How to Add scikit-plots to conda-forge Guidelines#

This guide provides a comprehensive, beginner-friendly walkthrough for adding scikit-plots to conda-forge, including best practices, important notes, and maintenance instructions.

Prerequisites#

Ensure you have:

  • A GitHub account.

  • scikit-plots published one of:

  • Installed tools: git, conda, and grayskull.

Important

Your package must be available via a valid source distribution like GitHub, PyPI, or Anaconda.org before it can be added to conda-forge. But can be use direct github repo, read below more info.

Hint

Install grayskull easily:

conda install -c conda-forge grayskull

or:

pip install grayskull

Choosing a Source Distribution#

When updating the recipe (meta.yaml), you must choose one of the valid source distribution types. Conda-forge supports four typical options:

  1. GitHub archive (tagged release)

    source:
  url: https://github.com/scikit-plots/scikit-plots/archive/v{{ version }}.tar.gz
  sha256: <fill-this-sha>

    Important

    • Ensure the tag exists on GitHub.

    • Use a specific release version to avoid build inconsistencies.

  2. PyPI source distribution

    source:
  url: https://pypi.org/packages/source/s/scikit-plots/scikit_plots-{{ version }}.tar.gz
  sha256: <fill-this-sha>

    Note

    This is the most common and preferred method when available on PyPI.

  3. Anaconda.org source (e.g., staging wheels)

    source:
  url: https://pypi.anaconda.org/<channel>/simple/scikit-plots/{{ version }}/scikit_plots-{{ version }}.tar.gz
  sha256: <fill-this-sha>

    Hint

    Replace <channel> with your actual Anaconda channel, such as scikit-plots-wheels-staging-nightly.

  4. Direct Git repository source

    source:
  git_url: https://github.com/scikit-plots/scikit-plots.git
  git_rev: {{ tag }}  # 🔒 Use a tag or commit hash for reproducibility

    Important

    • Always use a tag or commit hash, not main or master.

    • This method is less common and should only be used when source tarballs are not available.

Publishing to PyPI#

If you use PyPI:

Before proceeding, confirm that scikit-plots is correctly published on PyPI.

Important

Conda-forge builds the package from PyPI releases, not from GitHub directly (unless you configure otherwise in your recipe).

Hint

If you need help publishing, check out the official PyPI packaging tutorial.

Creating the Recipe with Grayskull#

Use grayskull to generate the initial conda recipe:

grayskull pypi --strict-conda-forge scikit-plots

This generates a recipes/scikit-plots/ folder containing a meta.yaml.

Note

grayskull will automatically pull metadata from your PyPI release — but you must review it.

{% set name = "scikit-plots" %}
{% set version = "0.4.0rc4" %}
{% set tag = "v" ~ version %}

package:
  name: {{ name|lower }}
  version: {{ version }}

source:
  # Choose one valid source:

  # GitHub
  git_url: https://github.com/scikit-plots/scikit-plots.git
  git_rev: {{ tag }}       # 🔒 use a tag or commit hash for reproducibility

  # GitHub archive
  # url: https://github.com/{{ name }}/{{ name }}/archive/{{ tag }}.tar.gz

  # PyPI source
  # url: https://pypi.org/packages/source/{{ name[0] }}/{{ name }}/{{ name | replace("-", "_") }}-{{ version }}.tar.gz

  # Anaconda source (less common)
  # url: https://pypi.anaconda.org/scikit-plots-wheels-staging-nightly/simple/{{ name }}/{{ version }}/{{ name | replace("-", "_") }}-{{ version }}.tar.gz

  url: https://pypi.org/packages/source/{{ name[0] }}/{{ name }}/scikit_plots-{{ version }}.tar.gz
  sha256: cd6c8a3d11cfe0b9cc3e4ecc95399efe16ea242ddb4c02505031c6271f8876f8

Note

You must include the correct sha256 checksum for the chosen source archive. To compute it:

## wget https://github.com/scikit-plots/scikit-plots/archive/refs/tags/{{ tag }}.tar.gz
wget https://github.com/scikit-plots/scikit-plots/archive/refs/tags/v0.4.0rc4.tar.gz

## openssl sha256 {{ tag }}.tar.gz
openssl sha256 v0.4.0rc4.tar.gz

Submitting to Staged-Recipes#

  1. Fork the conda-forge/staged-recipes repository.

  2. Clone your fork locally:

    git clone https://github.com/<your-github-username>/staged-recipes.git
cd staged-recipes/recipes/

  3. Create a new branch:

    git checkout -b add-scikit-plots

  4. Add your recipe directory and license file.

  5. Commit and push your changes:

    git add scikit-plots/
git commit -m "Add scikit-plots recipe"
git push origin add-scikit-plots

  6. Open a Pull Request (PR) to conda-forge/staged-recipes.

Important

Carefully fill out the PR checklist in the pull request description!

  • Confirm the recipe builds locally with conda build (optional but strongly recommended).

  • Confirm your metadata is accurate.

  • Ping reviewers if needed (@conda-forge/help-python).

Hint

Your PR title should follow the format: Add package: scikit-plots.

Post-Merge: Feedstock Creation#

After your PR is merged:

  • A new feedstock repository will be created, for example: scikit-plots-feedstock.

  • You will be added as a maintainer.

  • CI (Continuous Integration) will build and upload the package across all platforms.

Important

Carefully watch the CI builds! Build errors may still appear even after the staged-recipes PR is merged.

Hint

Add the feedstock repo to your GitHub notifications (watch → participating) to stay informed!

Maintaining the Package#

Future updates for scikit-plots:

  • Publish a new release to PyPI.

  • The conda-forge bot will open a PR automatically to update the feedstock recipe.

Manual Updates#

If the bot fails to update your package or if you need to make manual changes:

  1. Fork and clone your feedstock repository:

    git clone https://github.com/<your-github-username>/scikit-plots-feedstock.git
cd scikit-plots-feedstock/recipe/

  2. Update meta.yaml manually (or use grayskull again).

  3. Commit and push your changes:

    git add recipe/meta.yaml
git commit -m "Update scikit-plots to version X.Y.Z"
git push origin update-scikit-plots

  4. Open a pull request back to the feedstock repository.

  5. Wait for CI builds to pass, then merge.

Important

Always let CI finish before merging. Merging broken recipes can cause major issues across conda-forge!

Congratulations 🎉#

After the update PR is merged:

  • The new version of scikit-plots will be built and uploaded automatically.

  • Users will be able to install the latest version with:

conda install -c conda-forge scikit-plots

You are now officially maintaining a package on conda-forge!

Quick Reference Summary#

Follow these steps to successfully add and maintain scikit-plots on conda-forge.

Steps Overview#

  1. Choose source distribution

  2. Publish to PyPI (if needed)

  3. Generate recipe with grayskull

  4. Submit PR to staged-recipes

  5. Wait for review & merge

  6. Feedstock repository created

  7. Maintain future updates (bot/manual)

Hint

You can click on any action to jump directly to the detailed explanation!