0.3.1

Internal

• Improve type annotations and use stricter pyright settings (#291)

Documentation

• Use sphinxcontrib.apidoc to automatically generate API docs from the source code (#296)
• Update hero image in the docs' landing page (#300)
• Update release process notes (#301 and #303)

CI/CD

• Fix regressions tests by comparing against JSON artifacts instead of flaky JPEGs (#299)

Thanks to @imprvhub for their contributions to this release! 🚀

---

0.3.0

Features

• Add support for histogram and bar traces (#287)

Documentation

• Small improvements to ridgeplot()'s docstring (#284)
• Misc improvements to the API docs and the getting-started and contributing guides (#287)

Internal

• Small improvements to type hints and annotations (#284)
• Introduce an internal ridgeplot._obj package to hold object-oriented interfaces (#287)

CI/CD

• Improve type annotations and switch from mypy to pyright with stricter settings (#287)
• Switch from black to the new ruff formatter (#287)

---

0.2.1

Bug fixes

• Fixed ZeroDivisionError for index-based colormodes when specifying single-trace or single-row plots (#268)

---

0.2.0

After almost 4 years, 30 "patch" releases, +200 pull-requests, and close to 1,000 commits, this is ridgeplot's first minor release (v0.1.30 -> v0.2.0)! 🚀

ridgeplot has been downloaded over 400k times (peaking at 102k downloads in a single month), is listed as a dependency in 135 public GitHub repositories, and - perhaps most relevantly - is a dependency of larger projects such as Shiny for Python, Ploomber, and NiMARE which further extends the impact and reach of the project.

This release marks a small milestone for ridgeplot, which we believe has now reached a level of maturity and stability that warrants a stricter and more structured, predictable, and standard release and versioning process. Even though we have managed to never publish breaking changes in the past (if you find any, please let us know!), we will from now on be even more careful and mindful of the impact of any changes that could affect downstream users and their applications.

We will make an effort to standardise and document our versioning policy. For now, we will try to simply adhere to the following general rules:

• We are explicitly not going to follow Semantic Versioning, as we believe it is not a good fit for this project yet.
• MAJOR.MINOR.PATCH versioning scheme:
  • MAJOR: We don't have any plans for this yet... we will probably use this in the future once we settle on a more stable API and feature set
  • MINOR: New features, significant improvements, and deprecations
  • PATCH: Backwards-compatible bug fixes, small improvements, internal changes, and documentation updates
• Breaking changes:
  • We might introduce breaking changes in minor releases!
  • However, this will never happen without a proper deprecation period and a clear upgrade path. i.e., we will always first deprecate the old API via a DeprecationWarning and provide a clear migration path to the new API.
  • Such instances will be kept to a minimum and will likely only show up in the form of deprecated or renamed parameters or the meaning/behaviour of their arguments/values.

Breaking changes

• Remove support for the deprecated show_annotations parameter and colormode='index' value (#254)
• The new default colormode is "fillgradient" (#244)
• The default value for line_width changed from 1 to 1.5 (#253)

Features

• Implement new "fillgradient" colormode (#244)
• Add new line_color parameter to the ridgeplot function (#253)
• Add a line_color='fill-color' option which automatically matches the trace's line color to the trace's fill color (#253)
• Add new norm parameter to the ridgeplot function to allow users to normalize the data before plotting (#255)
• Add sample_weights argument to ridgeplot() to allow users to pass sample weights to the KDE estimator (#259)

Deprecations

• Rename coloralpha to opacity for consistently with Plotly Express and deprecate the old parameter name (#245)
• Rename linewidth to line_width for consistency with Plotly's API and deprecate the old parameter name (#253)
• Deprecated colorscale='default' and list_all_colorscale_names() in favour or Plotly Express' px.colors.named_colorscales() (#262)

Dependencies

• The new minimum version of Plotly is 5.20 to leverage the new fillgradient feature (#244)

Optimizations

• Importing statsmodels, scipy, and numpy can be slow, so we now only import the ridgeplot._kde module when the user needs this functionality (#242)

Documentation

• Update examples in the getting-started guide to reflect the new default colormode (#244)
• Update the plotly.min.js version from 2.27 to 2.35.2 to leverage the fillgradient feature (#244)
• Fix the API reference docs for the internal ridgeplot._color module (#244)
• Tighten margins in generated examples (#257)
• Add the reference jupyter notebook used to generate the ridgeplot logo (#242)
• Update ridgeplot's logo to use Plotly's official colors (#243)

