v0.2.9

DataLab-Kernel v0.2.9 Release Notes

New Features

Plotly visualization backend

• Added a new Plotly interactive visualization backend as an alternative to the existing Matplotlib static backend. Plotly figures provide rich interactivity (zoom, pan, hover tooltips) directly in Jupyter notebooks.
• Install with pip install datalab-kernel[plotly].

Smart backend selection

• When both Plotly and Matplotlib are available, Plotly is automatically preferred for its richer interactive experience.
• Switch backends at runtime with plotter.set_backend("matplotlib") or plotter.set_backend("plotly").
• Override the default backend via the DATALAB_PLOTTER_BACKEND environment variable.

Enhanced Matplotlib backend

• Extracted the Matplotlib rendering logic into a dedicated matplotlib_backend module with improved support for:
  • Error bars
  • Curve styles (Sticks, Steps)
  • Per-object line styling from metadata
  • Log scale and axis bounds
  • Non-uniform image coordinates (pcolormesh)
  • Per-object colormaps

Figure height customization

• Enhanced figure height customization for both backends, with updated documentation.

MIME bundle support for Plotly

• Plotly figures now produce proper MIME bundles for native Jupyter display integration.

Improvements

Documentation

• Removed references to the legacy XML-RPC connection method; documentation now emphasizes the Web API as the primary (and recommended) connection method.
• Updated README with visualization backend documentation, usage examples, and runtime switching instructions.
• Updated online demo badge link.

Python version support

• Added Python 3.13 and 3.14 to the supported version classifiers.

Optional dependency

• Added plotly optional dependency group in pyproject.toml (pip install datalab-kernel[plotly]).

Dev tooling

• Enhanced wheel building script (serve_dev_wheels.py) to support package selection via --packages argument.
• New VS Code tasks for interactive plot tests, dev wheel serving, and backend-specific testing.

Bug Fixes

• Fixed test compatibility with Python 3.9 by using ExitStack instead of multiple mock.patch context managers.
• Fixed test module shadowing issue by using mock.patch.object instead of mock.patch with string targets.
• Fixed various Ruff and Pylint warnings across the codebase.
• Fixed test workflow branch references for DataLab installation.

Test Suite

• Added 65 tests for the Matplotlib backend covering all rendering features, backward-compatibility aliases, and edge cases.
• Added 40 tests for the Plotly backend covering all rendering features and edge cases.
• Added backend selection and resolution logic tests.
• Added --gui interactive gallery viewers for visual test validation (tkinter for Matplotlib, tabbed HTML for Plotly).