Hoffman2:Python

From Center for Cognitive Neuroscience
Jump to navigation Jump to search

Use Python in Hoffman2

Use Python modules

module av python

This will show all the available Python modules in Hoffman2

python/2.7.15  python/2.7.18  python/3.6.8(default)  python/3.7.3  python/3.9.6

Be aware that the default version is python/3.6.8. To use a specific version, use command for example

module load python/3.7.3


Install Python libraires

Here shows the way to install Python libraries into your user space, which is under ~/.local

pip3 install library_name --user

For example

pip3 install scipy --user

This will install Scipy library into your ~/.local. Add the following to your ~/.bashrc

export PATH=$HOME/.local/bin:$PATH

Use Conda for Python virtual environment

Conda is very handy for managing Python environment. One can create multiple Conda environment for different purpose.

For example, if one application you want to use requires Python 3.6 but another application requires Python 3.7, you can have two separate Conda environments with one providing Python 3.6 and the other with Python 3.7.


To use conda, first load the module

module load conda

Then run the following command to setup the path

source $CONDA_DIR/etc/profile.d/conda.sh

Next, generate a conda environment with your preferred Python version

conda create --prefix /u/project/YOUR_GROUP/YOUR_CONDA_ENV_NAME python=3.7.3

This will generate a conda environment with python 3.7.3 (you can choose other existing version)

Activate your Conda environment

conda activate /u/project/YOUR_GROUP/YOUR_CONDA_ENV_NAME

Then you should see the ssh prompt changes to (/u/project/YOUR_GROUP/YOUR_CONDA_ENV_NAME) in front of your cursor. This means you are inside of the Conda virtual environment

Try check the Python version and see if it matches what you have picked

python --version

Then you can install other Python libraries you need here by "pip install". For example,

pip install numpy

This command will install Numpy library into this very Conda environment.

To exit the Conda environment, just type in

conda deactivate

CCN existing Conda Environments

ccn_py37 (equipped with nibabel, nilearn, pydicom, pandas, scikit-learn, scipy and Python 3.7)

/u/project/CCN/apps/conda/rh7/ccn_py37

tf230 (for Tensorflow 2.3.0)

/u/project/CCN/apps/conda/rh7/bcgnet/tf230

Python 2 Statement

Python 2 has officially reached its End of Life (EOL). That means that it will no longer be receiving active support or updates from their core development team as of January 1st, 2020.

As a result, several open-source packages have pledged to drop support for Python 2.7. For more information about these participating projects, please refer to the Python 3 Statement https://python3statement.org/