Release Notes#

0.3.0 (February 17, 2025)#

Added --workers option to the CLI for running multiple morphs in parallel.
Renamed --ramp-in and --ramp-out as --ease-in and --ease-out.
Added --ease option to the CLI, which is shorthand for --ease-in --ease-out.

Added Tutorials page to the documentation.
Added Shape Creation tutorial.
Added section to the documentation with ideas of how to use Data Morph in the classroom.
Updated contribution instructions in Custom Datasets tutorial.
Included reference to article on the creation of Data Morph in the documentation header.
Included reference to Data Morph: A Cautionary Tale of Summary Statistics (conference talk) in the documentation header.
Added contributing guidelines and code of conduct.
Configured canonical URL.

Reworked GIF creation logic to write only one PNG for distinct frames in the animation and use frequency counts to determine frame duration for speed improvements and less I/O.
Added three new datasets: soccer, pi, and gorilla. See the DataLoader for visuals.
Added four new shapes: Club, FigureEight, Spade, and Spiral.
Refactored some shape logic to use more numpy for speed improvements.
Added center property to Interval and BoundingBox to simplify shape calculations.
Made it possible to unpack the bounds from the BoundingBox.
Switched from tqdm to progress bars powered by rich.
Renamed the ramp_in and ramp_out arguments to DataMorpher.morph() as ease_in and ease_out.
Added Shape.get_name() method and Shape.name attribute to replace __str__() shape naming logic, which removed some redundancies in ShapeFactory.
Refactored the shape code into separate modules per shape.
Renamed get_values() as get_summary_statistics().
Improved variable names in several spots for more readable code.

Added show_bounds parameter to Dataset.plot() to visualize the automatically-calculated bounds.
Added title parameter to Dataset.plot() with default showing the name and number of points.
Switched from using Matplotlib’s tight_layout() to constrained layout.
Added style_context() context manager for plotting in a Data Morph style.
Added test suite for plotting.style module.
Changed plot layouts for DataLoader.plot() and ShapeFactory.plot_available_shapes().
Adjusted grid of datasets in the documentation to use tighter plot bounds.

Added new GitHub Actions workflow to run Data Morph on new/altered datasets, shapes, or core logic.
Switched to trusted publishing on PyPI via GitHub Actions.
Added PR and issue templates.
Added pytest-xdist to dev dependencies and reworked the test suite to make it faster.
Switched from black to ruff formatting in pre-commit setup and included additional rulesets.
Added pyproject-fmt pre-commit hook.
Refactored shape tests.

Addressed compatibility issues with numpy>=2.0.0.
Assorted performance and code quality improvements after running additional ruff rules.
Fixed some incorrect return types.

Replaced tqdm dependency with rich.
Replaced pytweening dependency with functions in plotting.animation.
Replaced sphinxarg.ext Sphinx extension with sphinx_argparse_cli.
Changed the minimum Sphinx version to 7.2.6.
Changed the minimum pydata-sphinx-theme version to 0.15.3.
Updated pre-commit hooks to the latest versions.
Factored out scipy dependency.
Updated GitHub Actions workflows for Node 16 deprecation.
Enabled Dependabot on GitHub Actions and switched to use commit hashes for versioning.
Updated Python testing matrix to include Python 3.9 through 3.13.

Created 5 new target shapes: Diamond, Heart, LeftParabola, RightParabola, Rings.
Created 3 new datasets: bunny, Python logo (TM), SuperDataScience logo. Logos are used with permission.
Made it possible to install via conda.
Configured versioned documentation hosted in GitHub Pages.
Generated a CLI reference page in the documentation.
Provided a custom dataset creation tutorial to the documentation.
Included logo and badges in README.
Included an example using easing for the animation to the documentation.
Included reference to article on the creation of Data Morph in the documentation.
Configured codecov reporting on test coverage with actions to validate the config both periodically and upon change.
Incorporated metavar indicators for some CLI options.
Added a test of the frame freezing functionality in the DataMorpher.
Added a test for the ShapeFactory.plot_available_shapes() method used in the documentation.
Added a test for the PointCollection.plot() method used in the documentation.

Fixed links to Autodesk assets after change to their website.
Corrected missing type references in the documentation.
Addressed a bug in the DataMorpher that caused frame recording to miss first frame (initial dataset).
Fixed a bug in the DataMorpher that wasn’t properly decreasing the temperature in the simulated annealing process.
Reduced point size for datasets displayed in the documentation so points appear with some separation and stay true to their subjects.
Fixed bug in logic for padding statistics in plots that would use incorrect spacing in some cases.

Changed the minimum Sphinx version to 7.2.1.
Changed the minimum pytweening version to 1.0.5.
Switched from isort and flake8 to ruff in pre-commit setup.
Updated pre-commit hooks to the latest versions and to use the new upstream numpydoc validation hook.

This is the first release of Data Morph, which is a spinoff of jmatejka/same-stats-different-graphs. Data Morph extends the original premise from the Autodesk Research paper to be more general. The core improvements include:

Created a modular package, moving away from functional programming to object-oriented programming.
Introduction of Shape classes both for resuability of the code and to decouple the shapes from the original “datasaurus” dataset and its hardcoded values. Morphing is now possible from any input dataset to the target shapes. See the ShapeFactory documentation for visuals.
Creation of Dataset to hold the data along with bounds needed for morphing and plotting and the name.
Morph and plot bounding logic for automatic calculation of values needed for simulated annealing and plot limits.
Fun new starter datasets: cat, dog, music, panda, and sheep. See the DataLoader documentation for visuals.
Easier application of plotting theme.
Documentation with Sphinx.
Testing suite with pytest.
Replaced center shape with Scatter and included a new Rectangle shape.