1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
|
"""fontTools.ttLib -- a package for dealing with TrueType fonts.
This package offers translators to convert TrueType fonts to Python
objects and vice versa, and additionally from Python to TTX (an XML-based
text format) and vice versa.
Example interactive session:
Python 1.5.2c1 (#43, Mar 9 1999, 13:06:43) [CW PPC w/GUSI w/MSL]
Copyright 1991-1995 Stichting Mathematisch Centrum, Amsterdam
>>> from fontTools import ttLib
>>> tt = ttLib.TTFont("afont.ttf")
>>> tt['maxp'].numGlyphs
242
>>> tt['OS/2'].achVendID
'B&H\000'
>>> tt['head'].unitsPerEm
2048
>>> tt.saveXML("afont.ttx")
Dumping 'LTSH' table...
Dumping 'OS/2' table...
Dumping 'VDMX' table...
Dumping 'cmap' table...
Dumping 'cvt ' table...
Dumping 'fpgm' table...
Dumping 'glyf' table...
Dumping 'hdmx' table...
Dumping 'head' table...
Dumping 'hhea' table...
Dumping 'hmtx' table...
Dumping 'loca' table...
Dumping 'maxp' table...
Dumping 'name' table...
Dumping 'post' table...
Dumping 'prep' table...
>>> tt2 = ttLib.TTFont()
>>> tt2.importXML("afont.ttx")
>>> tt2['maxp'].numGlyphs
242
>>>
"""
from __future__ import print_function, division, absolute_import
from fontTools.misc.py23 import *
import os
import sys
haveMacSupport = 0
if sys.platform == "mac":
haveMacSupport = 1
elif sys.platform == "darwin" and sys.version_info[:3] != (2, 2, 0):
# Python 2.2's Mac support is broken, so don't enable it there.
haveMacSupport = 1
class TTLibError(Exception): pass
class TTFont(object):
"""The main font object. It manages file input and output, and offers
a convenient way of accessing tables.
Tables will be only decompiled when necessary, ie. when they're actually
accessed. This means that simple operations can be extremely fast.
"""
def __init__(self, file=None, res_name_or_index=None,
sfntVersion="\000\001\000\000", flavor=None, checkChecksums=False,
verbose=False, recalcBBoxes=True, allowVID=False, ignoreDecompileErrors=False,
recalcTimestamp=True, fontNumber=-1, lazy=False, quiet=False):
"""The constructor can be called with a few different arguments.
When reading a font from disk, 'file' should be either a pathname
pointing to a file, or a readable file object.
It we're running on a Macintosh, 'res_name_or_index' maybe an sfnt
resource name or an sfnt resource index number or zero. The latter
case will cause TTLib to autodetect whether the file is a flat file
or a suitcase. (If it's a suitcase, only the first 'sfnt' resource
will be read!)
The 'checkChecksums' argument is used to specify how sfnt
checksums are treated upon reading a file from disk:
0: don't check (default)
1: check, print warnings if a wrong checksum is found
2: check, raise an exception if a wrong checksum is found.
The TTFont constructor can also be called without a 'file'
argument: this is the way to create a new empty font.
In this case you can optionally supply the 'sfntVersion' argument,
and a 'flavor' which can be None, or 'woff'.
If the recalcBBoxes argument is false, a number of things will *not*
be recalculated upon save/compile:
1) glyph bounding boxes
2) maxp font bounding box
3) hhea min/max values
(1) is needed for certain kinds of CJK fonts (ask Werner Lemberg ;-).
Additionally, upon importing an TTX file, this option cause glyphs
to be compiled right away. This should reduce memory consumption
greatly, and therefore should have some impact on the time needed
to parse/compile large fonts.
If the recalcTimestamp argument is false, the modified timestamp in the
'head' table will *not* be recalculated upon save/compile.
