How ride-austin/server works
This project is a large, multi-module Maven server-side system implementing a ride-sharing backend as a set of JVM services and shared libraries. The root POM aggregates:
• Library modules (no ‘rest’ packages) providing shared domain objects, DSL repositories, utilities, event models, and generic services (common, common-rest, domain-base, domain, polygon, payments-domain, locations-domain, …).
• Service modules (contain ‘rest’, ‘jobs’, or ‘service’ packages) each exposing a narrowly scoped REST API or scheduled job via Spring/Jersey (app, admin-rest, active-drivers, avatar-management, documents, payments, promocodes, surge-pricing, notifications, reports, driver-statistics, user-tracking, etc.).
• An integration-test module (app-itests) that spins up MySQL and Redis via Docker for end-to-end test scenarios.
• A deploy module using Flyway for database migrations.
Main components
• Edge/API services:
• app (core rider/driver endpoints, authentication, ride-flow)
• admin-rest (admin dashboards, monitoring APIs)
• active-drivers, avatar-management, documents, payments, promocodes, surge-pricing, notifications, reports, driver-statistics, user-tracking, etc.
• Shared libraries: common, common-rest, domain-base, domain, payments-domain, locations-domain, events, cached-events, polygon, …
• Data stores:
• MySQL cluster (ride_db) – all services use JPA/DSL to persist domain data
• Redis cluster – session cache, location cache, surge areas, state-machine context
• Scheduler / batch jobs: either embedded (Quartz via scheduling module) or separate microservices (admin-jobs)
• Third-party integrations: Google Maps/Geocoding API, Twilio (SMS / phone lookup), Stripe / Payoneer (payments), Amazon SNS / S3, external config service, etc.
Relationships & data flows
• Clients (mobile/web) → HTTP/JSON → app or admin-rest
• app → shared libraries (common, domain) for business logic & persistence
• app / each microservice → MySQL / Redis
• app ↔ other services (e.g., ride-dispatch library, promotions, surge) via REST calls or shared message/events (Spring application events or pub-sub over Redis)
• scheduled jobs (Quartz or admin-jobs) trigger database cleanup, deactivation, report generation
• services → external HTTP → Google, Twilio, Stripe, Payoneer, AWS
• integration tests module spins up containers for MySQL/Redis and runs end-to-end scenarios
Architectural patterns
• Microservice decomposition by domain feature
• Layered architecture in each service: REST controllers → service layer → repository/DSL → domain models
• Shared library pattern for cross-cutting concerns (authentication, caching, utilities)
• Event-driven sub-system via Spring events / cached-events builder
• State machine for ride-dispatch in Redis
• Scheduled batch jobs via Quartz (scheduling module)
Significant technologies
• Java 8, Spring (Spring MVC / Jersey), JPA/Hibernate, Spring Data (DSL), Maven
• MySQL, Redis, Docker (app-itests), Flyway (deploy)
• Logback (logging), AOP for cross-cutting (app/aop), Freemarker templates (email & UI assets)