MacLinuxInstallation.md

Local Installation on Mac and Linux

If you have any problems with this installation, please file an issue and describe any problems, so we can improve the instructions.

Install the lastest version of Python using Anaconda

Download the Python 3.7 Anaconda installer and run the Anaconda installer.

Open a terminal window

Clone this repository

git clone /~https://github.com/sbl-sdsc/mmtf-proteomics.git

Create a conda environment

cd mmtf-proteomics

conda env create -f binder/environment.yml

Activate the conda environment

conda activate mmtf-proteomics

Launch Jupyter Notebook

jupyter notebook

After you are finished, deactivate the conda environment

conda deactivate

Anytime you want to use the environment, activate it again and start Jupyter Notebook

To permanently remove the benchmark environment

conda remove -n mmtf-proteomics --all

Setting Spark Configurations

When running PySpark on many cores (e.g., > 8), the memory for the Spark Driver and Workers may need to be increased. If necessary, set the environmental variable SPARK_CONF_DIR to the conf directory provided in this repository in your .bashrc (Linux) or .bash_profile (Mac) file.

export SPARK_CONF_DIR=<path>/mmtf-proteomics/conf

Then close the terminal window and reopen it to set the environment variable.

The conf directory contains the file spark-env.sh with the following default settings.

SPARK_DRIVER_MEMORY=4G
SPARK_WORKER_MEMORY=4G

When running this repo on 24 core machine, you may need to increase the memory settings to 20G.