How to install Conda Forge packages

Owned by Aaron Jackson, created
Last updated: Oct 28, 2024 by Laura Ward
1 min read

Introduction

TRE Users who program in Python have acces to additional packages within Conda Forge, a curated repository for installing packages via Conda. Conda is an open source package and environment management system that is used to managing packages in various programming languages. While people who program in R may use this, it is likely more applicable to Python TRE Users.

As of January 2023, we introduced new functionality which allows TRE users to install packages available within conda-forge, without needing to make a request to HIC Support to add packages for you. This works in the same way as our CRAN proxy, which performs live scanning of packages as they are downloaded. If your workspace was created prior to January 2023, some manual changes will be required to your workspace.

As of August 2024, we also allow package installation from the Bioconda repository for bioinformatics software, enabling installation of packages like plink.

 

If your workspace was created <Feb 2023 you will need to complete the following steps

Configuring an old workspace

You only need to complete this step if your workspace was created before February 2023. If you are unsure about when your workspace was created, repeating these steps will not break anything - so you are safe to repeat this portion if needed.

If you are on an older Windows workspace, you may need to open an Anaconda PowerShell prompt. Under Ubuntu workspaces, any terminal window should work fine. This instructs your conda environment to connect to the HIC conda proxy, rather than trying to reach out to the internet directly.

$ conda config --set channels http://conda.hic-tre.dundee.ac.uk/conda-forge/

on Windows, you should also remove the global condarc file. From PowerShell, this can be done with:

$ Remove-Item -Path C:\conda\.condarc

And that’s it! You should now be ready to install packages from conda-forge within your workspace.

If it doesn’t work, it might be due to an unusual installation of Conda on your Windows workspace. You can override the default repo path in the install command. For example, to install numpy:

conda install -c http://conda.hic-tre.dundee.ac.uk/conda-forge/ --override-channels numpy

Enabling bioconda

As of August 2024, we also allow package installation from the bioconda repository. You will need to manually enable this channel.

 Instructions

Installation of packages should be as you are used to with conda. For example, to install numpy, you would simply run:

For queries or comments regarding HIC How To Articles contact, HICSupport@dundee.ac.uk

 

 Related articles