René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"Simplify glyph stream logic in ps backend ","articleBody":"Seems good, but perhaps simpler to use a groupby here?  Something like (untested)\r\n```patch\r\ndiff --git a/lib/matplotlib/backends/backend_ps.py b/lib/matplotlib/backends/backend_ps.py\r\nindex f209e811f1..4d31fd881a 100644\r\n--- a/lib/matplotlib/backends/backend_ps.py\r\n+++ b/lib/matplotlib/backends/backend_ps.py\r\n@@ -7,6 +7,7 @@ import datetime\r\n from enum import Enum\r\n import functools\r\n from io import StringIO\r\n+import itertools\r\n import logging\r\n import os\r\n import pathlib\r\n@@ -626,10 +627,13 @@ grestore\r\n         if ismath:\r\n             return self.draw_mathtext(gc, x, y, s, prop, angle)\r\n \r\n+        stream = []  # list of (ps_name, x, char_name)\r\n+\r\n         if mpl.rcParams['ps.useafm']:\r\n             font = self._get_font_afm(prop)\r\n+            ps_name = (font.postscript_name\r\n+                       .encode(\"ascii\", \"replace\").decode(\"ascii\"))\r\n             scale = 0.001 * prop.get_size_in_points()\r\n-            stream = []\r\n             thisx = 0\r\n             last_name = None  # kerns returns 0 for None.\r\n             xs_names = []\r\n@@ -643,37 +647,24 @@ grestore\r\n                 kern = font.get_kern_dist_from_name(last_name, name)\r\n                 last_name = name\r\n                 thisx += kern * scale\r\n-                xs_names.append((thisx, name))\r\n+                xs_names.append((ps_name, thisx, name))\r\n                 thisx += width * scale\r\n-            ps_name = (font.postscript_name\r\n-                       .encode(\"ascii\", \"replace\").decode(\"ascii\"))\r\n-            stream.append((ps_name, xs_names))\r\n \r\n         else:\r\n             font = self._get_font_ttf(prop)\r\n             self._character_tracker.track(font, s)\r\n-            stream = []\r\n-            prev_font = curr_stream = None\r\n             for item in _text_helpers.layout(s, font):\r\n                 ps_name = (item.ft_object.postscript_name\r\n                            .encode(\"ascii\", \"replace\").decode(\"ascii\"))\r\n-                if item.ft_object is not prev_font:\r\n-                    if curr_stream:\r\n-                        stream.append(curr_stream)\r\n-                    prev_font = item.ft_object\r\n-                    curr_stream = [ps_name, []]\r\n-                curr_stream[1].append(\r\n-                    (item.x, item.ft_object.get_glyph_name(item.glyph_idx))\r\n-                )\r\n-            # append the last entry\r\n-            stream.append(curr_stream)\r\n+                glyph_name = item.ft_object.get_glyph_name(item.glyph_idx)\r\n+                stream.append((ps_name, item.x, glyph_name))\r\n \r\n         self.set_color(*gc.get_rgb())\r\n \r\n-        for ps_name, xs_names in stream:\r\n+        for ps_name, group in itertools.groupby(stream, lambda entry: entry[0]):\r\n             self.set_font(ps_name, prop.get_size_in_points(), False)\r\n             thetext = \"\\n\".join(f\"{x:g} 0 m /{name:s} glyphshow\"\r\n-                                for x, name in xs_names)\r\n+                                for _, x, name in group)\r\n             self._pswriter.write(f\"\"\"\\\r\n gsave\r\n {self._get_clip_cmd(gc)}\r\n```\r\n\r\nEdit: In fact set_font itself already checks if the font has already been set previously and skips redundant setfont calls, so we could just entirely get rid of the groupby logic too.\r\n\r\n_Originally posted by @anntzer in https://github.com/matplotlib/matplotlib/issues/23910#issuecomment-1252201631_","author":{"url":"https://github.com/tacaswell","@type":"Person","name":"tacaswell"},"datePublished":"2022-09-20T14:28:30.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":5},"url":"https://github.com/23965/matplotlib/issues/23965"}