Installation

Requirements

• Python ≥ 3.10

• A working C compiler for Numba JIT (usually ships with your OS)

• Approximately 4 GB of RAM for typical datasets; larger genome-wide screens benefit from 16–64 GB

Install from PyPI

pip install crispyx

Install from source (development)

Clone the repository and install in editable mode:

git clone https://github.com/jaydu1/crispyx.git
cd crispyx
pip install -e .

Optional extras

crispyx offers optional dependency groups for different use cases:

# Testing dependencies (pytest, statsmodels, pydeseq2)
pip install -e ".[test]"

# Benchmarking dependencies (pertpy, pyyaml, psutil)
pip install -e ".[benchmark]"

# Documentation dependencies (sphinx, nbsphinx, etc.)
pip install -e ".[docs]"

# Install everything
pip install -e ".[test,benchmark,docs]"

Conda environment

A env.yml file is provided for setting up a full Conda environment including R packages for edgeR benchmarking:

conda env create -f env.yml
conda activate pert
pip install -e .

This installs R 4.1.1 with Bioconductor edgeR and rpy2 for cross-language benchmarking.

Verifying the installation

import crispyx as cx
print(cx.__version__)

# Quick check: open a dataset without loading into memory
adata = cx.read_h5ad_ondisk("data/demo_benchmark.h5ad")
print(adata)

Docker and Singularity

For HPC environments, pre-built container images are available. See Benchmarking and the benchmarking/singularity/ directory for Dockerfile and Singularity definition files.