If the allowVID argument is set to true, then virtual GID's are
supported. Asking for a glyph ID with a glyph name or GID that is not in
the font will return a virtual GID. This is valid for GSUB and cmap
tables. For SING glyphlets, the cmap table is used to specify Unicode
values for virtual GI's used in GSUB/GPOS rules. If the gid N is requested
and does not exist in the font, or the glyphname has the form glyphN
and does not exist in the font, then N is used as the virtual GID.
Else, the first virtual GID is assigned as 0x1000 -1; for subsequent new
virtual GIDs, the next is one less than the previous.
If ignoreDecompileErrors is set to True, exceptions raised in
individual tables during decompilation will be ignored, falling
back to the DefaultTable implementation, which simply keeps the
binary data.
If lazy is set to True, many data structures are loaded lazily, upon
access only.
"""
from fontTools.ttLib import sfnt
self.verbose = verbose
self.quiet = quiet
self.lazy = lazy
self.recalcBBoxes = recalcBBoxes
self.recalcTimestamp = recalcTimestamp
self.tables = {}
self.reader = None
# Permit the user to reference glyphs that are not int the font.
self.last_vid = 0xFFFE # Can't make it be 0xFFFF, as the world is full unsigned short integer counters that get incremented after the last seen GID value.
self.reverseVIDDict = {}
self.VIDDict = {}
self.allowVID = allowVID
self.ignoreDecompileErrors = ignoreDecompileErrors
if not file:
self.sfntVersion = sfntVersion
self.flavor = flavor
self.flavorData = None
return
if not hasattr(file, "read"):
# assume file is a string
if haveMacSupport and res_name_or_index is not None:
# on the mac, we deal with sfnt resources as well as flat files
from . import macUtils
if res_name_or_index == 0:
if macUtils.getSFNTResIndices(file):
# get the first available sfnt font.
file = macUtils.SFNTResourceReader(file, 1)
else:
file = open(file, "rb")
else:
file = macUtils.SFNTResourceReader(file, res_name_or_index)
else:
file = open(file, "rb")
else:
pass # assume "file" is a readable file object
self.reader = sfnt.SFNTReader(file, checkChecksums, fontNumber=fontNumber)
self.sfntVersion = self.reader.sfntVersion
self.flavor = self.reader.flavor
self.flavorData = self.reader.flavorData
def close(self):
"""If we still have a reader object, close it."""
if self.reader is not None:
self.reader.close()
def save(self, file, makeSuitcase=False, reorderTables=True):
"""Save the font to disk. Similarly to the constructor,
the 'file' argument can be either a pathname or a writable
file object.
On the Mac, if makeSuitcase is true, a suitcase (resource fork)
file will we made instead of a flat .ttf file.
"""
from fontTools.ttLib import sfnt
if not hasattr(file, "write"):
closeStream = 1
if os.name == "mac" and makeSuitcase:
from . import macUtils
file = macUtils.SFNTResourceWriter(file, self)
else:
file = open(file, "wb")
if os.name == "mac":
from fontTools.misc.macCreator import setMacCreatorAndType
setMacCreatorAndType(file.name, 'mdos', 'BINA')
else:
# assume "file" is a writable file object
closeStream = 0
tags = list(self.keys())
if "GlyphOrder" in tags:
tags.remove("GlyphOrder")
numTables = len(tags)
if reorderTables:
import tempfile
tmp = tempfile.TemporaryFile(prefix="ttx-fonttools")
else:
tmp = file
writer = sfnt.SFNTWriter(tmp, numTables, self.sfntVersion, self.flavor, self.flavorData)
done = []
for tag in tags:
self._writeTable(tag, writer, done)
writer.close()
if reorderTables:
tmp.flush()
tmp.seek(0)
reorderFontTables(tmp, file)
tmp.close()
if closeStream:
file.close()
def saveXML(self, fileOrPath, progress=None, quiet=False,
tables=None, skipTables=None, splitTables=False, disassembleInstructions=True,
bitmapGlyphDataFormat='raw'):
"""Export the font as TTX (an XML-based text file), or as a series of text
files when splitTables is true. In the latter case, the 'fileOrPath'
argument should be a path to a directory.
