maryam-batool-sleepy-cat/semiconductor-platform

Purpose and entry points

This is a semiconductor-fabrication operations platform for wafer production, equipment health, maintenance, yield analytics, and executive/admin views. Runtime entry points are the FastAPI service at backend/app/main.py, the Vite/React client at frontend/src/main.tsx, and simulated plant telemetry at data-simulator/simulate.py. Local composition is defined by docker-compose.yml.

Web client

The TypeScript React SPA initializes through frontend/src/main.tsx and composes application navigation and role-oriented views in frontend/src/App.tsx. Domain screens—equipment health, wafer lifecycle/production, maintenance, yield, and dashboards—form the user-facing boundary; their API calls are centralized in frontend/src/services/api.ts. The browser communicates with the backend HTTP API rather than the database or industrial protocols directly.

API and application boundary

FastAPI is assembled in backend/app/main.py, which exposes API versioned endpoint modules under backend/app/api/v1/endpoints/. Endpoint ownership follows the manufacturing domains: equipment, wafers, maintenance, yield analytics, dashboard, and ISA-95. API request/response contracts reside in backend/app/schemas/, separating external representations from persistence models. backend/app/core/security.py is the security boundary for protected API behavior.

Manufacturing domain and persistence

The relational manufacturing model is defined in database/schema.sql and represented in backend ORM/domain modules such as backend/app/models/equipment.py, backend/app/models/wafer.py, and backend/app/models/maintenance.py. backend/app/core/database.py owns database connectivity. Core flows persist equipment state and telemetry-derived observations, wafer production/lifecycle records, maintenance activity, and yield metrics; API endpoints read or mutate these records for the UI and analytics views.

Industrial telemetry and scheduled processing

Plant-side integration is abstracted by backend/app/services/opcua.py, making OPC UA the backend-facing industrial protocol boundary. The simulator at data-simulator/simulate.py provides development/runtime telemetry, with MQTT infrastructure configured in mosquitto/config/mosquitto.conf. Airflow defines a separate scheduled equipment-monitoring workflow in airflow/dags/equipment_monitoring.py; this is the batch/orchestration boundary distinct from request-time FastAPI operations.

Runtime infrastructure and observability

The material runtime stack includes Python/FastAPI, Node-based Vite/React, PostgreSQL, Redis, MQTT/Mosquitto, Airflow, Prometheus, and Grafana. Container topology is declared in docker-compose.yml; the backend image is built from backend/Dockerfile. Prometheus scraping configuration is monitoring/prometheus/prometheus.yml, while Grafana alerting assets reside in monitoring/grafana/alerting/1/. Packaged Grafana plugins are vendor artifacts, not application components.

Deployment automation

Kubernetes deployment manifests are in k8s/deployment.yaml. Infrastructure provisioning is represented by terraform/main.tf; host/service deployment automation is in ansible/deploy.yml. CI/CD orchestration is configured in .github/workflows/ci-cd.yml. These layers deploy and operate the service stack without changing the API, telemetry, or domain-data boundaries.