🌒 Eclipsera

Modern Machine Learning, Built from Scratch

Version 1.2.0 • 68+ Algorithms • 100% Python • Scikit-learn Compatible

📖 Documentation • 🚀 Quick Start • 📦 Installation • 🤝 Contributing

---

🎯 About Eclipsera

Eclipsera is a comprehensive machine learning framework built from the ground up in Python. Designed for researchers, developers, and data scientists who need a unified, powerful, and transparent ML ecosystem without the complexity of heavy dependencies.

✨ Key Highlights

• 🧠 68+ Algorithms - From classical ML to modern AutoML
• 🔧 Drop-in Compatible - 100% Scikit-learn API compatibility
• ⚡ Performance Optimized - FastAlloc object pooling for 5-15% speedup
• 🛡️ Production Ready - 88% test coverage with 618 passing tests
• 🔍 Explainable AI - Built-in model interpretation tools
• 🎨 Modern Python - Full type hints, Python 3.11+ support

---

🚀 Quick Start

Installation

# Core package
pip install eclipsera

# With performance optimizations (recommended)
pip install eclipsera[perf]

# For plotting and visualization
pip install eclipsera[plot]

# Everything
pip install eclipsera[all]

First Steps

import numpy as np
from eclipsera.ml import RandomForestClassifier
from eclipsera.model_selection import train_test_split

# Generate sample data
X = np.random.randn(150, 4)
y = np.random.randint(0, 3, 150)
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)

# Train and evaluate
model = RandomForestClassifier(n_estimators=100, random_state=42)
model.fit(X_train, y_train)
score = model.score(X_test, y_test)

print(f"✅ Accuracy: {score:.3f}")

---

🌟 Advanced Features

🤖 AutoML - Automatic Algorithm Selection

from eclipsera.automl import AutoClassifier

# Let Eclipsera find the best algorithm
auto_clf = AutoClassifier(cv=5, verbose=1)
auto_clf.fit(X_train, y_train)

print(f"🏆 Best: {auto_clf.best_algorithm_} (score: {auto_clf.best_score_:.4f})")

🔍 Model Explainability

from eclipsera.explainability import permutation_importance

# Understand what drives your model
result = permutation_importance(model, X_test, y_test, n_repeats=10)
for i, importance in enumerate(result['importances_mean']):
    print(f"📊 Feature {i}: {importance:.4f}")

🔗 Complete ML Pipeline

from eclipsera.pipeline import Pipeline
from eclipsera.preprocessing import StandardScaler
from eclipsera.feature_selection import SelectKBest
from eclipsera.decomposition import PCA
from eclipsera.ml import LogisticRegression

# Build sophisticated workflows
pipe = Pipeline([
    ('scaler', StandardScaler()),
    ('selector', SelectKBest(k=20)),
    ('pca', PCA(n_components=10)),
    ('classifier', LogisticRegression())
])

pipe.fit(X_train, y_train)
print(f"🎯 Pipeline Score: {pipe.score(X_test, y_test):.4f}")

---

📚 Algorithm Library

🎓 Supervised Learning (28 Algorithms)

• Linear Models: LogisticRegression, LinearRegression, Ridge, Lasso
• Tree-Based: DecisionTree, RandomForest, GradientBoosting
• Support Vector Machines: SVC, SVR (linear, rbf, poly, sigmoid)
• Naive Bayes: GaussianNB, MultinomialNB, BernoulliNB
• Nearest Neighbors: KNeighborsClassifier, KNeighborsRegressor
• Neural Networks: MLPClassifier, MLPRegressor

🎯 Clustering (7 Methods)

• K-Means (Standard & MiniBatch)
• DBSCAN (Density-based)
• Agglomerative (4 linkage methods)
• Spectral (RBF & k-NN affinity)
• MeanShift (Kernel density)
• Gaussian Mixture (Probabilistic EM)

📉 Dimensionality Reduction

• PCA, TruncatedSVD, NMF

🗺️ Manifold Learning

• t-SNE, Isomap, LocallyLinearEmbedding

⚙️ Preprocessing & Feature Selection

• Scalers: StandardScaler, MinMaxScaler, RobustScaler
• Imputers: SimpleImputer, KNNImputer
• Encoders: LabelEncoder, OneHotEncoder, OrdinalEncoder
• Selection: VarianceThreshold, SelectKBest, RFE

🤖 AutoML & Explainability

• AutoClassifier, AutoRegressor
• permutation_importance, partial_dependence
• plot_partial_dependence, get_feature_importance

---

📊 Project Metrics

Metric	Value
🎯 Total Algorithms	68+
📝 Lines of Code	~10,500
✅ Test Coverage	88%
🧪 Tests Passing	618/618
📦 Python Version	3.11+
🔗 Dependencies	NumPy, SciPy, Pandas, Joblib

---

🔄 Version History

🛡️ Version 1.2.0 - Security Hardening (Current)

Security & Code Quality Focus

• ✅ Security hardening with pickle deserialization controls
• ✅ GitHub Actions supply chain hardening (pinned SHAs)
• ✅ Added Bandit and detect-secrets security scanning
• ✅ Fixed 522 flake8 linting errors
• ✅ SBOM generation for supply chain transparency
• ✅ Enhanced CLI security with confirmation gates

📈 Previous Versions

• v1.1.0 - Performance optimizations with FastAlloc
• v1.0.0 - Initial stable release

📋 Full Changelog

---

🛠️ Development

Setup for Contributors

git clone https://github.com/tiverse/eclipsera.git
cd eclipsera
pip install -e ".[dev,perf]"
pytest tests/

Contributing Guidelines

We welcome all contributions! Here's how you can help:

• 🐛 Report bugs - Found an issue? Open an issue
• ✨ Request features - Have an idea? Share it
• 📝 Improve docs - Help others understand Eclipsera
• 🧪 Add tests - Improve our coverage
• 💻 Write code - Add algorithms or optimize existing ones

---

📄 License & Citation

License

This project is licensed under the MIT License - see the LICENSE file for details.

Citation

If you use Eclipsera in your research, please cite:

@software{eclipsera2024,
  title = {Eclipsera: A Modern Machine Learning Framework},
  author = {Roy, Eshan},
  year = {2024},
  url = {https://github.com/tiverse/eclipsera},
  version = {1.2.0}
}

---

🔗 Useful Links

---

🌟 Why Eclipsera?

🎯 Comprehensive	🔧 Compatible	🛡️ Reliable
68 algorithms covering all major ML workflows	100% Scikit-learn compatible API	88% test coverage with 618 passing tests

🎨 Modern	⚡ Lightweight	🔍 Transparent
Built for Python 3.11+ with complete type hints	Minimal dependencies - only NumPy, SciPy, Pandas, Joblib	Clear, documented implementations

---

Built with ❤️ by Eshan Roy

Empowering the next generation of machine learning applications

⭐ If you find Eclipsera useful, consider giving it a star! ⭐