tgmm 0.1.5__tar.gz

tgmm-0.1.5/PKG-INFO

@@ -0,0 +1,123 @@
+Metadata-Version: 2.2
+Name: tgmm
+Version: 0.1.5
+Summary: A Gaussian Mixture Model (GMM) based on Expectation-Maximisation (EM) implemented in PyTorch
+Home-page: https://github.com/adriansousapoza/TorchGMM
+Author: Your Name
+Author-email: adrian.sousapoza@gmail.com
+Classifier: Programming Language :: Python :: 3
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Requires-Python: >=3.6
+Description-Content-Type: text/markdown
+Requires-Dist: torch>=2.5.1
+Requires-Dist: numpy>=1.23.0
+Requires-Dist: matplotlib>=3.10.1
+Provides-Extra: docs
+Requires-Dist: Sphinx>=8.2.3; extra == "docs"
+Requires-Dist: Pygments>=2.19.1; extra == "docs"
+Dynamic: author
+Dynamic: author-email
+Dynamic: classifier
+Dynamic: description
+Dynamic: description-content-type
+Dynamic: home-page
+Dynamic: provides-extra
+Dynamic: requires-dist
+Dynamic: requires-python
+Dynamic: summary
+
+# TorchGMM: A Gaussian Mixture Model Implementation with PyTorch
+
+![TorchGMM logo](docs/_static/logo-no-background.png)
+
+**TorchGMM** is a flexible implementation of Gaussian Mixture Models in PyTorch, supporting:
+
+- EM Algorithm
+- MAP Estimation with Priors
+- Multiple Covariance Types
+- Various Initialization Methods
+- Comprehensive Clustering & Evaluation Metrics
+
+## Features
+
+1. **GaussianMixture**  
+   - Full, diag, spherical, tied covariances  
+   - MLE or MAP estimation with weight, mean, or covariance priors  
+
+2. **GMMInitializer**  
+   - `kmeans`, `kpp` (k-means++), `random`, `points`, `maxdist`  
+
+3. **ClusteringMetrics**  
+   - Unsupervised metrics (Silhouette, Davies-Bouldin, etc.)  
+   - Supervised metrics (ARI, NMI, Purity, Confusion Matrix, etc.)  
+
+## Installation
+
+```bash
+git clone https://github.com/YourUser/TorchGMM.git
+cd TorchGMM
+pip install -r requirements.txt
+```
+
+Make sure you have PyTorch installed. For GPU usage, install the CUDA-enabled version of PyTorch as per the official instructions.
+
+## Documentation
+We use Sphinx to build documentation. The generated HTML pages live under docs/_build/html/. You can also read them online if you host them (e.g., on GitHub Pages).
+
+```bash
+cd docs
+make clean
+make html
+# Open _build/html/index.html in a browser
+# Linux
+xdg-open _build/html/index.html 
+```
+
+The docs include:
+
+API Reference for all modules (see GaussianMixture, GMMInitializer, and ClusteringMetrics).
+Tutorials that walk through different usage scenarios (basic GMM, metrics, using priors).
+Tutorials
+We provide three Jupyter notebooks in the tutorials/ folder:
+
+GMM Tutorial: Basic usage of the GaussianMixture class.
+Metrics Tutorial: Demonstrates ClusteringMetrics and how to compare models.
+Priors Tutorial: Shows how to use weight/mean/covariance priors (MAP).
+To view or run them locally, just open them in Jupyter or VS Code.
+
+## Basic Usage Example
+Here’s a short snippet:
+
+```python
+import torch
+from utils.gmm import GaussianMixture
+
+# Generate random 2D data
+X = torch.randn(500, 2)
+
+# Create and fit the GMM
+gmm = GaussianMixture(
+    n_features=2,
+    n_components=3,
+    covariance_type='full',
+    max_iter=200
+)
+gmm.fit(X)
+
+print("Converged?:", gmm.converged_)
+print("Cluster Weights:", gmm.weights_)
+print("Cluster Means:", gmm.means_)
+```
+
+You can also run on GPU by specifying device='cuda' in the GaussianMixture constructor (assuming a CUDA-capable device).
+
+## Contributing
+Fork the repository and create your feature branch.
+Make changes and add tests or notebooks as appropriate.
+Submit a pull request (PR) for review.
+We welcome improvements to both the code and the documentation.
+
+## License
+Released under the MIT License.
+© 2025, Adrián A. Sousa-Poza

