Switch from fontforge to fonttools

The dependencies of fonttools are much lighter than fontforge, and since all
we need are some metrics, fonttools is very much up to that task.

This addresses issue #288.

metrics/README.md

@@ -7,11 +7,11 @@ There are several requirements for generating the metrics used by KaTeX.
   this by running `tex --version`, and seeing if it has a line that looks like
   > kpathsea version 6.2.0
 
-- You need the JSON module for perl. You can install this either from CPAN or with
+- You need the JSON module for perl. You can install this either from CPAN
-  your package manager.
+  (possibly using the `cpan` command line tool) or with your package manager.
 
-- You need the python fontforge module. This is probably either installed with
+- You need the python module fonttools. You can install this either from PyPi
-  fontforge or can be installed from your package manager.
+  (using `easy_install` or `pip`) or with your package manager.
 
 Once you have these things, run
 

metrics/extract_ttfs.py

@@ -1,6 +1,6 @@
 #!/usr/bin/env python
 
-import fontforge
+from fontTools.ttLib import TTFont
 import sys
 import json
 
@@ -60,33 +60,47 @@ def main():
     start_json = json.load(sys.stdin)
 
     for font, chars in metrics_to_extract.iteritems():
-        fontInfo = fontforge.open("../static/fonts/KaTeX_" + font + ".ttf")
+        fontInfo = TTFont("../static/fonts/KaTeX_" + font + ".ttf")
+        glyf = fontInfo["glyf"]
+        unitsPerEm = float(fontInfo["head"].unitsPerEm)
 
-        for glyph in fontInfo.glyphs():
+        # We keep ALL Unicode cmaps, not just fontInfo["cmap"].getcmap(3, 1).
-            try:
+        # This is playing it extra safe, since it reports inconsistencies.
-                char = unichr(glyph.unicode)
+        # Platform 0 is Unicode, platform 3 is Windows. For platform 3,
-            except ValueError:
+        # encoding 1 is UCS-2 and encoding 10 is UCS-4.
+        cmap = [t.cmap for t in fontInfo["cmap"].tables
+                if (t.platformID == 0)
+                or (t.platformID == 3 and t.platEncID in (1, 10))]
+
+        for char, base_char in chars.iteritems():
+            code = ord(char)
+            names = set(t.get(code) for t in cmap)
+            if not names:
+                sys.stderr.write(
+                    "Codepoint {} of font {} maps to no name\n"
+                    .format(code, font))
                 continue
+            if len(names) != 1:
+                sys.stderr.write(
+                    "Codepoint {} of font {} maps to multiple names: {}\n"
+                    .format(code, font, ", ".join(sorted(names))))
+                continue
+            name = names.pop()
 
-            if char in chars:
+            height = depth = italic = skew = 0
-                _, depth, _, height = glyph.boundingBox()
+            glyph = glyf[name]
-
+            if glyph.numberOfContours:
-                depth = -depth
+                height = glyph.yMax
-
+                depth = -glyph.yMin
-                base_char = chars[char]
             if base_char:
                 base_char_str = str(ord(base_char))
                 base_metrics = start_json[font][base_char_str]
-
                 italic = base_metrics["italic"]
                 skew = base_metrics["skew"]
-                else:
-                    italic = 0
-                    skew = 0
 
-                start_json[font][str(ord(char))] = {
+            start_json[font][str(code)] = {
-                    "height": height / fontInfo.em,
+                "height": height / unitsPerEm,
-                    "depth": depth / fontInfo.em,
+                "depth": depth / unitsPerEm,
                 "italic": italic,
                 "skew": skew,
             }

src/fontMetricsData.json

@@ -632,7 +632,7 @@
   "8463": {"depth": 0.0, "height": 0.68889, "italic": 0.0, "skew": 0.0}
 },
 "Main-Regular": {
-  "32": {"depth": -0.0, "height": 0.0, "italic": 0, "skew": 0},
+  "32": {"depth": 0.0, "height": 0.0, "italic": 0, "skew": 0},
   "33": {"depth": 0.0, "height": 0.69444, "italic": 0.0, "skew": 0.0},
   "34": {"depth": 0.0, "height": 0.69444, "italic": 0.0, "skew": 0.0},
   "35": {"depth": 0.19444, "height": 0.69444, "italic": 0.0, "skew": 0.0},
@@ -727,7 +727,7 @@
   "124": {"depth": 0.25, "height": 0.75, "italic": 0.0, "skew": 0.0},
   "125": {"depth": 0.25, "height": 0.75, "italic": 0.0, "skew": 0.0},
   "126": {"depth": 0.35, "height": 0.31786, "italic": 0.0, "skew": 0.0},
-  "160": {"depth": -0.0, "height": 0.0, "italic": 0, "skew": 0},
+  "160": {"depth": 0.0, "height": 0.0, "italic": 0, "skew": 0},
   "168": {"depth": 0.0, "height": 0.66786, "italic": 0.0, "skew": 0.0},
   "172": {"depth": 0.0, "height": 0.43056, "italic": 0.0, "skew": 0.0},
   "175": {"depth": 0.0, "height": 0.56778, "italic": 0.0, "skew": 0.0},
@@ -778,7 +778,7 @@
   "8221": {"depth": 0.0, "height": 0.69444, "italic": 0.0, "skew": 0.0},
   "8224": {"depth": 0.19444, "height": 0.69444, "italic": 0.0, "skew": 0.0},
   "8225": {"depth": 0.19444, "height": 0.69444, "italic": 0.0, "skew": 0.0},
-  "8230": {"depth": -0.0, "height": 0.12, "italic": 0, "skew": 0},
+  "8230": {"depth": 0.0, "height": 0.12, "italic": 0, "skew": 0},
   "8242": {"depth": 0.0, "height": 0.55556, "italic": 0.0, "skew": 0.0},
   "8407": {"depth": 0.0, "height": 0.71444, "italic": 0.15382, "skew": 0.0},
   "8463": {"depth": 0.0, "height": 0.68889, "italic": 0.0, "skew": 0.0},