Learn OOP the Hard Way

Modeling and Implementing a Machine Learning Library
Object-Oriented Programming with Python — Université Paris-Saclay

---

⚠️ View our project report.

This project focuses on the design and implementation of a small machine learning library in Python, with an emphasis on clean, modular, and extensible software architecture.

Inspired by existing machine learning libraries such as scikit-learn, the objective is to model core components including estimators, datasets, transformers, and pipelines without relying on external machine learning frameworks.

---

Repository structure

.
├── docs/
│   ├── class_diagrams/
│   │   ├── class_diagram_v1.png        # First version of the UML class diagram
│   │   ├── class_diagram_v2.png        # Intermediate version of the UML class diagram
│   │   ├── class_diagram_v3.png        # Intermediate version of the UML class diagram
│   │   └── ClassDiagram_VFINAL.png     # Final UML class diagram used in the report
│   │
│   └── report.pdf                      # Project report (PDF)                  
|
├── upsaclay-ml/
│   ├── upsaclay_ml/
│   │   ├── models/
│   │   │   ├── estimator.py            # Base estimator abstraction
│   │   │   ├── linear_model.py         # Linear models
│   │   │   ├── tree.py                 # Tree-based models
│   │   │   └── cluster.py              # Clustering models
│   │   │
│   │   ├── datasets.py                 # Dataset loading utilities
│   │   ├── transformers.py             # Data preprocessing transformers
│   │   ├── pipeline.py                 # Pipeline implementation
│   │   ├── model_selection.py          # Train/test split utilities
│   │   ├── metrics.py                  # Evaluation metrics
│   │   └── test_dataset.csv            # Example dataset
│   │
│   ├── test_calling_pipeline.py        # Pipeline tests
│   ├── test_supervised_learning.py     # Supervised learning tests
│   ├── test_unsupervised_learning.py   # Unsupervised learning tests
│   ├── test_MNIST_TCGA.py              # MNIST and TCGA dataset tests
│   ├── test_BrunOOP.py                 # OOP behavior tests
│   └── test_BrunOOP_SK.py              # Comparison with scikit-learn
│
├── 2526-m1geniomhe-OOPwithPython-Project.pdf      # Project instructions (PDF)
├── requirements.txt                               # Project dependencies
└── README.md                                      # Project documentation

---

How to use

1. Install the required dependencies

pip install -r requirements.txt

2. Run the test scripts

The library can be tested using the provided test_*.py scripts located in the upsaclay-ml/ directory.

cd upsaclay-ml
python test_calling_pipeline.py

Other test scripts can be run in the same way to test different parts of the library.