ml-robust-eval 0.1.7

ML evaluation, validation, and test case generation toolkit with production monitoring.

ML Robust Eval

---

ML Eval Robust is a pure Python, object-oriented library for comprehensive machine learning model evaluation, validation, and robustness testing.
It's an all-in-one toolkit that features:

• 📊 Metrics for classification, regression, NLP, and computer vision tasks
• 🔁 Cross-validation and A/B testing helpers
• 📈 Visualization tools for confusion matrices and ROC curves (stdout-based, no dependencies!)
• 🦾 Automated test case generation: edge cases, adversarial samples, and boundary value tests
• 📡 Production Monitoring (optional): Real-time model degradation tracking with Prometheus & Grafana
• 🧩 Zero dependencies (core) – works anywhere Python runs!

---

🚀 Installation

Basic installation:

pip install ml_robust_eval

With production monitoring support (Prometheus/Grafana):

pip install ml_robust_eval[monitoring]

---

🧠 Features

• Classification, Regression, NLP, and CV Metrics
  • Accuracy, Precision, Recall, F1, MAE, MSE, R², BLEU, IoU, and more!
• Cross-Validation & A/B Testing
  • K-fold splitting, group comparison, and statistical difference calculation
• Visualization
  • Confusion matrices and ROC curves printed directly to your console
• Robustness Test Case Generation
  • Edge, boundary, and adversarial sample generation for any tabular data
• Production Monitoring (Optional)
  • Real-time model performance tracking with Prometheus
  • Automatic degradation detection compared to baseline
  • Pre-configured Grafana dashboards
  • Docker Compose setup for easy deployment
• Zero Dependencies (Core)
  • Core library uses only standard library, OOP-based, and lightweight
  • Monitoring module is optional and requires prometheus-client

---

📚 Documentation

• API Reference
• Examples & Tutorials
• Production Monitoring Guide - Real-time model degradation tracking

---

💡 Why ML Eval Robust?

• Universal: No dependencies, works in any Python environment
• Educational: Clear, readable OOP code for learning and teaching
• Robust: Covers the full ML evaluation and validation pipeline, including adversarial and edge testing
• Production-Ready: Optional monitoring module for real-time model tracking in production

---

🤝 Contributing

All contributions, bug reports, and suggestions are welcome!
See the contributing guide.

---

📜 License

MIT License

---

📬 Contact

Questions? Open an issue or reach out at [[e-mail]].

---

Let your models earn their confidence. Test, validate, and challenge them with ML Robust Eval!