Installation

Requirements

  • Python 3.10 or newer

  • Git (required to install the DCA dependency from source)

Standard install

Clone the repository and install in editable mode so local changes are reflected immediately:

git clone https://github.com/dulithaprasanna/ALSEBO.git
cd ALSEBO
pip install -e .

This installs all dependencies declared in setup.cfg, including the dca package which is fetched directly from GitHub:

dca @ git+https://github.com/utdal/py-mfdca.git@vae_dependency

Note

Because dca is a Git-only dependency, make sure Git is available in the environment where you run pip install.

Dependencies

The following packages are installed automatically:

Package

Purpose

numpy / pandas

Array operations and data I/O

scikit-learn

Gaussian Process Regression, t-SNE, PCA, k-means

biopython

FASTA parsing in sequence space and VAE data loading

transformers / torch

ESM protein language model embeddings

tensorflow

VAE training (encoder–decoder architecture)

matplotlib / seaborn

Plotting utilities

dca

Mean-field Direct Coupling Analysis features

Verify the installation

import alsebo
print(alsebo.__version__)

Building the documentation locally

pip install tox
tox -e viewdocs   # builds HTML and serves at http://localhost:8000