CI/CD

• Stop sending coverage reports to Codacy (#265)
• Improve local development experience and optimise the CI pipeline (#273)

Internal

• Simplify and refactor interpolate_color to not depend on px.colors.find_intermediate_color (#253)
• Improve type narrowing using typing.TypeIs (#259)
• Refactor community health files (#260)

Thanks to @sstephanyy for their contributions to this release! 🚀

---

0.1.30

Features

• Add support for named CSS colors (#229)
• Allow users to define color scales as a collection of colors (Collection[Color]) (#231)
• Dynamically infer the default colorscale from the active Plotly template (#237)

Documentation

• Improve the documentation for the colormode parameter (#232)

Internal

• Refactor _figure_factory.py to use a functional approach (#228)
• Stop using the term "midpoints" to refer to the "interpolation values" when dealing with continuous color scales (#232)
• Refactor color validation logic to use helpers provided by Plotly (#233)
• Drop colors.json and use Plotly's ColorscaleValidator and named_colorscales directly (#234)
• Refactor color utilities into ridgeplot._color (#235)

---

0.1.29

Features

• Add new "trace-index-row-wise" colormode (#224)

Improvements

• Remove duplicated labels when plotting multiple traces on the same y-axis row (#223)

Documentation

• Update and improve the "Contributing" guide (#218 and #221)

Internal

• Eagerly validate input shapes in RidgeplotFigureFactory (#222)
• Vendor _zip_equal() from more-itertools (#222)
• Improve overall test coverage (#222)

Bug fixes

• Support edge case in get_collection_array_shape where the input array is empty or contains nested empty arrays (#222)

CI/CD

• Add new "Greet new users" workflow to welcome new contributors to the project (#210)
• Add concurrency entries to relevant GitHub workflows (#211)
• Add Dependabot configuration file (#211)
• Add GitHub issue templates (#211)
• Add support for Python 3.13 (#217)
• Add a CodeQL GitHub workflow (#220)

---

0.1.27

Breaking changes

• Dropped support for Python 3.8, in accordance with the official Python support policy¹. (#204)
• Removed deprecated function get_all_colorscale_names() in favor of list_all_colorscale_names() (#200)

CI/CD

• Adopt setuptools-scm for package versioning (#200)
• Add actionlint pre-commit hook (#201)
• Improve logic in .github/workflows/check-release-notes.yaml to post comments to the PR (#201)
• Handle footnotes in the automatically generated release notes (#209)

---

1. https://devguide.python.org/versions/ ↩

0.1.26

Breaking changes

• Dropped support for statsmodels==0.14.2 due to import-time issue. See #197 for more details. (#198)

CI/CD

• Refactor test coverage logic (#193)
• Replace pip and venv with uv (#189)
• Move all CI/CD utilities to the cicd_utils/ directory (#186)
• Publish to PyPi as a Trusted Publisher (#187)
• Add check-jsonschema pre-commit hooks and define timeout-minutes for all GitHub workflows (#187)

0.1.25

This release contains a number of improvements to the docs, API reference, CI/CD logic (incl. official support for Python 3.12), and other minor internal changes.

Documentation

• Misc documentation improvements (#180)
• Move changelog to ./docs/reference/changelog.md (#180)

Internals

• Migrate from setup.cfg from pyproject.toml (#176)
• Use importlib.resources to load data assets from within the package - to be PEP-302 compliant (#176)
• Enforce "strict" mypy mode (mostly improved type annotations for generic types) (#177)

CI/CD

• Add support for Python 3.12 (#182)

---

0.1.24

Breaking changes

• Dropped support for Python 3.7. (#154)

Features

• Add hoverinfo by default to the Plotly traces. (#174)

Documentation

• Use the {raw} html :file: _static/charts/<PLOT-ID>.html directive to load the interactive Plotly graphs in the generated Sphinx docs. The generated HTML artefacts only include a <div> wrapper block now and the plotly.min.js is now vendored and automatically loaded via the html_js_files Sphinx config. (#132)
• Small adjustments to the example plots in the documentation. (#132)
• Reformat markdown files, removing all line breaks. (#132)

Internals

• Define a ridgeplot._missing.MISSING sentinel object for internal use (this replaces the multiple module-level _MISSING = object() sentinels). (#154)
• Add an internal extras/ directory to place helper modules and packages used in different CI tasks. (#154 and #161)

CI/CD

• Replace isort, flake8, and pyupgrade with ruff. (#131)
• Add regression tests for the figure artifacts generated by the examples in _ridgeplot_examples. (#154)
• Remove the Python locked dependency files. (#163)

---