The 'tables' argument must either be false (dump all tables) or a
list of tables to dump. The 'skipTables' argument may be a list of tables
to skip, but only when the 'tables' argument is false.
"""
from fontTools import version
from fontTools.misc import xmlWriter
self.disassembleInstructions = disassembleInstructions
self.bitmapGlyphDataFormat = bitmapGlyphDataFormat
if not tables:
tables = list(self.keys())
if "GlyphOrder" not in tables:
tables = ["GlyphOrder"] + tables
if skipTables:
for tag in skipTables:
if tag in tables:
tables.remove(tag)
numTables = len(tables)
if progress:
progress.set(0, numTables)
idlefunc = getattr(progress, "idle", None)
else:
idlefunc = None
writer = xmlWriter.XMLWriter(fileOrPath, idlefunc=idlefunc)
writer.begintag("ttFont", sfntVersion=repr(self.sfntVersion)[1:-1],
ttLibVersion=version)
writer.newline()
if not splitTables:
writer.newline()
else:
# 'fileOrPath' must now be a path
path, ext = os.path.splitext(fileOrPath)
fileNameTemplate = path + ".%s" + ext
for i in range(numTables):
if progress:
progress.set(i)
tag = tables[i]
if splitTables:
tablePath = fileNameTemplate % tagToIdentifier(tag)
tableWriter = xmlWriter.XMLWriter(tablePath, idlefunc=idlefunc)
tableWriter.begintag("ttFont", ttLibVersion=version)
tableWriter.newline()
tableWriter.newline()
writer.simpletag(tagToXML(tag), src=os.path.basename(tablePath))
writer.newline()
else:
tableWriter = writer
self._tableToXML(tableWriter, tag, progress, quiet)
if splitTables:
tableWriter.endtag("ttFont")
tableWriter.newline()
tableWriter.close()
if progress:
progress.set((i + 1))
writer.endtag("ttFont")
writer.newline()
writer.close()
if self.verbose:
debugmsg("Done dumping TTX")
def _tableToXML(self, writer, tag, progress, quiet):
if tag in self:
table = self[tag]
report = "Dumping '%s' table..." % tag
else:
report = "No '%s' table found." % tag
if progress:
progress.setLabel(report)
elif self.verbose:
debugmsg(report)
else:
if not quiet:
print(report)
if tag not in self:
return
xmlTag = tagToXML(tag)
if hasattr(table, "ERROR"):
writer.begintag(xmlTag, ERROR="decompilation error")
else:
writer.begintag(xmlTag)
writer.newline()
if tag in ("glyf", "CFF "):
table.toXML(writer, self, progress)
else:
table.toXML(writer, self)
writer.endtag(xmlTag)
writer.newline()
writer.newline()
def importXML(self, file, progress=None, quiet=False):
"""Import a TTX file (an XML-based text format), so as to recreate
a font object.
"""
if "maxp" in self and "post" in self:
# Make sure the glyph order is loaded, as it otherwise gets
# lost if the XML doesn't contain the glyph order, yet does
# contain the table which was originally used to extract the
# glyph names from (ie. 'post', 'cmap' or 'CFF ').
self.getGlyphOrder()
from fontTools.misc import xmlReader
reader = xmlReader.XMLReader(file, self, progress, quiet)
reader.read()
def isLoaded(self, tag):
"""Return true if the table identified by 'tag' has been
decompiled and loaded into memory."""
