OfflineDataSci

This package helps you download and configure common tools for teaching and doing data science without an internet connection. This includes:

• Installers for data science languages: Currently R and Python
• Installers for common data science IDEs: Currently RStudio
• Partial local mirrors of package repositories: Currently CRAN (for R) and PyPI (for Python)
• Locally browseable clones of data science teaching websites: Currently Data Carpentry and Software Carpentry

Status

Early stage experiment

Installation

Using pip:

From PyPI

pip install offlinedatasci

From GitHub (latest development version)

pip install git+https://git@github.com/carpentriesoffline/offlinedatasci.git

For local development:

Clone the repository and from the root directory run:

pip install .

On macOS make sure wheel package is installed first.

Usage

Download and setup everything

offlinedatasci install all /install/path

Create just the local CRAN mirror with basic data science packages

offlinedatasci install minicran /install/path

Add packages to repository mirrors:

To add packages beyond those included in the basic data science teaching focused mirrors use add-packages The command structure is offlinedatasci add-packages followed by the language you want to add packages to, followed by the names of the packages, followed by the path of the mirror. The path should be the same as where the mirror was originally setup, so the same install path you used for setup offlinedatasci.

For example, to add the sf, terra, and stars geospatial packages to the CRAN mirror:

offlinedatasci add-packages r sf terra stars /install/path

Developer docs

Creating a release

1. Increment the version numbers in pyproject.toml, __init__.py, and docs/conf.py
2. Commit and push the changes to GitHub
3. Create a tag for the new version
4. Push the tag to GitHub (git push upstream <tag_name>)
5. Make sure the build package is installed
6. Make sure your PyPI credentials are stored in ~/.pypirc
7. Build the source distribution: python -m build --sdist
8. Build the universal wheel: python -m build --wheel
9. Upload the new release to PyPI: twine upload dist/*