Template for python projects based on devcontainer

Dependencies

• Visual Studio Code (with the "Remote containers" extension installed)
• Docker
• (Optional) Devcontainer CLI (allows to directly launch devcontainer from commandline, instead of opening VSCode and then clicking on "Reopen in container")

How to use

1. Clone this repo (or use it as a template for a new one)
2. Open VSCode
3. Either click on "Reopen in container" when the extension detects the devcontainer, or search for it in the command palette (can be opened with CTRL + SHIFT + P)

How to setup access for DVC remote storage (S3)

• generate an AWS credentials file by copying the example one (cp .aws/credentials.example .aws/credentials)
• fill the credentials file with your personal information (AWS access key and secret key)
• initialize DVC in the repo with dvc init
• configure the DVC default remote storage on AWS S3 with dvc remote add s3-bucket s3://{BUCKET_NAME} && dvc remote default s3-bucket (note that you need to change {BUCKET_NAME} with the actual name of the S3 bucket you want to use)

Jupytext guide

Jupytext extension allow to translate common text files (e.g. python or markdown) in jupyter notebooks and viceversa. Jupytext is integrated via this Jupytext VSCode extension. Below a simple guide on how to use it:

• create a .py file
• insert some code
• when you want to run it as a notebook:
  • move to file explorer section (top left icon in vs code)
  • right-clik on the file you want to open
  • click "Open as a Jupyter Notebook"
• modifying the notebook will change also the original file
• modifying the original file will not change the notebook already opened (re-launch to see changes on notebook)

Additional notes:

• jupytext syntax:
  • when the file is opened as a notebook the first time will be formatted automatically
  • to divide cells just add this comment

  # %%
  ... code here ...
  

• it works with different extensions (.py, .md etc.)
• NB: By default the jupytext notebook file opened in this way does not exist on disk, thus modifying the file once the notebook is opened will not affect the file.
• metadata that can be visible on the original file after opening it as a notebook are not considered by this extension (this enhacement is in the roadmap)