Construct glyph data from unicodes as a fallback

Lib/glyphsLib/builder/features.py

@@ -170,7 +170,7 @@ def _build_gdef(ufo, skipExportGlyphs=None):
 
         # First check glyph.lib for category/subCategory overrides. Otherwise,
         # use global values from GlyphData.
-        glyphinfo = glyphdata.get_glyph(glyph.name)
+        glyphinfo = glyphdata.get_glyph(glyph.name, [f"{c:04X}" for c in glyph.unicodes])
         category = glyph.lib.get(category_key) or glyphinfo.category
         subCategory = glyph.lib.get(subCategory_key) or glyphinfo.subCategory
 

Lib/glyphsLib/builder/glyph.py

@@ -143,7 +143,7 @@ def to_ufo_glyph(self, ufo_glyph, layer, glyph):  # noqa: C901
             ufo_glyph.lib[GLYPHLIB_PREFIX + "Export"] = export
 
     # FIXME: (jany) next line should be an API of GSGlyph?
-    glyphinfo = glyphsLib.glyphdata.get_glyph(ufo_glyph.name)
+    glyphinfo = glyphsLib.glyphdata.get_glyph(ufo_glyph.name, glyph.unicodes)
     if glyph.production:
         production_name = glyph.production
         # Make sure production names of bracket glyphs also get a BRACKET suffix.
@@ -283,7 +283,7 @@ def to_glyphs_glyph(self, ufo_glyph, ufo_layer, master):  # noqa: C901
         # glyphinfo = glyphsLib.glyphdata.get_glyph(ufo_glyph.name)
         # production_name = glyph.production or glyphinfo.production_name
 
-    glyphinfo = glyphsLib.glyphdata.get_glyph(ufo_glyph.name)
+    glyphinfo = glyphsLib.glyphdata.get_glyph(ufo_glyph.name, glyph.unicodes)
 
     layer = self.to_glyphs_layer(ufo_layer, glyph, master)
 

Lib/glyphsLib/glyphdata.py

@@ -50,26 +50,29 @@ class GlyphData:
     dictionaries.
     """
 
-    __slots__ = ["names", "alternative_names", "production_names"]
+    __slots__ = ["names", "alternative_names", "production_names", "unicodes"]
 
-    def __init__(self, name_mapping, alt_name_mapping, production_name_mapping):
+    def __init__(self, name_mapping, alt_name_mapping, production_name_mapping, unicodes_mapping):
         self.names = name_mapping
         self.alternative_names = alt_name_mapping
         self.production_names = production_name_mapping
+        self.unicodes = unicodes_mapping
 
     @classmethod
     def from_files(cls, *glyphdata_files):
         """Return GlyphData holding data from a list of XML file paths."""
         name_mapping = {}
         alt_name_mapping = {}
         production_name_mapping = {}
+        unicodes_mapping = {}
 
         for glyphdata_file in glyphdata_files:
             glyph_data = xml.etree.ElementTree.parse(glyphdata_file).getroot()
             for glyph in glyph_data:
                 glyph_name = glyph.attrib["name"]
                 glyph_name_alternatives = glyph.attrib.get("altNames")
                 glyph_name_production = glyph.attrib.get("production")
+                glyph_unicode = glyph.attrib.get("unicode")
 
                 name_mapping[glyph_name] = glyph.attrib
                 if glyph_name_alternatives:
@@ -78,16 +81,18 @@ def from_files(cls, *glyphdata_files):
                         alt_name_mapping[glyph_name_alternative] = glyph.attrib
                 if glyph_name_production:
                     production_name_mapping[glyph_name_production] = glyph.attrib
+                if glyph_unicode:
+                    unicodes_mapping[glyph_unicode] = glyph.attrib
 
-        return cls(name_mapping, alt_name_mapping, production_name_mapping)
+        return cls(name_mapping, alt_name_mapping, production_name_mapping, unicodes_mapping)
 
 
-def get_glyph(glyph_name, data=None):
+def get_glyph(glyph_name, unicodes=[], data=None):
     """Return a named tuple (Glyph) containing information derived from a glyph
     name akin to GSGlyphInfo.
 
     The information is derived from an included copy of GlyphData.xml
-    and GlyphData_Ideographs.xml, going purely by the glyph name.
+    and GlyphData_Ideographs.xml, going by the glyph name or unicode fallback.
     """
 
     # Read data on first use.
@@ -105,6 +110,13 @@ def get_glyph(glyph_name, data=None):
     # Look up data by full glyph name first.
     attributes = _lookup_attributes(glyph_name, data)
 
+    # Look up by unicode
+    if attributes == {} and unicodes:
+        for unicode in unicodes:
+            attributes = _lookup_attributes_by_unicode(unicode, data)
+            if attributes:
+                break
+
     production_name = attributes.get("production")
     if production_name is None:
         production_name = _construct_production_name(glyph_name, data=data)
@@ -147,6 +159,14 @@ def _lookup_attributes(glyph_name, data):
     return attributes
 
 
+def _lookup_attributes_by_unicode(unicode, data):
+    """Look up glyph attributes in data by unicode
+    or return empty dictionary.
+    """
+    attributes = data.unicodes.get(unicode) or {}
+    return attributes
+
+
 def _agl_compliant_name(glyph_name):
     """Return an AGL-compliant name string or None if we can't make one."""
     MAX_GLYPH_NAME_LENGTH = 63

tests/glyphdata_test.py

@@ -139,6 +139,13 @@ def cat(n):
         self.assertEqual(cat("dal_alef-ar"), ("Letter", "Ligature"))
         self.assertEqual(cat("dal_lam-ar.dlig"), ("Letter", "Ligature"))
 
+    def test_category_buy_unicode(self):
+        def cat(n, u):
+            return get_glyph(n, u).category, get_glyph(n, u).subCategory
+
+        # "SignU.bn" is a non-standard name not defined in GlyphData.xml
+        self.assertEqual(cat("SignU.bn", ["09C1"]), ("Mark", "Nonspacing"))
+
     def test_bug232(self):
         # https://github.com/googlefonts/glyphsLib/issues/232
         u, g = get_glyph("uni07F0"), get_glyph("longlowtonecomb-nko")