How lunka81/simsssg works
Smart Access is a prototype face-verification gate: an edge device captures a live face, computes an embedding, asks the backend for an access decision, and actuates a lock. The backend also supports admin authentication, employee enrollment and management, access logs, and a live recognition feed to the React application. The graph includes the separately documented offline model-evaluation capability. Some backend and frontend connections are documented in the README or source index, but their implementation details were not sampled.
Main entry points: database.py PostgreSQL and pgvector · recognition_service.py Face, liveness, embedding · recognition_routes.py Identify endpoint · app.py API application
Components 32
Edge access device
- Camera capture
edge-device/camera/camera.py - Face, liveness, embedding
edge-device/services/recognition_service.py - Edge vision models
edge-device/recognition - Decision uploader
edge-device/clients/uploader.py - Lock and LEDs
edge-device/hardware/lock.py - Simulated lock
edge-device/hardware/fake_lock.py - Device configuration
edge-device/config/config.py
Backend and recognition
- API application
backend/app/app.py - Identify endpoint
backend/routes/recognition_routes.py - Match and decision
backend/services/comparison_services.py - Employee enrollment
backend/services/employee_services.py - Image embedding
backend/services/img_services.py - Employee API
backend/routes/employee_routes.py - Admin authentication
backend/services/auth_services.py - Admin login API
backend/routes/admin_login_routes.py - Access log API
backend/routes/employee_log_routes.py - Access log service
backend/services/employee_log_services.py - Recognition WebSocket
backend/services/recognition_socket.py - PostgreSQL and pgvector
backend/db/database.py - Employee records
backend/models/employee.py - Admin records
backend/models/admin.py - Access log records
backend/models/employee_log.py
Admin web application
- Web app routes
frontend/src/App.tsx - Live access view
frontend/src/pages/Home - Admin console
frontend/src/pages/Admin - Frontend auth
frontend/src/auth.ts - Employee API client
frontend/src/api/employees.ts
Model evaluation
- Evaluation pipeline
model-evaluation/pipeline - Evaluation models
model-evaluation/pipeline/models.py - Evaluation presentation
model-evaluation/presentation
Other components
- Gate visitor
- Model evaluator
Connections 39
- Gate visitor → Camera capture: approaches gate
README.md - Camera capture → Face, liveness, embedding: supplies frames
edge-device/services/recognition_service.py - Face, liveness, embedding → Edge vision models: uses models
edge-device/services/recognition_service.py - Face, liveness, embedding → Decision uploader: sends embedding
edge-device/services/recognition_service.py - Decision uploader → Identify endpoint: posts embedding
README.md - Identify endpoint → Match and decision: requests decision
backend/routes/recognition_routes.py - Match and decision → PostgreSQL and pgvector: searches matches
README.md - Match and decision → Access log service: logs approval
backend/services/comparison_services.py - Identify endpoint → Recognition WebSocket: pushes result
backend/routes/recognition_routes.py - Identify endpoint → Employee records: loads matched employee
backend/routes/recognition_routes.py - Recognition WebSocket → Live access view: streams recognition
README.md - Identify endpoint → Lock and LEDs: returns decision
README.md - Face, liveness, embedding → Lock and LEDs: actuates decision
edge-device/services/recognition_service.py - Face, liveness, embedding → Simulated lock: selects in demo mode
edge-device/services/recognition_service.py - Device configuration → Face, liveness, embedding: configures runtime
edge-device/services/recognition_service.py - API application → Identify endpoint: registers route
backend/app/app.py - API application → Employee API: registers route
backend/app/app.py - API application → Admin login API: registers route
backend/app/app.py - API application → Access log API: registers route
backend/app/app.py - Employee API → Employee enrollment: delegates operations
backend/routes/employee_routes.py - Employee enrollment → Image embedding: extracts embedding
backend/services/employee_services.py - Employee enrollment → PostgreSQL and pgvector: stores employees
backend/services/employee_services.py - Image embedding → Edge vision models: uses backend models
- Employee API → Admin authentication: requires token
backend/routes/employee_routes.py - Admin login API → Admin authentication: handles login
backend/routes/admin_login_routes.py - Admin authentication → PostgreSQL and pgvector: reads admin
backend/services/auth_services.py - Access log API → Access log service: delegates log access
backend/routes/employee_log_routes.py - Access log service → PostgreSQL and pgvector: reads and writes logs
backend/services/employee_log_services.py - PostgreSQL and pgvector → Employee records: stores
backend/models/employee.py - PostgreSQL and pgvector → Admin records: stores
backend/models/admin.py - PostgreSQL and pgvector → Access log records: stores
backend/models/employee_log.py - Web app routes → Live access view: routes to
frontend/src/App.tsx - Web app routes → Admin console: routes to
frontend/src/App.tsx - Admin console → Employee API client: uses employee API
- Employee API client → Employee API: calls REST API
- Frontend auth → Admin login API: authenticates via REST
README.md - Model evaluator → Evaluation pipeline: runs evaluation
README.md - Evaluation pipeline → Evaluation models: uses model definitions
model-evaluation/pipeline/__main__.py - Evaluation pipeline → Evaluation presentation: presents results