return tag in self.tables
def has_key(self, tag):
if self.isLoaded(tag):
return True
elif self.reader and tag in self.reader:
return True
elif tag == "GlyphOrder":
return True
else:
return False
__contains__ = has_key
def keys(self):
keys = list(self.tables.keys())
if self.reader:
for key in list(self.reader.keys()):
if key not in keys:
keys.append(key)
if "GlyphOrder" in keys:
keys.remove("GlyphOrder")
keys = sortedTagList(keys)
return ["GlyphOrder"] + keys
def __len__(self):
return len(list(self.keys()))
def __getitem__(self, tag):
tag = Tag(tag)
try:
return self.tables[tag]
except KeyError:
if tag == "GlyphOrder":
table = GlyphOrder(tag)
self.tables[tag] = table
return table
if self.reader is not None:
import traceback
if self.verbose:
debugmsg("Reading '%s' table from disk" % tag)
data = self.reader[tag]
tableClass = getTableClass(tag)
table = tableClass(tag)
self.tables[tag] = table
if self.verbose:
debugmsg("Decompiling '%s' table" % tag)
try:
table.decompile(data, self)
except:
if not self.ignoreDecompileErrors:
raise
# fall back to DefaultTable, retaining the binary table data
print("An exception occurred during the decompilation of the '%s' table" % tag)
from .tables.DefaultTable import DefaultTable
file = StringIO()
traceback.print_exc(file=file)
table = DefaultTable(tag)
table.ERROR = file.getvalue()
self.tables[tag] = table
table.decompile(data, self)
return table
else:
raise KeyError("'%s' table not found" % tag)
def __setitem__(self, tag, table):
self.tables[Tag(tag)] = table
def __delitem__(self, tag):
if tag not in self:
raise KeyError("'%s' table not found" % tag)
if tag in self.tables:
del self.tables[tag]
if self.reader and tag in self.reader:
del self.reader[tag]
def get(self, tag, default=None):
try:
return self[tag]
except KeyError:
return default
def setGlyphOrder(self, glyphOrder):
self.glyphOrder = glyphOrder
def getGlyphOrder(self):
try:
return self.glyphOrder
except AttributeError:
pass
if 'CFF ' in self:
cff = self['CFF ']
self.glyphOrder = cff.getGlyphOrder()
elif 'post' in self:
# TrueType font
glyphOrder = self['post'].getGlyphOrder()
if glyphOrder is None:
#
# No names found in the 'post' table.
# Try to create glyph names from the unicode cmap (if available)
# in combination with the Adobe Glyph List (AGL).
#
self._getGlyphNamesFromCmap()
else:
self.glyphOrder = glyphOrder
else:
self._getGlyphNamesFromCmap()
return self.glyphOrder
def _getGlyphNamesFromCmap(self):
#
# This is rather convoluted, but then again, it's an interesting problem:
# - we need to use the unicode values found in the cmap table to
# build glyph names (eg. because there is only a minimal post table,
# or none at all).
# - but the cmap parser also needs glyph names to work with...
# So here's what we do:
# - make up glyph names based on glyphID
# - load a temporary cmap table based on those names
# - extract the unicode values, build the "real" glyph names
# - unload the temporary cmap table
#
if self.isLoaded("cmap"):
# Bootstrapping: we're getting called by the cmap parser
# itself. This means self.tables['cmap'] contains a partially
# loaded cmap, making it impossible to get at a unicode
# subtable here. We remove the partially loaded cmap and
# restore it later.
# This only happens if the cmap table is loaded before any
# other table that does f.getGlyphOrder() or f.getGlyphName().
cmapLoading = self.tables['cmap']
del self.tables['cmap']
else:
cmapLoading = None
# Make up glyph names based on glyphID, which will be used by the
# temporary cmap and by the real cmap in case we don't find a unicode
# cmap.
numGlyphs = int(self['maxp'].numGlyphs)
glyphOrder = [None] * numGlyphs
glyphOrder[0] = ".notdef"
for i in range(1, numGlyphs):
