How to install a Python package into Sandbox

Follow the instructions here to load your package into FinnGen green bucket. After your package is in the green bucket follow these steps to install the Python package to your home/ivm.

If you have third-party packages and many dependencies installation from the source may get tricky. You may consider using the Anaconda Python environment within a Docker image instead.

Step 1:

In sandbox open Terminal Emulator. Check that you are in /home/ivm

pwd
/home/ivm/

Step 2:

Copy the source code from /finngen/green to /home/ivm/. Here we will load source code for yhaplo software.

cp /finngen/green/yhaplo-master.zip .

Step 3:

Extract the source code

unzip yhaplo-master.zip

Step 4:

Change directory

cd yhaplo-master 

Step 5:

Check the status of installation in sandbox environment

pip show yhaplo

Step 6:

Information about the package appear indicating that the package is installed on python2.

To uninstal Python package

pip uninstall yhaplo

Step 7:

Check the status on python3

pip3 show yhaplo

Step 8:

No output appears indicating that the package is not installed on python3.

To install the Python package with python3 move to the folder where your package is (here /home/ivm/yhaplo-master).

cd /home/ivm/yhaplo-master

Step 9:

Install with python3. Note that the code must be run in the package's folder where setup.py exists (here /home/ivm/yhaplo-master).

pip3 install --editable .

Step 10:

Check where the package is installed

pip3 show yhaplo

Step 11:

You have to add the local installed package directory into PYTHONPATH.

There are two ways to do this:

First way on command line (This command needs to be run every time you open Sandbox instance)

export PYTHONPATH = $PYTHONPATH:/home/ivm/.local/lib/python3.7/site-packages

Or paste the same command into ~/.bashrc

export PYTHONPATH = "${PYTHONPATH}:/home/ivm/.local/lib/python3.7/site-packages"

and then source the file by typing the below command

source ~/.bashrc

This way, every Sandbox instance will automatically add the PYTHONPATH

Second way is to add them in Python itself. Doing this one time enough and it will remember the path next time you run Python or even open new Sandbox instance. Follow the instruction below

python3

import sys

sys.path.insert(0,'/home/ivm/.local/lib/python3.7/site-packages')

import yhalpo

Either of them will add the path and will load the installed packages.

Note: If the package is not installed it might be incompatible with other Python packages in the Sandbox. In that case, you can make Python Virtual Environment and load your package there.