FastAPI entry · Dashboard.tsx Analysis dashboard

Components 21

Web client

App routing frontend/src/App.tsx
Login page frontend/src/pages/Login.tsx
Analysis dashboard frontend/src/pages/Dashboard.tsx
HTTP API client frontend/src/services/api.ts
WebSocket client frontend/src/services/websocket.ts

API and identity

FastAPI entry backend/app/main.py
Authentication routes backend/app/routers/auth.py
JWT security backend/app/security.py
Analysis routes backend/app/routers/analisis.py
WebSocket routes backend/app/routers/ws.py

Analysis processing

Analysis service backend/app/services/analisis_service.py
Analysis task backend/app/tasks.py
Celery worker backend/app/worker.py

Persistence and messaging

PostgreSQL
Redis
MinIO object store
ORM models backend/app/models/analisis.py

AI and object storage

LLM adapter backend/app/adapters/llm_adapter.py
MinIO adapter backend/app/adapters/minio_adapter.py

Other components

Doctor / Admin
Nginx gateway nginx.conf

Connections 31

Doctor / Admin → Nginx gateway: uses web README.md
Nginx gateway → App routing: serves SPA README.md
Nginx gateway → FastAPI entry: proxies API README.md
Nginx gateway → WebSocket routes: proxies WebSocket README.md
App routing → Login page: renders frontend/src/App.tsx
App routing → Analysis dashboard: renders frontend/src/App.tsx
Login page → HTTP API client: calls frontend/src/pages/Login.tsx
Analysis dashboard → HTTP API client: calls frontend/src/pages/Dashboard.tsx
Analysis dashboard → WebSocket client: subscribes frontend/src/pages/Dashboard.tsx
HTTP API client → FastAPI entry: sends HTTP README.md
WebSocket client → WebSocket routes: connects README.md
FastAPI entry → Authentication routes: mounts routes backend/app/main.py
FastAPI entry → Analysis routes: mounts routes backend/app/main.py
FastAPI entry → WebSocket routes: mounts routes backend/app/main.py
Authentication routes → PostgreSQL: queries users backend/app/routers/auth.py
Authentication routes → JWT security: verifies and signs backend/app/routers/auth.py
Analysis routes → Analysis service: delegates backend/app/routers/analisis.py
Analysis service → MinIO adapter: stores image backend/app/services/analisis_service.py
Analysis service → PostgreSQL: creates and reads backend/app/services/analisis_service.py
Analysis service → Analysis task: enqueues backend/app/services/analisis_service.py
Analysis service → ORM models: uses model backend/app/services/analisis_service.py
Analysis task → Celery worker: uses app backend/app/tasks.py
Analysis task → MinIO adapter: stores Grad-CAM backend/app/tasks.py
Analysis task → LLM adapter: requests explanation backend/app/tasks.py
Analysis task → PostgreSQL: updates result backend/app/tasks.py
Analysis task → Redis: publishes event backend/app/tasks.py
WebSocket routes → Redis: forwards events README.md
MinIO adapter → MinIO object store: reads and writes backend/app/adapters/minio_adapter.py
Analysis routes → PostgreSQL: uses session backend/app/routers/analisis.py
Analysis routes → ORM models: uses model backend/app/routers/analisis.py
Analysis dashboard → HTTP API client: polls status README.md

Drawn by GitDiagram on November 10, 2026 from the repository's file tree, README and source files. Read as Markdown ·