Working locally

On Mac or Linux machine

1. Install a conda based package manager. We recommend Miniforge3 if you are setting up for the first time.
   • miniforge3
   • micromamba
2. Install VS Code
3. Install 'code' command in PATH

Clone this repo and navigate into the dir:

git clone https://github.com/Adamtaranto/python-novice-dataframes.git
cd python-novice-dataframes

Set up a python environment with the packages we will be using:

Micromamba

Setup new micromamba environment.

# Create env and install packages from yml
micromamba env create --name pandas-workshop --file environment.yml

# Activate the environment
micromamba activate pandas-workshop

Or using Conda

# Create env and install packages from yml
conda env create --name pandas-workshop --file environment.yml

# Activate the environment
conda activate pandas-workshop

Or using pip

pip install matplotlib numpy pandas scipy seaborn jupyterlab

Now you can launch Jupyter Lab.

jupyter lab

Setup for Windows

1. Install GitBash
2. Install VS Code If you are using VS Code on a windows machine you will also need to set your default shell as "GitBash".
3. Install miniforge3
4. Open a GitBash shell and run this command to enable Conda: ~/miniforge3/Scripts/conda.exe init bash
5. From a new GitBash shell use conda to install packages: conda install -c conda-forge jupyterlab pandas numpy matplotlib seaborn scipy

Clone this repo and navigate into the dir:

git clone https://github.com/Adamtaranto/python-novice-dataframes.git
cd python-novice-dataframes

Launch Jupyter Lab.

jupyter lab

Working on WEHI Ondemand

You can launch a in interactive Jupyter Lab session on the WEHI HPC (Milton) via Ondemand.

You must have a Milton HPC account and a VAST scratch workspace set up before using this option.

• Sign in to WEHI Ondemand.
• From Apps select Jupyter
• Under the option "Extra Jupyter arguments" enter: --notebook-dir=/vast/scratch/users/$USER
• Open the Jupyter session. From the menu bar select View >> Open JupyterLab

To install packages in a Jupyter session running on Milton you will need to install into a target location that is visible from the notebook.

# Define user package location
from os import environ
username = environ["USER"]
pkgdir = f"/vast/scratch/users/{username}/workshop-pkgs"

# Prepend PYTHONPATH
import sys
sys.path.insert(0, pkgdir)

# Install package
!pip install --target $pkgdir pandas

# Package should now be visible
import pandas

Working in Gitpod

Click the Gitpod button at the top of this README to launch a gitpod workspace with all the required software pre-installed.

If you have a paid Gitpod account you can increase the timout limit on your workspace. Otherwise you will need to restart the workspace periodically.

# Increase gitpod timeout setting
gp timeout set 6h

Manually start Jupyter session

# Launch jupyter-lab
jupyter lab --NotebookApp.allow_origin='*' --NotebookApp.allow_remote_access=True --NotebookApp.token='' --NotebookApp.password='' --no-browser --port=8888

Note: See other gitpod settings here.