tgmm-0.1.5/README.md

@@ -0,0 +1,94 @@
+# TorchGMM: A Gaussian Mixture Model Implementation with PyTorch
+
+![TorchGMM logo](docs/_static/logo-no-background.png)
+
+**TorchGMM** is a flexible implementation of Gaussian Mixture Models in PyTorch, supporting:
+
+- EM Algorithm
+- MAP Estimation with Priors
+- Multiple Covariance Types
+- Various Initialization Methods
+- Comprehensive Clustering & Evaluation Metrics
+
+## Features
+
+1. **GaussianMixture**  
+   - Full, diag, spherical, tied covariances  
+   - MLE or MAP estimation with weight, mean, or covariance priors  
+
+2. **GMMInitializer**  
+   - `kmeans`, `kpp` (k-means++), `random`, `points`, `maxdist`  
+
+3. **ClusteringMetrics**  
+   - Unsupervised metrics (Silhouette, Davies-Bouldin, etc.)  
+   - Supervised metrics (ARI, NMI, Purity, Confusion Matrix, etc.)  
+
+## Installation
+
+```bash
+git clone https://github.com/YourUser/TorchGMM.git
+cd TorchGMM
+pip install -r requirements.txt
+```
+
+Make sure you have PyTorch installed. For GPU usage, install the CUDA-enabled version of PyTorch as per the official instructions.
+
+## Documentation
+We use Sphinx to build documentation. The generated HTML pages live under docs/_build/html/. You can also read them online if you host them (e.g., on GitHub Pages).
+
+```bash
+cd docs
+make clean
+make html
+# Open _build/html/index.html in a browser
+# Linux
+xdg-open _build/html/index.html 
+```
+
+The docs include:
+
+API Reference for all modules (see GaussianMixture, GMMInitializer, and ClusteringMetrics).
+Tutorials that walk through different usage scenarios (basic GMM, metrics, using priors).
+Tutorials
+We provide three Jupyter notebooks in the tutorials/ folder:
+
+GMM Tutorial: Basic usage of the GaussianMixture class.
+Metrics Tutorial: Demonstrates ClusteringMetrics and how to compare models.
+Priors Tutorial: Shows how to use weight/mean/covariance priors (MAP).
+To view or run them locally, just open them in Jupyter or VS Code.
+
+## Basic Usage Example
+Here’s a short snippet:
+
+```python
+import torch
+from utils.gmm import GaussianMixture
+
+# Generate random 2D data
+X = torch.randn(500, 2)
+
+# Create and fit the GMM
+gmm = GaussianMixture(
+    n_features=2,
+    n_components=3,
+    covariance_type='full',
+    max_iter=200
+)
+gmm.fit(X)
+
+print("Converged?:", gmm.converged_)
+print("Cluster Weights:", gmm.weights_)
+print("Cluster Means:", gmm.means_)
+```
+
+You can also run on GPU by specifying device='cuda' in the GaussianMixture constructor (assuming a CUDA-capable device).
+
+## Contributing
+Fork the repository and create your feature branch.
+Make changes and add tests or notebooks as appropriate.
+Submit a pull request (PR) for review.
+We welcome improvements to both the code and the documentation.
+
+## License
+Released under the MIT License.
+© 2025, Adrián A. Sousa-Poza

tgmm-0.1.5/setup.cfg

@@ -0,0 +1,4 @@
+[egg_info]
+tag_build = 
+tag_date = 0
+

tgmm-0.1.5/setup.py

@@ -0,0 +1,30 @@
+from setuptools import setup, find_packages
+
+setup(
+    name="tgmm",
+    version="0.1.5",
+    author="Your Name",
+    author_email="adrian.sousapoza@gmail.com",
+    description="A Gaussian Mixture Model (GMM) based on Expectation-Maximisation (EM) implemented in PyTorch",
+    long_description=open('README.md').read(),
+    long_description_content_type="text/markdown",
+    url="https://github.com/adriansousapoza/TorchGMM",
+    packages=find_packages(),
+    install_requires=[
+        "torch>=2.5.1",
+        "numpy>=1.23.0",
+        "matplotlib>=3.10.1"
+    ],
+    extras_require={
+        "docs": [
+            "Sphinx>=8.2.3",
+            "Pygments>=2.19.1",
+        ],
+    },
+    classifiers=[
+        "Programming Language :: Python :: 3",
+        "License :: OSI Approved :: MIT License",
+        "Operating System :: OS Independent",
+    ],
+    python_requires='>=3.6',
+)

tgmm-0.1.5/tests/test_gmm.py

@@ -0,0 +1,21 @@
+import torch
+import pytest
+from tgmm import GaussianMixture
+
+def test_gaussian_mixture_fit_predict():
+    # Generate a synthetic dataset
+    X = torch.randn(100, 2)
+    
+    # Instantiate a GMM with 3 components
+    gmm = GaussianMixture(n_components=3, n_features=2)
+    gmm.fit(X)
+    
+    # Predict cluster labels
+    labels = gmm.predict(X)
+    
+    # Check that labels have the correct shape and values in the expected range
+    assert labels.shape == (100,)
+    assert labels.max() < 3 and labels.min() >= 0
+
+if __name__ == '__main__':
+    pytest.main([__file__])

tgmm-0.1.5/tgmm/__init__.py

@@ -0,0 +1,4 @@
+from .gmm import GaussianMixture
+from .gmm_init import GMMInitializer
+from .metrics import ClusteringMetrics
+from .plotting import plot_gmm, dynamic_figsize