Add support for custom style sheets

Author: dstansby

docs/user_guide.rst

@@ -36,3 +36,14 @@ To use these:
 1. Open the desired widget using the ``Plugins > napari-matplotlib`` menu in napari.
 2. Select a single layer that has a features table using the napari layers list in the bottom left-hand side of the window.
 3. Use the drop down menu(s) under the Matplotlib figure to select the feature(s) to plot.
+
+Customising plots
+-----------------
+`Matplotlib style sheets <https://matplotlib.org/stable/tutorials/introductory/customizing.html#defining-your-own-style>`__ can be used to customise
+the plots generated by ``napari-matplotlib``.
+To use a custom style sheet:
+
+1. Save it as ``napari-matplotlib.mplstyle``
+2. Put it in the Matplotlib configuration directory.
+   The location of this directory varies on different computers,
+   and can be found by calling :func:`matplotlib.get_configdir()`.

src/napari_matplotlib/base.py

@@ -2,6 +2,7 @@
 from pathlib import Path
 from typing import List, Optional, Tuple
 
+import matplotlib
 import matplotlib.style as mplstyle
 import napari
 from matplotlib.backends.backend_qtagg import (
@@ -16,6 +17,10 @@
 
 __all__ = ["BaseNapariMPLWidget", "NapariMPLWidget", "SingleAxesWidget"]
 
+_CUSTOM_STYLE_PATH = (
+    Path(matplotlib.get_configdir()) / "napari-matplotlib.mplstyle"
+)
+
 
 class BaseNapariMPLWidget(QWidget):
     """
@@ -46,7 +51,6 @@ def __init__(
         with mplstyle.context(self.mpl_style_sheet_path):
             self.canvas = FigureCanvas()
 
-        self.canvas.figure.patch.set_facecolor("none")
         self.canvas.figure.set_layout_engine("constrained")
         self.toolbar = NapariNavigationToolbar(
             self.canvas, parent=self
@@ -73,6 +77,8 @@ def mpl_style_sheet_path(self) -> Path:
         """
         if self._mpl_style_sheet_path is not None:
             return self._mpl_style_sheet_path
+        elif (_CUSTOM_STYLE_PATH).exists():
+            return _CUSTOM_STYLE_PATH
         elif self._napari_theme_has_light_bg():
             return Path(__file__).parent / "styles" / "light.mplstyle"
         else:

src/napari_matplotlib/tests/test_theme.py

@@ -1,6 +1,5 @@
-from copy import deepcopy
-
 import shutil
+from copy import deepcopy
 from pathlib import Path
 
 import matplotlib
@@ -152,27 +151,36 @@ def find_mpl_stylesheet(name: str) -> Path:
     return Path(matplotlib.__path__[0]) / f"mpl-data/stylelib/{name}.mplstyle"
 
 
-def test_stylesheet_in_cwd(tmpdir, make_napari_viewer, image_data):
+def test_custom_stylesheet(make_napari_viewer, image_data):
     """
     Test that a stylesheet in the current directory is given precidence.
 
     Do this by copying over a stylesheet from matplotlib's built in styles,
     naming it correctly, and checking the colours are as expected.
     """
-    with tmpdir.as_cwd():
-        # Copy Solarize_Light2 to current dir as if it was a user-overriden stylesheet.
-        shutil.copy(find_mpl_stylesheet("Solarize_Light2"), "./user.mplstyle")
-        viewer = make_napari_viewer()
-        viewer.add_image(image_data[0], **image_data[1])
-        widget = HistogramWidget(viewer)
-        ax = widget.figure.gca()
-
-        # The axes should have a light brownish grey background:
-        assert ax.get_facecolor() == to_rgba("#eee8d5")
-        assert ax.patch.get_facecolor() == to_rgba("#eee8d5")
-
-        # The figure background and axis gridlines are light yellow:
-        assert widget.figure.patch.get_facecolor() == to_rgba("#fdf6e3")
-        for gridline in ax.get_xgridlines() + ax.get_ygridlines():
-            assert gridline.get_visible() is True
-            assert gridline.get_color() == "#fdf6e3"
+    # Copy Solarize_Light2 as if it was a user-overriden stylesheet.
+    style_sheet_path = (
+        Path(matplotlib.get_configdir()) / "napari-matplotlib.mplstyle"
+    )
+    if style_sheet_path.exists():
+        pytest.skip("Won't ovewrite existing custom style sheet.")
+    shutil.copy(
+        find_mpl_stylesheet("Solarize_Light2"),
+        style_sheet_path,
+    )
+
+    viewer = make_napari_viewer()
+    viewer.add_image(image_data[0], **image_data[1])
+    widget = HistogramWidget(viewer)
+    assert widget.mpl_style_sheet_path == style_sheet_path
+    ax = widget.figure.gca()
+
+    # The axes should have a light brownish grey background:
+    assert ax.get_facecolor() == to_rgba("#eee8d5")
+    assert ax.patch.get_facecolor() == to_rgba("#eee8d5")
+
+    # The figure background and axis gridlines are light yellow:
+    assert widget.figure.patch.get_facecolor() == to_rgba("#fdf6e3")
+    for gridline in ax.get_xgridlines() + ax.get_ygridlines():
+        assert gridline.get_visible() is True
+        assert gridline.get_color() == "#fdf6e3"