Update mpl_draw() to fix multigraph plots (#1204)

* Update #1: Fixing  mpl_draw() for multigraphs

* Update matplotlib.py for formatting

* Update rustworkx/visualization/matplotlib.py

Co-authored-by: Ivan Carvalho <[email protected]>

* Update matplotlib.py to remove the loop

* Add releasenotes

* Reformat connectionstyle string in rustworkx/visualization/matplotlib.py

Co-authored-by: Ivan Carvalho <[email protected]>

* Fixes #774

* Optimize edge search by using sets

---------

Co-authored-by: Ivan Carvalho <[email protected]>

releasenotes/notes/fix-mpl-draw-digraph-plots-aecf86738ab9b0db.yaml

@@ -0,0 +1,29 @@
+---
+fixes:
+  - |
+    Fixed the plots of multigraphs using :func:`.mpl_draw`. Previously, parallel edges of 
+    multigraphs were plotted on top of each other, with overlapping arrows and labels. 
+    The radius of parallel edges of the multigraph was fixed to be `0.25` for 
+    `connectionstyle` supporting this argument in :func:`.draw_edges`. The edge lables 
+    were offset to `0.25` in :func:`.draw_edge_labels` to align with their respective 
+    edges. This fix can be tested using the following code: 
+
+        .. jupyter-execute::
+
+            import rustworkx
+            from rustworkx.visualization import mpl_draw
+
+            graph = rustworkx.PyDiGraph()
+            graph.add_node('A')
+            graph.add_node('B')
+            graph.add_node('C')
+
+            graph.add_edge(1, 0, 2)
+            graph.add_edge(0, 1, 3)
+            graph.add_edge(1, 2, 4)
+
+            mpl_draw(graph, with_labels=True, labels=str, edge_labels=str, alpha=0.5)
+
+  - |
+    Refer to `#774 <https://github.com/Qiskit/rustworkx/issues/774>` for more
+    details.

rustworkx/visualization/matplotlib.py

@@ -636,7 +636,9 @@ def draw_edges(
         edge_color = "k"
 
     # set edge positions
-    edge_pos = np.asarray([(pos[e[0]], pos[e[1]]) for e in edge_list])
+    edge_pos = set()
+    for e in edge_list:
+        edge_pos.add((tuple(pos[e[0]]), tuple(pos[e[1]])))
 
     # Check if edge_color is an array of floats and map to edge_cmap.
     # This is the only case handled differently from matplotlib
@@ -670,58 +672,17 @@ def to_marker_edge(marker_size, marker):
     arrow_collection = []
     mutation_scale = arrow_size  # scale factor of arrow head
 
-    # compute view
-    mirustworkx = np.amin(np.ravel(edge_pos[:, :, 0]))
-    maxx = np.amax(np.ravel(edge_pos[:, :, 0]))
-    miny = np.amin(np.ravel(edge_pos[:, :, 1]))
-    maxy = np.amax(np.ravel(edge_pos[:, :, 1]))
-    w = maxx - mirustworkx
-    h = maxy - miny
-
     base_connectionstyle = mpl.patches.ConnectionStyle(connectionstyle)
 
     # Fallback for self-loop scale. Left outside of _connectionstyle so it is
     # only computed once
     max_nodesize = np.array(node_size).max()
 