glyphOrder[i] = "glyph%.5d" % i
# Set the glyph order, so the cmap parser has something
# to work with (so we don't get called recursively).
self.glyphOrder = glyphOrder
# Get a (new) temporary cmap (based on the just invented names)
tempcmap = self['cmap'].getcmap(3, 1)
if tempcmap is not None:
# we have a unicode cmap
from fontTools import agl
cmap = tempcmap.cmap
# create a reverse cmap dict
reversecmap = {}
for unicode, name in list(cmap.items()):
reversecmap[name] = unicode
allNames = {}
for i in range(numGlyphs):
tempName = glyphOrder[i]
if tempName in reversecmap:
unicode = reversecmap[tempName]
if unicode in agl.UV2AGL:
# get name from the Adobe Glyph List
glyphName = agl.UV2AGL[unicode]
else:
# create uni<CODE> name
glyphName = "uni%04X" % unicode
tempName = glyphName
n = 1
while tempName in allNames:
tempName = glyphName + "#" + repr(n)
n = n + 1
glyphOrder[i] = tempName
allNames[tempName] = 1
# Delete the temporary cmap table from the cache, so it can
# be parsed again with the right names.
del self.tables['cmap']
else:
pass # no unicode cmap available, stick with the invented names
self.glyphOrder = glyphOrder
if cmapLoading:
# restore partially loaded cmap, so it can continue loading
# using the proper names.
self.tables['cmap'] = cmapLoading
def getGlyphNames(self):
"""Get a list of glyph names, sorted alphabetically."""
glyphNames = sorted(self.getGlyphOrder()[:])
return glyphNames
def getGlyphNames2(self):
"""Get a list of glyph names, sorted alphabetically,
but not case sensitive.
"""
from fontTools.misc import textTools
return textTools.caselessSort(self.getGlyphOrder())
def getGlyphName(self, glyphID, requireReal=False):
try:
return self.getGlyphOrder()[glyphID]
except IndexError:
if requireReal or not self.allowVID:
# XXX The ??.W8.otf font that ships with OSX uses higher glyphIDs in
# the cmap table than there are glyphs. I don't think it's legal...
return "glyph%.5d" % glyphID
else:
# user intends virtual GID support
try:
glyphName = self.VIDDict[glyphID]
except KeyError:
glyphName ="glyph%.5d" % glyphID
self.last_vid = min(glyphID, self.last_vid )
self.reverseVIDDict[glyphName] = glyphID
self.VIDDict[glyphID] = glyphName
return glyphName
def getGlyphID(self, glyphName, requireReal=False):
if not hasattr(self, "_reverseGlyphOrderDict"):
self._buildReverseGlyphOrderDict()
glyphOrder = self.getGlyphOrder()
d = self._reverseGlyphOrderDict
if glyphName not in d:
if glyphName in glyphOrder:
self._buildReverseGlyphOrderDict()
return self.getGlyphID(glyphName)
else:
if requireReal:
raise KeyError(glyphName)
elif not self.allowVID:
# Handle glyphXXX only
if glyphName[:5] == "glyph":
try:
return int(glyphName[5:])
except (NameError, ValueError):
raise KeyError(glyphName)
else:
# user intends virtual GID support
try:
glyphID = self.reverseVIDDict[glyphName]
except KeyError:
# if name is in glyphXXX format, use the specified name.
if glyphName[:5] == "glyph":
try:
glyphID = int(glyphName[5:])
except (NameError, ValueError):
glyphID = None
if glyphID is None:
glyphID = self.last_vid -1
self.last_vid = glyphID
self.reverseVIDDict[glyphName] = glyphID
self.VIDDict[glyphID] = glyphName
return glyphID
glyphID = d[glyphName]
if glyphName != glyphOrder[glyphID]:
self._buildReverseGlyphOrderDict()
return self.getGlyphID(glyphName)
return glyphID
def getReverseGlyphMap(self, rebuild=False):
if rebuild or not hasattr(self, "_reverseGlyphOrderDict"):
self._buildReverseGlyphOrderDict()
return self._reverseGlyphOrderDict
def _buildReverseGlyphOrderDict(self):
self._reverseGlyphOrderDict = d = {}
glyphOrder = self.getGlyphOrder()
for glyphID in range(len(glyphOrder)):
d[glyphOrder[glyphID]] = glyphID
def _writeTable(self, tag, writer, done):
"""Internal helper function for self.save(). Keeps track of
inter-table dependencies.
