Step 0

If you are on a Windows machine or plan to work on a windows machine, note: most/all instruction will be linux/ubuntu-based. You should download WSL (Windows Subsystem for Linux), which will provide you an ubuntu terminal on your windows machine.

Step 1 - Setup

Installing VSCode and Python Setup

• Install VScode.
• Install a package manager:
  • (Nice, but not necessary) Install conda
    • once installed, create a virtual environment, and then activate that virtual environment

conda create --name python_sandbox
conda activate python_sandbox

• Make sure that pip is installed. If not, install it
• After pip is installed, installed all of the necessary python packages used by this script

pip install matplotlib numpy

Installing Extensions

• Open up VScode.
  • If you are on a Windows machine, follow these instructions to make WSL your terminal in VSCoode
  • Select your conda environment as your 'python kernel' - follow these instructions
• Go to the Extensions Marketplace on the left-hand side
• Search for and install the following extensions:
  • Jupyter Keymap -- lets you develop jupyter notebooks locally, without any network connection.
  • Markdown All in One -- useful for making markdown files readable
  • Remote SSH -- this will be useful once we have GPU resources that we can log into

Step 2 - Inspecting the code

• Open up the folder containing this README and go to the .ipynb file. This is a jupyter notebook that references classes and functions in neighboring files.
• Read metadata.py to see how the Metadata class stores certain variables.
• Read the functions in plotting.py to see how plots are generated from input variables that are passed into each of the functions

Step 3 - Running the code

• You can run the code either by running each of the cells in the Assignment0_DataExploration.ipynb notebook or by running the Assignment0_DataExploration.py file. It would be good to run each of them to see how a python script behaves differently from a jupyter notebook.
  • Jupyter notebooks are typically good for data inspection and debugging, while python scripts are typically useful when running one process after another.