zephyrus2822/sih2025

Purpose and Entry Points

This is a railway operations platform with role-specific web dashboards, passenger services, predictive maintenance/ML, complaints, chatbot support, and automated engineer job assignment. It consists of a Next.js frontend and Python backend. Main entry points are src/app/page.tsx, src/app/layout.tsx, and src/backend/main.py.

Frontend Application Shell and Localization

The frontend uses the Next.js App Router. Global layout and styling originate at src/app/layout.tsx and src/app/globals.css. Locale-scoped pages use src/app/[locale]/layout.tsx and src/app/[locale]/page.tsx, backed by translation assets in public/locales/en/common.json and public/locales/hi/common.json. Shared i18n setup is in src/components/I18nProvider.tsx and src/lib/i18n.ts.

Role-Oriented Operational Portals

Separate route groups expose tailored interfaces for employees, engineers, locomotive pilots, maintainers, and passengers. Each dashboard generally has a server route page that mounts a client-side interaction component:

• Engineer work management: src/app/engineer_dashboard/page.tsx, src/app/engineer_dashboard/EngineerDashboardClient.tsx

• Employee job view: src/app/employee_dashboard/page.tsx, src/app/employee_dashboard/EmployeeJobsClient.tsx

• Loco-pilot and maintainer views: src/app/locopilot_dashboard/LocoPilotDashboardClient.tsx, src/app/maintainer_dashboard/MaintainerDashboardClient.tsx

• Passenger-facing services: src/app/passenger_dashboard/page.tsx, src/app/passenger_dashboard/complaints/page.tsx, src/app/passenger_dashboard/schedule/page.tsx

Backend API Boundary

The Python service is organized by domain routers registered from src/backend/main.py. Domain APIs isolate authentication, chat, complaints, personnel workflows, contact/footer data, and pilot workflows:

• src/backend/auth/routes.py

• src/backend/chatbot/routes.py

• src/backend/complaints/routes.py

• src/backend/engineer/routes.py

• src/backend/locopilot/routes.py

The frontend’s client components are the primary consumer boundary for these APIs; API implementation, persistence, and business rules remain under src/backend.

Persistence and Validation

Relational database connectivity and ORM-style data models are centralized in src/backend/db/database_connection.py and src/backend/db/database_models.py. The validation domain has its own data-access, models, schemas, routes, and service layer, separating request contracts from validation business logic: src/backend/validation/routes.py, src/backend/validation/service.py, and src/backend/validation/db.py. SQL bootstrap/snapshot assets are provided by backup.sql and backup_clean.sql.

Predictive Model Pipeline

The deep-learning subsystem serves model-driven predictions through src/backend/dl_model/routes.py. It contains request/database schemas, model persistence, data import, training, metrics, and reporting stages:

• Model artifacts/implementation: src/backend/dl_model/model.h5, src/backend/dl_model/model/model.py

• Training and evaluation: src/backend/dl_model/services/training_service.py, src/backend/dl_model/services/metrics.py

• Dataset ingestion/reporting: src/backend/dl_model/scripts/import_csv.py, src/backend/dl_model/services/report_service.py

Its UI is mounted at src/app/predictor/page.tsx and src/app/predictor/PredictorClient.tsx.

Asynchronous Job Distribution

Engineer-job allocation is a distinct asynchronous subsystem. API requests enter through src/backend/job_distributor/api.py; assignment decisions are implemented in src/backend/job_distributor/assigner.py; work is dispatched and executed through Celery configuration/tasks in src/backend/job_distributor/celery_app.py and src/backend/job_distributor/tasks.py. Worker deployment is represented by src/backend/job_distributor/worker_entrypoint.sh. This boundary prevents allocation work from blocking normal API requests.

Runtime, Deployment, and Observability

The frontend runs in the Node/Next.js ecosystem (package.json, next.config.ts). The backend is containerized with src/backend/Dockerfile and Python dependencies in src/backend/requirements.txt; compose definitions exist at docker-compose.yml and src/backend/backend-docker-compose.yml. Monitoring infrastructure uses Prometheus and Grafana assets in observibility/prometheus.yml, observibility/docker-compose.yml, and observibility/grafana_dashboard.json. The frontend includes a Grafana embedding component at src/components/ui/grafanaembed.tsx.