First..

1. install poetry, pyenv
2. create project directory

Folder structure

1. src(with current project, subprojects)
2. scripts
3. data, experiments(tensorboard), configs(hydra), ...

Dealing with python version with pyenv

1. pyenv install {preferred_version}
2. pyenv global {preferred_version}

Dependency and virtualenv control with poetry

1. poetry init
2. poetry env use python # activate virtualenv using current python version
3. poetry shell # activate virtualenv using poetry
4. poetry install # install dependencies + current project (as editable package, so that modules in src/ can be cross-imported or imported in scripts/)

submodules

1. Basically, there are two ways to add submodules.
   1. git submodule add {url} {external/submodule1} and make symlink using ln -s {/path/to/external/submodule1} {/path/to/src/submodule1}
   2. git submodule add {url} {src/submodule2} if python files(and init.py) are exposed in the main directory

debug

You can easily debug your code (using script or module directly) in vscode since we installed this package.

github setting

1. create repo in github
2. follow the instruction in README.md