XGBoost API

Flask API for XGBoost ML predictions.

What is this?

This repository provides a machine learning API using XGBoost, a popular gradient boosting library. It trains a model on sample data and serves predictions via HTTP endpoints. XGBoost is used for tasks like classification or regression, making this a machine learning application.

Keywords: python, flask, xgboost, machine-learning, api, docker, kubernetes, gpu

Quick Start

1. Install dependencies:

   pip install -r requirements.txt

2. Train the model:

   python train.py

3. Run tests:

   pip install -r requirements.txt
   pytest

4. Run the API:

   python app.py

Conventional Commits

This project follows conventional commit standards for commit messages.

Setup

To enable commit message validation:

cp scripts/commit-msg .git/hooks/commit-msg
chmod +x .git/hooks/commit-msg

Commit Message Format

Commit messages must:

• Start with a type: feat:, fix:, docs:, style:, refactor:, test:, chore:, perf:, ci:, build:, revert:
• Be entirely lowercase
• Have first line ≤60 characters

Example: feat: add user authentication endpoint

Rewriting History

To clean up existing commit messages:

git filter-branch --msg-filter 'bash scripts/rewrite_msg.sh' -- --all
git push --force-with-lease

API Endpoints

• GET /health: Health check
• POST /predict: Make predictions
  • Body: {"features": [array_of_features]}
  • Response: {"prediction": [values], "shape": [dims]}

Usage Examples

Health check:

curl http://localhost:8000/health

Prediction:

curl -X POST http://localhost:8000/predict \
  -H "Content-Type: application/json" \
  -d '{"features": [1.0, 2.0, 3.0, 4.0]}'

Response: {"id":"uuid","prediction":[0.99...],"shape":[1,4]}

Requirements

• Python 3.9+ (due to scikit-learn 1.5.0 requirements)
• Optional: NVIDIA GPU with CUDA support
• For GPU: Docker with NVIDIA Container Toolkit, Kubernetes with NVIDIA device plugin

Mac Setup

Install OpenMP:

brew install libomp
export LDFLAGS="-L/opt/homebrew/opt/libomp/lib"
export CPPFLAGS="-I/opt/homebrew/opt/libomp/include"

Deployment

Docker:

docker build -t bniladridas/flask-xgboost-api .
docker run --gpus all -p 8000:8000 bniladridas/flask-xgboost-api

Test:

1. docker run -p 8000:8000 bniladridas/flask-xgboost-api
2. curl http://localhost:8000/health → {"status":"healthy"}
3. curl -X POST http://localhost:8000/predict -H "Content-Type: application/json" -d '{"features": [1.0, 2.0, 3.0, 4.0]}' → prediction JSON

Kubernetes:

# Install minikube: https://minikube.sigs.k8s.io/docs/start/
minikube start
kubectl apply -f kubernetes/

Requires a Kubernetes cluster (group of nodes for container management).