From 9f4cc2f1cb8cdc2d86c102c13a2e9f95f1a44580 Mon Sep 17 00:00:00 2001
From: Simon Cozens <simon@simon-cozens.org>
Date: Tue, 7 Jul 2020 12:38:15 +0100
Subject: [PATCH] Introduce the concept of a ClassContextualRuleSet

Currently unused but will be super helpful for optimizing the formats of contextual lookups.

* Splits the rules of a class contextual lookup based on explicit subtable breaks
* Returns various properties on the ruleset to help determine appropriate layout format.
* (More properties, such as "touched glyphs", planned - will be added when needed.)
---
 Lib/fontTools/otlLib/builder.py | 59 +++++++++++++++++++++++++++++++++
 Tests/otlLib/builder_test.py    | 57 +++++++++++++++++++++++++++++++
 2 files changed, 116 insertions(+)

diff --git a/Lib/fontTools/otlLib/builder.py b/Lib/fontTools/otlLib/builder.py
index 5731f51c4..968a327a6 100644
--- a/Lib/fontTools/otlLib/builder.py
+++ b/Lib/fontTools/otlLib/builder.py
@@ -244,11 +244,70 @@ class AlternateSubstBuilder(LookupBuilder):
         self.alternates[(self.SUBTABLE_BREAK_, location)] = self.SUBTABLE_BREAK_
 
 
+class ChainContextualRuleset():
+    def __init__(self):
+        self.rules = []
+
+    def addRule(self, prefix, glyphs, suffix, lookups):
+        self.rules.append((prefix, glyphs, suffix, lookups))
+
+    @property
+    def hasPrefixOrSuffix(self):
+        # Do we have any prefixes/suffixes? If this is False for all
+        # rulesets, we can express the whole lookup as GPOS5/GSUB7.
+        for (prefix, glyphs, suffix, lookups) in self.rules:
+            if len(prefix) > 0 or len(suffix) > 0: return True
+        return False
+
+    @property
+    def hasAnyGlyphClasses(self):
+        # Do we use glyph classes anywhere in the rules? If this is False
+        # we can express this subtable as a Format 1.
+        for (prefix, glyphs, suffix, lookups) in self.rules:
+            for coverage in (prefix, glyphs, suffix):
+                if any([len(x) > 1 for x in coverage]):
+                    return True
+        return False
+
+    def format1Classdefs(self):
+        PREFIX, GLYPHS, SUFFIX = 0,1,2
+        classdefbuilders = []
+        for ix in [PREFIX, GLYPHS, SUFFIX]:
+            context = []
+            for r in self.rules:
+                context.append(r[ix])
+            classes = self._classBuilderForContext(context)
+            if not classes:
+                return None
+            classdefbuilders.append(classes)
+        return classdefbuilders
+
+    def _classBuilderForContext(self, context):
+        classdefbuilder = ClassDefBuilder(useClass0=False)
+        for position in context:
+            for glyphset in position:
+                if not classdefbuilder.canAdd(glyphset):
+                    return None
+                classdefbuilder.add(glyphset)
+        return classdefbuilder
+
 class ChainContextualBuilder(LookupBuilder):
     def equals(self, other):
         return (LookupBuilder.equals(self, other) and
                 self.rules == other.rules)
 
+    def rulesets(self):
+        # Return a list of ChainContextRuleset objects, taking explicit
+        # subtable breaks into account
+        ruleset = [ ChainContextualRuleset() ]
+        for (prefix, glyphs, suffix, lookups) in self.rules:
+            if prefix == self.SUBTABLE_BREAK_:
+                ruleset.append(ChainContextualRuleset() )
+                continue
+            ruleset[-1].addRule(prefix, glyphs, suffix, lookups)
+        # Squish any empty subtables
+        return [x for x in ruleset if len(x.rules) > 0]
+
     def build(self):
         """Build the lookup.
 
diff --git a/Tests/otlLib/builder_test.py b/Tests/otlLib/builder_test.py
index 667e0826f..27acba70d 100644
--- a/Tests/otlLib/builder_test.py
+++ b/Tests/otlLib/builder_test.py
@@ -1392,6 +1392,63 @@ def test_stat_infinities():
     assert struct.pack(">l", posInf) == b"\x7f\xff\xff\xff"
 
 
+class ChainContextualRulesetTest(object):
+    def test_makeRulesets(self):
+        font = ttLib.TTFont()
+        font.setGlyphOrder(["a","b","c","d","A","B","C","D","E"])
+        sb = builder.ChainContextSubstBuilder(font, None)
+        prefix = [ ["a"], ["b"] ]
+        input_ = [ ["c"] ]
+        suffix = [ ]
+        lookups = [ None ]
+        sb.rules.append((prefix, input_, suffix, lookups))
+
+        prefix = [ ["a"], ["d"] ]
+        sb.rules.append((prefix, input_, suffix, lookups))
+
+        sb.add_subtable_break(None)
+
+        # Second subtable has some glyph classes
+        prefix = [ ["A"] ]
+        input_ = [ ["E"] ]
+        sb.rules.append((prefix, input_, suffix, lookups))
+        input_ = [ ["C","D"] ]
+        sb.rules.append((prefix, input_, suffix, lookups))
+        prefix = [ ["A", "B"] ]
+        input_ = [ ["E"] ]
+        sb.rules.append((prefix, input_, suffix, lookups))
+
+        sb.add_subtable_break(None)
+
+        # Third subtable has no pre/post context
+        prefix = []
+        suffix = []
+        sb.rules.append((prefix, input_, suffix, lookups))
+        input_ = [ ["C","D"] ]
+        sb.rules.append((prefix, input_, suffix, lookups))
+
+        rulesets = sb.rulesets()
+        assert len(rulesets) == 3
+        assert rulesets[0].hasPrefixOrSuffix
+        assert not rulesets[0].hasAnyGlyphClasses
+        cd = rulesets[0].format1Classdefs()
+        assert set(cd[0].classes()[1:]) == set([("d",),("b",),("a",)])
+        assert set(cd[1].classes()[1:]) == set([("c",)])
+        assert set(cd[2].classes()[1:]) == set()
+
+        assert rulesets[1].hasPrefixOrSuffix
+        assert rulesets[1].hasAnyGlyphClasses
+        assert not rulesets[1].format1Classdefs()
+
+        assert not rulesets[2].hasPrefixOrSuffix
+        assert rulesets[2].hasAnyGlyphClasses
+        assert rulesets[2].format1Classdefs()
+        cd = rulesets[2].format1Classdefs()
+        assert set(cd[0].classes()[1:]) == set()
+        assert set(cd[1].classes()[1:]) == set([("C","D"), ("E",)])
+        assert set(cd[2].classes()[1:]) == set()
+
+
 if __name__ == "__main__":
     import sys