ArchitectureSystem Overview

System Overview

Architecture at a Glance

Lineo-PM is a full-stack web application composed of four main components, all orchestrated using Docker Compose:

┌─────────────────────────────────────────────────────────┐
│                     Browser (User)                      │
└────────────────────────┬────────────────────────────────┘
                         │ HTTP / REST
┌────────────────────────▼────────────────────────────────┐
│              Frontend (React + Vite)                    │
│                   localhost:5173                        │
└────────────────────────┬────────────────────────────────┘
                         │ HTTP / REST + JSON
┌────────────────────────▼────────────────────────────────┐
│             Backend API (FastAPI)                       │
│                   localhost:8000                        │
└──────────┬─────────────────────────┬───────────────────┘
           │ async SQLAlchemy        │ Task Queue (Redis)
┌──────────▼──────────┐   ┌──────────▼──────────────────┐
│   PostgreSQL DB     │   │   Celery Worker             │
│   (persistent data) │   │   (background simulations)  │
└─────────────────────┘   └─────────────────────────────┘

Component Descriptions

Frontend

A React + TypeScript single-page application built with Vite. It serves the Gantt chart, scenario panels, Monte Carlo results dashboard, and all user interaction. It communicates with the backend exclusively via REST API calls.

The frontend runs on localhost:5173 in development (Vite dev server with hot module replacement).

Backend API

A FastAPI application providing the REST API for all data operations: projects, tasks, milestones, scenarios, relations, and updates. It uses async SQLAlchemy to interact with PostgreSQL and exposes automatic OpenAPI documentation at /docs.

The backend runs on localhost:8000.

PostgreSQL

The primary persistent data store. All project data — schedules, scenarios, simulation results, task risk levels, relations, updates — is stored in a PostgreSQL database. Schema migrations are managed by Alembic.

Celery Worker

An asynchronous background worker that handles computationally intensive jobs — primarily Monte Carlo simulations. When a simulation is requested, the API queues a Celery task, and the worker processes it independently. Results are written back to PostgreSQL and surfaced in the frontend.

The Celery broker (message queue) is Redis, also running as a container.

How Components Interact

  1. The user interacts with the Frontend (drag a task, create a scenario, request a simulation)
  2. The frontend sends REST API requests to the Backend API
  3. The backend reads/writes data from PostgreSQL via async SQLAlchemy
  4. For simulation requests, the backend enqueues a task via Redis to the Celery Worker
  5. The worker runs the Monte Carlo simulation, writes results to PostgreSQL
  6. The frontend polls or receives notification that results are ready and displays the dashboard

Docker-Based Deployment

All components are defined in docker-compose.dev.yml. The full stack is started with:

docker compose -f docker-compose.dev.yml up -d --build

Each service runs in its own container, isolated by Docker networking, with shared volumes for data persistence. This makes the development environment fully reproducible with no host-level dependencies beyond Docker itself.

Risk AnalysisBackend