Skip to content

Latest commit

 

History

History
166 lines (115 loc) · 7.62 KB

CHANGELOG.md

File metadata and controls

166 lines (115 loc) · 7.62 KB
Raw

Changelog

v0.6.0 (unreleased)

Breaking changes

  • Removed support for python 3.9 (#130).

  • The minimum versions of some dependencies were changed (#132).

    Package Old New
    cartopy 0.21 0.22
    matplotlib 3.6 3.8
    numpy 1.22 1.24
    seaborn 0.12 0.13
    xarray 2022.12 2023.9

  • The modules cartopy_utils, colormaps, map_layout, mpl, and xrcompat were renamed (added a leading underscore) to indicate that they are private (#141 and #142).

Enhancements

  • Added convenience functions to draw hatches and add stippling:

    1. mpu.hatch: for regular axes
    2. mpu.hatch_map: for cartopy GeoAxes
    3. mpu.hatch_map_global: as 2. but also adds a cyclic point to the array

    all three functions expect a 2D boolean xr.DataArray and a hatch pattern. Values that are True are hatched (#123 and #143).

  • Enable passing AxesGrid (from mpl_toolkits.axes_grid1) to set_map_layout (#116).

  • Raise more informative error when a wrong type is passed to set_map_layout (#121).

  • set_map_layout now raises an explicit error when the figure contains SubFigure (#121).

  • Test upstream dependencies and fix compatibility with the upcoming pandas v3 (#133).

Bug fixes

  • Ensure the current axes (plt.gca()) is not changed by calling mpu.colorbar(...) (#136).

Internal changes

v0.5.0 (27.03.2024)

Version v0.5.0 aligns passing multiple axes to colorbar with matplotlib.

Breaking changes

  • The ax1 and ax2 arguments of mpu.colorbar have been combined into ax (#107) To update

    - mpu.colorbar(h, axs[0], axs[1])
+ mpu.colorbar(h, [axs[0], axs[1]])

    or

    - mpu.colorbar(h, axs[0], axs[1])
+ mpu.colorbar(h, axs)

  • When passing size to mpu.colorbar it now uses the height/ width of all passed axes to scale the colorbar. This is consistent with plt.colorbar but may lead to differences compared to the previous version (#107).

  • Similarly for pad, which is also scaled by the height/ width of all passed axes. This is consistent with plt.colorbar but may change the padding of the colorbar compared to the previous version (#107).

v0.4.0 (23.02.2024)

Version 0.4.0 simplifies the way figures with a mpu.colorbar have to be saved and bumps the supported versions.

Breaking changes

  • Removed support for python 3.7 and python 3.8 (#88).

  • The minimum versions of some dependencies were changed (#88, #91, and #96).

    Package Old New
    cartopy 0.18 0.21
    matplotlib* 3.4 3.6
    numpy 1.17 1.22
    seaborn 0.11 0.12
    xarray 0.15 2022.12

  • sample_data_map now offsets the lon data such that the first grid cell does not wrap around (#87).

Enhancements

  • Calling f.canvas.draw() before plt.savefig is no longer necessary. This now happens automatically (#98).
  • Add python 3.12 to list of supported versions (#89).

v0.3.1 (09.02.2023)

Version v0.3.1 fixes a regression and an additional bug from v0.3.0. It's recommended to use this version.

Bug fixes

  • Fixed a bug in set_map_layout: the data ratio of a cartopy GeoAxesSubplot requires a draw to be correct (#61).
  • Fix a regression introduced in #33: cyclic_dataarray now correctly extrapolates the coordinates (#58).

v0.3.0 (15.01.2023)

Version 0.3.0 includes some long-overdue bug fixes, modernization of the code, much better test coverage and some enhancements.

mplotutils now uses the MIT license instead of GPL-3.0 (#51)

Deprecations

  • Deprecated mpu.infer_interval_breaks as this is no longer necessary with matplotlib v3.2 and cartopy v0.21 (#32).
  • Deprecated a number of positional arguments, these are now keyword only, e.g. in mpu.colorbar (#54).

Enhancements

  • Added nrow and ncol parameters to set_map_layout for use with a gridspec.
  • Allow passing list of axes to set_map_layout, renamed the files and extended the test coverage (#42 and #43).
  • Add function to create xr.DataArray sample data (#53).

Bug fixes

  • Fixed compatibility of mpu.colorbar with bbox_inches="tight" for matplotlib 3.4 and newer and refactor and extend tests (#26 and #40).

Internal changes

  • Replaced ax.get_geometry() with ax.get_subplotspec().get_geometry() as the former was deprecated in matplotlib (#8).
  • Refactor mpu.cyclic_dataarray using obj.pad (#33).
  • Enabled CI on github actions (#9).
  • Formatted with black and isort, checked with flake8.

v0.2.0 (01.06.2018)

Mayor release, mostly introducing the new colorbar functionality.

  • add _color_palette that selects colors from the whole range of the colormap. Previously we used the seaborn logic that excludes colors at the edge of the colorbar (#1).
  • the colorbar functionality was entirely redesigned:
    • new top-level function mpu.colorbar
    • added aspect keyword which defines the ratio of long to short side
    • the default width/ height is now aspect=20 (was size=0.04)
    • pad and size are now scaled by the width/ height of the axes (was width/ height of the figure)
    • pad now uses the matplotlib default (0.05 for vertical bars and 0.15 for horizontal colorbars)
  • add _is_monotonic to infer_interval_breaks
  • renamed sample_data_3d to sample_data_map

v0.1.0 (13.03.2018)

  • Functionality as developed for pyvis (see below), packed into a module.