← All projects// 07 — UML modelling, design and development

G-SERVICES — geolocated services and products

Delivered

2026

A platform to manage geolocated services and products: a full administration back-office (RBAC) and a cinematic public front, built on a UML model of about 28 classes.

Screenshot of the G-SERVICES — geolocated services and products website
Context

A complete UML model — about 28 classes, 2 packages, 4 association classes — had to become a real, manageable and tested application.

Mission

Implement the whole model in the database, in JPA and in services, manage it through a web interface and expose a public front.

Role

UML modelling, design and development

Technologies
  • Java
  • Spring Boot
  • Spring Security
  • JPA/Hibernate
  • PrimeFaces
  • PostgreSQL
  • Flyway
  • Docker

Technical implementation

  • Stack

    Java 21, Spring Boot 3.3; Jakarta Faces (Mojarra) and PrimeFaces 14 interface via JoinFaces.

  • UML modelling

    Class diagram of about 28 classes in 2 packages (GestServices, GestUser) with association classes: geography (continent to city), services, products and articles, batch stock, people, roles, profiles and permissions.

  • Persistence

    Spring Data JPA / Hibernate 6; PostgreSQL 16; 19 Flyway migrations; DTO mapping with MapStruct.

  • Security

    Spring Security 6: roles and profiles, per-person permission overrides, account locking.

  • Tests

    20 unit tests on business rules and a Testcontainers integration test that applies every migration.

  • Delivery

    One command is enough: docker compose up --build (application, PostgreSQL, pgAdmin).

  • Mobile

    Flutter client as a complement (mobile folder of the repository).

Actions
  • I modelled the domain in UML then implemented every class in JPA, with versioned Flyway migrations.
  • I built the back-office with Jakarta Faces and PrimeFaces, secured by Spring Security (roles and per-person permission overrides).
  • I built the animated public front and the business rules: net price, stock by batch, moderated reviews.
  • I tested the business rules (JUnit 5) and the full startup (Testcontainers), then containerised everything with Docker Compose.
Features
  • Full back-office with roles and permissions
  • Catalogue of geolocated services and products
  • Stock by batch and threshold alerts
  • Promotions and net price
  • Moderated reviews
  • Animated public front
Results
  • UML classes implemented
  • Flyway migrations
  • unit tests + 1 integration test
ArchitectureJSF · PrimeFacesSpring Boot · SecurityPostgreSQL 16 · FlywayDocker Compose
  • JSF · PrimeFaces → Spring Boot · Security
  • Spring Boot · Security → PostgreSQL 16 · Flyway
  • Docker Compose → Spring Boot · Security