Include a ligatures sample, and add TR TH and TI ligatures from the original handwriting sample

xkcd-script/generator/pt2_character_classification.py

@@ -102,6 +102,18 @@ def merge(img1, img1_bbox, img2, img2_bbox):
 
 
 
+_LIGATURE_SOURCES = [
+    (5, 4, 5, 'TH'),   # UNAUTHORITATIVENESS: T + H
+    (5, 31, 32, 'TR'),  # INTRACOLIC: T + R
+    (6, 24, 25, 'TI'),  # ALBERTITE: T + I
+]
+for line_no, left_no, right_no, lig in _LIGATURE_SOURCES:
+    _, left_bbox, left_img = characters_by_line[line_no][left_no]
+    _, right_bbox, right_img = characters_by_line[line_no][right_no]
+    lig_img, lig_bbox = merge(left_img, left_bbox, right_img, right_bbox)
+    characters_by_line[line_no].append([lig, lig_bbox, lig_img])
+
+
 import skimage.io
 
 if not os.path.isdir('../generated/characters'):

xkcd-script/samples/gen_ligatures.py

@@ -0,0 +1,62 @@
+#!/usr/bin/env python3
+"""Validate expected ligatures against the built OTF and emit preview text."""
+import os
+import sys
+
+from fontTools.ttLib import TTFont
+
+HERE = os.path.dirname(os.path.abspath(__file__))
+OTF = os.path.join(HERE, '../font/xkcd-script.otf')
+
+# (display_text, input_glyph_sequence)
+EXPECTED = [
+    ('co', ('c', 'o')),
+    ('oo', ('o', 'o')),
+    ('EA', ('E', 'A')),
+    ('CO', ('C', 'O')),
+    ('CR', ('C', 'R')),
+    ('OCH', ('O', 'C', 'H')),
+    ('TH', ('T', 'H')),
+    ('TR', ('T', 'R')),
+    ('TI', ('T', 'I')),
+    ('EE', ('E', 'E')),
+    ('TT', ('T', 'T')),
+    ('LB', ('L', 'B')),
+    ('ER', ('E', 'R')),
+    ('RR', ('R', 'R')),
+    ('LA', ('L', 'A')),
+    ('LL', ('L', 'L')),
+    ('ON', ('O', 'N')),
+    ('CA', ('C', 'A')),
+    ('PS', ('P', 'S')),
+    ('TO', ('T', 'O')),
+    ('|>', ('bar', 'greater')),
+    ('<|', ('less', 'bar')),
+    ('I-pronoun', ('I', 'hyphen', 'p', 'r', 'o', 'n', 'o', 'u', 'n')),
+]
+
+tt = TTFont(OTF)
+gsub = tt['GSUB'].table
+
+ligatures = set()
+for record in gsub.FeatureList.FeatureRecord:
+    if record.FeatureTag != 'liga':
+        continue
+    for idx in record.Feature.LookupListIndex:
+        lookup = gsub.LookupList.Lookup[idx]
+        if lookup.LookupType != 4:
+            continue
+        for sub in lookup.SubTable:
+            for first, ligset in sub.ligatures.items():
+                for lig in ligset:
+                    ligatures.add((first,) + tuple(lig.Component))
+
+expected_seqs = {seq for _, seq in EXPECTED}
+uncovered = sorted(str(seq) for seq in ligatures if seq not in expected_seqs)
+if uncovered:
+    sys.exit('ERROR: font ligatures not in EXPECTED — add to gen_ligatures.py:\n' + '\n'.join(uncovered))
+
+cols = 6
+tokens = [display for display, _ in EXPECTED]
+rows = [tokens[i:i + cols] for i in range(0, len(tokens), cols)]
+print('Ligatures:\n' + '\n'.join('  ' + '  '.join(row) for row in rows))

xkcd-script/samples/preview.sh

@@ -41,6 +41,13 @@ ${RUN_CTXT} convert -strip ${NAME}.pgm ${NAME}.png
 ${RUN_CTXT} rm ${NAME}.pgm
 
 
+NAME=ligatures
+CONTENT=$(${RUN_CTXT} python3 gen_ligatures.py) || exit 1
+${SIMPLE_CTXT} -o ${NAME}.pgm --text "${CONTENT}"
+${RUN_CTXT} convert -strip ${NAME}.pgm ${NAME}.png
+${RUN_CTXT} rm ${NAME}.pgm
+
+
 if [ "$?" == "141" ] ; then
     # Unexplained exit code from the handwriting sample.
     exit 0