ENH: first draft of MPL artist

datashader/mpl_ext.py

@@ -0,0 +1,66 @@
+from __future__ import absolute_import
+import datashader as ds
+
+
+import numpy as np
+
+
+import matplotlib.image as mimage
+from matplotlib.transforms import (Bbox, TransformedBbox, BboxTransform)
+
+
+class DSArtist(mimage._ImageBase):
+    def __init__(self, ax, df, x, y, agg=None, **kwargs):
+        super().__init__(ax, **kwargs)
+        self.df = df
+        self.x = x
+        self.y = y
+        self.agg = agg
+        self.axes = ax
+        ax.set_ylim((df[y].min(), df[y].max()))
+        ax.set_xlim((df[x].min(), df[x].max()))
+        self.set_array([[1, 1], [1, 1]])
+
+    def make_image(self, renderer, magnification=1.0,
+                   unsampled=False):
+        trans = self.get_transform()
+        (x1, x2), (y1, y2) = self.axes.get_xlim(), self.axes.get_ylim()
+        bbox = Bbox(np.array([[x1, y1], [x2, y2]]))
+        transformed_bbox = TransformedBbox(bbox, trans)
+
+        dims = self.axes.axesPatch.get_window_extent().bounds
+        plot_width = int(dims[3] + 0.5)
+        plot_height = int(dims[2] + 0.5)
+
+        cvs = ds.Canvas(plot_width=plot_width,
+                        plot_height=plot_height,
+                        x_range=(x1, x2),
+                        y_range=(y1, y2))
+        img = cvs.points(self.df, self.x, self.y, self.agg)
+        img = np.flipud(img)
+        self.set_clim(np.min(img)+1, np.max(img))
+        self.set_array(img)
+        return self._make_image(
+            img, bbox, transformed_bbox, self.axes.bbox, magnification,
+            unsampled=unsampled)
+
+    def get_extent(self):
+        return (*self.axes.get_xlim(), *self.axes.get_ylim())
+
+    def get_cursor_data(self, event):
+        """Get the cursor data for a given event"""
+        xmin, xmax, ymin, ymax = self.get_extent()
+        if self.origin == 'upper':
+            ymin, ymax = ymax, ymin
+        arr = self.get_array()
+        data_extent = Bbox([[ymin, xmin], [ymax, xmax]])
+        array_extent = Bbox([[0, 0], arr.shape[:2]])
+        trans = BboxTransform(boxin=data_extent,
+                              boxout=array_extent)
+        y, x = event.ydata, event.xdata
+        i, j = trans.transform_point([y, x]).astype(int)
+        # Clip the coordinates at array bounds
+        if not (0 <= i < arr.shape[0]) or not (0 <= j < arr.shape[1]):
+            return None
+        else:
+            return arr[i, j]