-    def _connectionstyle(posA, posB, *args, **kwargs):
-        # check if we need to do a self-loop
-        if np.all(posA == posB):
-            # Self-loops are scaled by view extent, except in cases the extent
-            # is 0, e.g. for a single node. In this case, fall back to scaling
-            # by the maximum node size
-            selfloop_ht = 0.005 * max_nodesize if h == 0 else h
-            # this is called with _screen space_ values so covert back
-            # to data space
-            data_loc = ax.transData.inverted().transform(posA)
-            v_shift = 0.1 * selfloop_ht
-            h_shift = v_shift * 0.5
-            # put the top of the loop first so arrow is not hidden by node
-            path = [
-                # 1
-                data_loc + np.asarray([0, v_shift]),
-                # 4 4 4
-                data_loc + np.asarray([h_shift, v_shift]),
-                data_loc + np.asarray([h_shift, 0]),
-                data_loc,
-                # 4 4 4
-                data_loc + np.asarray([-h_shift, 0]),
-                data_loc + np.asarray([-h_shift, v_shift]),
-                data_loc + np.asarray([0, v_shift]),
-            ]
-
-            ret = mpl.path.Path(ax.transData.transform(path), [1, 4, 4, 4, 4, 4, 4])
-        # if not, fall back to the user specified behavior
-        else:
-            ret = base_connectionstyle(posA, posB, *args, **kwargs)
-
-        return ret
-
     # FancyArrowPatch doesn't handle color strings
     arrow_colors = mpl.colors.colorConverter.to_rgba_array(edge_color, alpha)
-    for i, (src, dst) in enumerate(edge_pos):
-        x1, y1 = src
-        x2, y2 = dst
+    for i, edge in enumerate(edge_pos):
+        x1, y1 = edge[0][0], edge[0][1]
+        x2, y2 = edge[1][0], edge[1][1]
         shrink_source = 0  # space from source to tail
         shrink_target = 0  # space from  head to target
         if np.iterable(node_size):  # many node sizes
@@ -754,6 +715,12 @@ def _connectionstyle(posA, posB, *args, **kwargs):
         else:
             line_width = width
 
+        # radius of edges
+        if tuple(reversed(edge)) in edge_pos:
+            rad = 0.25
+        else:
+            rad = 0.0
+
         arrow = mpl.patches.FancyArrowPatch(
             (x1, y1),
             (x2, y2),
@@ -763,14 +730,57 @@ def _connectionstyle(posA, posB, *args, **kwargs):
             mutation_scale=mutation_scale,
             color=arrow_color,
             linewidth=line_width,
-            connectionstyle=_connectionstyle,
+            connectionstyle=connectionstyle + f", rad = {rad}",
             linestyle=style,
             zorder=1,
         )  # arrows go behind nodes
 
         arrow_collection.append(arrow)
         ax.add_patch(arrow)
 
+    edge_pos = np.asarray(tuple(edge_pos))
+
+    # compute view
+    mirustworkx = np.amin(np.ravel(edge_pos[:, :, 0]))
+    maxx = np.amax(np.ravel(edge_pos[:, :, 0]))
+    miny = np.amin(np.ravel(edge_pos[:, :, 1]))
+    maxy = np.amax(np.ravel(edge_pos[:, :, 1]))
+    w = maxx - mirustworkx
+    h = maxy - miny
+
+    def _connectionstyle(posA, posB, *args, **kwargs):
+        # check if we need to do a self-loop
+        if np.all(posA == posB):
+            # Self-loops are scaled by view extent, except in cases the extent
+            # is 0, e.g. for a single node. In this case, fall back to scaling
+            # by the maximum node size
+            selfloop_ht = 0.005 * max_nodesize if h == 0 else h
+            # this is called with _screen space_ values so covert back
+            # to data space
+            data_loc = ax.transData.inverted().transform(posA)
+            v_shift = 0.1 * selfloop_ht
+            h_shift = v_shift * 0.5
+            # put the top of the loop first so arrow is not hidden by node
+            path = [
+                # 1
+                data_loc + np.asarray([0, v_shift]),
+                # 4 4 4
+                data_loc + np.asarray([h_shift, v_shift]),
+                data_loc + np.asarray([h_shift, 0]),
+                data_loc,
+                # 4 4 4
+                data_loc + np.asarray([-h_shift, 0]),
+                data_loc + np.asarray([-h_shift, v_shift]),
+                data_loc + np.asarray([0, v_shift]),
+            ]
+
+            ret = mpl.path.Path(ax.transData.transform(path), [1, 4, 4, 4, 4, 4, 4])
+        # if not, fall back to the user specified behavior
+        else:
+            ret = base_connectionstyle(posA, posB, *args, **kwargs)
+
+        return ret
+
     # update view
     padx, pady = 0.05 * w, 0.05 * h
     corners = (mirustworkx - padx, miny - pady), (maxx + padx, maxy + pady)
@@ -1001,6 +1011,12 @@ def draw_edge_labels(
             x1 * label_pos + x2 * (1.0 - label_pos),
             y1 * label_pos + y2 * (1.0 - label_pos),
         )
+        if (n2, n1) in labels.keys():  # loop
+            dy = np.abs(y2 - y1)
+            if n2 > n1:
+                y -= 0.25 * dy
+            else:
+                y += 0.25 * dy
 
         if rotate:
             # in degrees