Skip to content

Process-based Deployment

This guide explains how to run Data-Continuum services as standalone OS processes. This is the simplest way to run the project without Docker or Kubernetes.

Prerequisites

  • Python 3.12+
  • Node.js
  • Local instances of PostgreSQL, MongoDB, and Redis.

Setup

  1. Install Python Dependencies: 

    uv venv
source .venv/bin/activate  # Or .\.venv\Scripts\activate on Windows
uv pip install -r api/requirements.txt
uv pip install -r ml/requirements.txt

  2. Install UI Dependencies: 

    cd ui
npm install
cd ..

Running Services

Using Helper Scripts (Windows)

If you are on Windows, you can use the PowerShell scripts in the scripts/ directory:

.\scripts\start_all.ps1

Manual Execution

1. Start the API

export POSTGRES_URL="postgresql+asyncpg://postgres:postgres@localhost:5432/continuum"
export MONGO_URL="mongodb://admin:admin@localhost:27017/"
cd api
uvicorn main:app --reload --port 8000

2. Start the ML Service

export MLFLOW_TRACKING_URI="http://localhost:5000"
cd ml
uvicorn main:app --reload --port 8001

3. Start the UI

cd ui
npm run dev

Configuration

Services use environment variables for configuration. You can also create a .env file in the root directory, though the current implementation primarily relies on direct environment variables or defaults in the code.