Standard Operating Procedure for Creating a new Python Project

1. Create a new Github repository on own account or on organization account
2. Create a new pycharm project by using the pycharm github integration and cloning the repo from github into pycharm
3. Add any pre-exisiting files to the pycharm project that is now automatically set up with a git remote
4. Once inside pycharm, create a local virtual environment using pycharm's own venv creation interface

Legend

1. Project specific .py files go into src or sub-dir of src. All plotting scripts go in src/plotting. Avoid any matplotlib or related imports in main src files
2. All scripts (think .sh files, and template .py files) that write to results should be in scripts/
3. All .ipynb files go into notebooks. Using google colab is preffered but files should be downloaded into this folder once done
4. All figures/videos/plots go in the corresponding sub-dir of results. Example all .png files in results/figures
5. All models (bullet/mujoco physcis based, trained, or otherwise) go into models or sub-dir of models
6. Any documentation that does not belong in the README.md (text files for minor references etc.) and things like referred PDFs go into notes
7. All local datasets and problem instance files go into data or sub-dir of data

Append project .gitignore with

notes/
data/

Command to be able to add an empty directory tree to git

Command used to push this empty template repo to git/guthub was. Run this before pushing whenever a new dir is added to this template

find . -type d -empty -not -path "./.git/*" -exec touch {}/.gitkeep \;

Once cloned, run the following to get rid of these .gitkeep files and keep your project pristine

find . -name '*.gitkeep' -delete

Why this exists

When a python-project reaches a stage where it is ready to transition from a collection of scripts to a "code-base" then at that point (this should happen once the viability of the project has been established) a remote (typically a GitHub) should be created for it using this template to save time and maintain standardization in relative paths.