sharepy

Want to share your overly complicated conda environment? Or want to delete all those old conda environments but appease your neuroticism and save a backup? Use this shell script to easily save conda environments!

Requires conda >= 4.8.0. If script does not work, try updating conda.

Tested in x86_64-pc-linux-gnu (64-bit) and x86_64-w64-mingw32/x64 (64-bit) distributions.

Install and run by adding script to path

1. Download the sharepy bash script.

sudo wget https://gist.githubusercontent.com/Yoyomanzoor/0ae9008abd4e72384e21098580a9d75b/raw/7d9f9129398bf2f7e8f6a1aabde3177177049316/sharepy.sh -O /usr/local/bin/sharepy
sudo chmod +x /usr/local/bin/sharepy

2. Run sharepy -e <environment name> -o <output directory>

Install and run without adding script to path

This method requires no administrator privileges (although it is still possible to add to path without admin privileges)

1. Download the sharepy bash script via wget https://gist.githubusercontent.com/Yoyomanzoor/0ae9008abd4e72384e21098580a9d75b/raw/c7bea25d5b8f57fa94a18e7218604e3b48fc2c8f/sharepy.sh.
2. Run path/to/sharepy/sharepy.sh -e <environment name> -o <output directory>

Install and run by adding to ~/.bashrc

In your ~/.bashrc file, you can add this function:

# sharepy by Yoyomanzoor https://github.com/Yoyomanzoor/sharepy
function sharepy() {
    conda activate $1
    mkdir -p $1
    cd $1
    conda list > $1.txt
    conda list --explicit > $1_explicit.txt 
    conda env export > $1.yml
    conda env export --from-history > $1_cp.yml
    cd ..
}

To use this version, run sharepy <environment name> in the directory you would like the output.

That's it!

Output files

Say we have a conda environment called myenv.
There are 4 output files:

File	contents	re-install with conda
myenv/myenv.txt	human-readable table with environment info
myenv/myenv_explicit.txt	OS-specific list of tar files for all items in the environment	$ conda create --name <env> --file <this file>
myenv/myenv.yml	OS-specific yaml file of environment	$ conda env create -f <this file>
myenv/myenv_crossplatform.yml	Cross platform yaml file of environment. Be careful! There are known issues sometimes with this, such as not including pip dependencies.	$ conda env create -f <this file>