Installation

Requirements

  • Python >= 3.6

  • pandas

  • requests

  • tqdm

  • typing_extensions

  • pyyaml

  • scikit-learn

Installing improvelib with pip

pip install improvelib

Installing improvelib with git

Clone the IMPROVE repository to a directory of your preference (outside of your drug response prediction (DRP) model’s directory).

git clone https://github.com/JDACS4C-IMPROVE/IMPROVE
cd IMPROVE
git checkout develop

Note

You can git checkout older versions of IMPROVE if needed. See here for details on older versions.

To use improvelib, specify the full path to the IMPROVE library with $PYTHONPATH.

export PYTHONPATH=$PYTHONPATH:/your/path/to/IMPROVE

Downloading IMPROVE benchmark datasets

We currently have benchmark data for Drug Response Prediction. This data can be downloaded from this link or to a specified definition with the following:

./scripts/get-benchmarks $DESTINATION/csa_data/raw_data

Installing conda

To check if you have conda, use conda --version.

If you do not have conda installed, install with the following commands:

wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
chmod +x Miniconda3-latest-Linux-x86_64.sh
./Miniconda3-latest-Linux-x86_64.sh
source ~/.bashrc

Now, using the conda --version command should give the version you’ve downloaded. If not, close the terminal session and open again, then try conda --version.