[vis] Label axes for points with item names

phi/vis/_matplotlib/_matplotlib_plots.py

@@ -631,23 +631,35 @@ def _plot_points(axis: Axes, data: Field, dims: tuple, vector: Shape, color: Ten
                         axis.scatter(xs, ys, s=3, marker='o', c=col, alpha=alphas)
 
         if any(non_channel(data).item_names):
-            PointCloud2D._annotate_points(axis, data.points, color, alpha)
+            PointCloud2D._annotate_points(axis, data.points, color, alpha, dims)
 
     @staticmethod
-    def _annotate_points(axis, points: math.Tensor, color: Tensor, alpha: Tensor):
-        labelled_dims = non_channel(points)
-        labelled_dims = math.concat_shapes(*[d for d in labelled_dims if d.item_names[0]])
-        if not labelled_dims:
+    def _annotate_points(axis, points: math.Tensor, color: Tensor, alpha: Tensor, dims: Tuple[str], label_axis=True, max_axis_labels=10):
+        labeled_dims = non_channel(points)
+        labeled_dims = math.concat_shapes(*[d for d in labeled_dims if d.item_names[0]])
+        if not labeled_dims:
             return
-        if all(dim.name in points.shape.get_item_names('vector') for dim in labelled_dims):
+        if all(dim.name in points.shape.get_item_names('vector') for dim in labeled_dims):
+            if label_axis:
+                for labeled_dim in labeled_dims:
+                    if len(labeled_dim.item_names[0]) <= max_axis_labels:
+                        which_axis = dims.index(labeled_dim.name)
+                        if which_axis == 0:
+                            axis.set_xticks(reshaped_numpy(points.vector[labeled_dim.name], [shape]))
+                            if axis.get_xscale() == 'log':
+                                axis.get_xaxis().set_major_formatter(matplotlib.ticker.ScalarFormatter())
+                        elif which_axis == 1:
+                            axis.set_yticks(reshaped_numpy(points.vector[labeled_dim.name], [shape]))
+                            if axis.get_yscale() == 'log':
+                                axis.get_yaxis().set_major_formatter(matplotlib.ticker.ScalarFormatter())
             return  # The point labels match one of the figure axes, so they are redundant
         if points.shape['vector'].size == 2:
             xs, ys = reshaped_numpy(points, ['vector', points.shape.without('vector')])
             x_view = axis.get_xlim()[1] - axis.get_xlim()[0]
             y_view = axis.get_ylim()[1] - axis.get_ylim()[0]
             x_c = .95 * axis.get_xlim()[1] + .1 * axis.get_xlim()[0]
             y_c = .95 * axis.get_ylim()[1] + .1 * axis.get_ylim()[0]
-            for x, y, idx, idx_n in zip(xs, ys, labelled_dims.meshgrid(), labelled_dims.meshgrid(names=True)):
+            for x, y, idx, idx_n in zip(xs, ys, labeled_dims.meshgrid(), labeled_dims.meshgrid(names=True)):
                 if axis.get_xscale() == 'log':
                     offset_x = x * (1 + .0003 * x_view) if x < x_c else x * (1 - .0003 * x_view)
                 else: