raphaelquast/EOmaps: EOmaps v5.4

A minor release with some awesome new features!

🌦️ changes

• ❗ m.plot_map(pick_distance=...) is depreciated, use m.cb.pick.set_props(search_radius=...) instead ❗

🌳 New 🌲 Improvements for PICK callbacks

The internals for pick-callbacks have been re-worked to improve performance and incorporate some new features.

• you can now use m.cb.pick.set_props(...) to set the picking behavior for each Maps-object

🌱 Pick N nearest neighbors

• Starting with EOmaps v5.4 it is possible to pick N nearest neighbors instead of a single point!
• Checkout the docs for more details: 🍭Picking N nearest neighbours

  m = Maps()
  m.set_data([1,2,3,4,5], [1,2,3,4,5], [1,2,3,4,5])
  m.plot_map()
  m.cb.pick.set_props(n=3, pick_relative_to_closest=True, consecutive_pick=False)
  m.cb.pick.attach.mark(fc="none", ec="r")
  

🌲 Improvements for shapes 🌱 There is a new plot-shape: m.set_shape.scatter_points

• basically the same as using plt.scatter() but supercharged so that EOmaps functionalities work as expected
• use it to draw datapoints with dimensions defined in (figure-points)² (rather than actual physical dimensions)

  m = Maps()
  m.add_feature.preset.coastline()
  m.set_data([1, 2, 3], [-45, 20, 52], [-10, 20, 30])
  m.set_shape.scatter_points(size=[200, 50, 500], marker="*")
  m.plot_map(cmap="tab10")
  

🌱 The shape-radius can now be defined for each datapoint individually!

• The radius argument of the shapes ellipses, rectangles and geod_circles now properly broadcasts multiple values (e.g. you can assign an individual radius for each datapoint)

  m = Maps()
  m.add_feature.preset.coastline()
  m.set_data([1, 2, 3], [-45, 20, 52], [-10, 20, 30])
  m.set_shape.ellipses(radius=[1, 2, 3], radius_crs=4326)
  m.plot_map()
  

🔨 fixes

• fix parsing n=None as shape-argument in the companion-widget
• allow using the companion-widget with matplotlibs "QtAgg" backend
• fix performance issues with m.get_crs
• fix using m.BM.blit_artist if the renderer is not yet cached
• make evaluation of scalebar background patch size more robust
• ensure that scalebar is correctly updated if properties change
• switch to mambaforge to avoid installation-issues in github-actions
• lazily initialize pick-search tree only if a pick callback is attached
  (provides a nice speedup for large datasets if no pick callbacks are assigned)
• allow passing None to m._decode_values
• fix picking GeoDataFrames if no value-column was assigned (e.g. when using m.add_gdf(column=None))
• remove obsolete dependencies and unused requirements.txt from the docs