"""
if tag in done:
return
tableClass = getTableClass(tag)
for masterTable in tableClass.dependencies:
if masterTable not in done:
if masterTable in self:
self._writeTable(masterTable, writer, done)
else:
done.append(masterTable)
tabledata = self.getTableData(tag)
if self.verbose:
debugmsg("writing '%s' table to disk" % tag)
writer[tag] = tabledata
done.append(tag)
def getTableData(self, tag):
"""Returns raw table data, whether compiled or directly read from disk.
"""
tag = Tag(tag)
if self.isLoaded(tag):
if self.verbose:
debugmsg("compiling '%s' table" % tag)
return self.tables[tag].compile(self)
elif self.reader and tag in self.reader:
if self.verbose:
debugmsg("Reading '%s' table from disk" % tag)
return self.reader[tag]
else:
raise KeyError(tag)
def getGlyphSet(self, preferCFF=True):
"""Return a generic GlyphSet, which is a dict-like object
mapping glyph names to glyph objects. The returned glyph objects
have a .draw() method that supports the Pen protocol, and will
have an attribute named 'width', but only *after* the .draw() method
has been called.
If the font is CFF-based, the outlines will be taken from the 'CFF '
table. Otherwise the outlines will be taken from the 'glyf' table.
If the font contains both a 'CFF ' and a 'glyf' table, you can use
the 'preferCFF' argument to specify which one should be taken.
"""
if preferCFF and "CFF " in self:
return list(self["CFF "].cff.values())[0].CharStrings
if "glyf" in self:
return _TTGlyphSet(self)
if "CFF " in self:
return list(self["CFF "].cff.values())[0].CharStrings
raise TTLibError("Font contains no outlines")
class _TTGlyphSet(object):
"""Generic dict-like GlyphSet class, meant as a TrueType counterpart
to CFF's CharString dict. See TTFont.getGlyphSet().
"""
# This class is distinct from the 'glyf' table itself because we need
# access to the 'hmtx' table, which could cause a dependency problem
# there when reading from XML.
def __init__(self, ttFont):
self._ttFont = ttFont
def keys(self):
return list(self._ttFont["glyf"].keys())
def has_key(self, glyphName):
return glyphName in self._ttFont["glyf"]
__contains__ = has_key
def __getitem__(self, glyphName):
return _TTGlyph(glyphName, self._ttFont)
def get(self, glyphName, default=None):
try:
return self[glyphName]
except KeyError:
return default
class _TTGlyph(object):
"""Wrapper for a TrueType glyph that supports the Pen protocol, meaning
that it has a .draw() method that takes a pen object as its only
argument. Additionally there is a 'width' attribute.
"""
def __init__(self, glyphName, ttFont):
self._glyphName = glyphName
self._ttFont = ttFont
self.width, self.lsb = self._ttFont['hmtx'][self._glyphName]
def draw(self, pen):
"""Draw the glyph onto Pen. See fontTools.pens.basePen for details
how that works.
"""
glyfTable = self._ttFont['glyf']
glyph = glyfTable[self._glyphName]
if hasattr(glyph, "xMin"):
offset = self.lsb - glyph.xMin
else:
offset = 0
if glyph.isComposite():
for component in glyph:
glyphName, transform = component.getComponentInfo()
pen.addComponent(glyphName, transform)
else:
coordinates, endPts, flags = glyph.getCoordinates(glyfTable)
if offset:
coordinates = coordinates + (offset, 0)
start = 0
for end in endPts:
end = end + 1
contour = coordinates[start:end].tolist()
cFlags = flags[start:end].tolist()
start = end
if 1 not in cFlags:
# There is not a single on-curve point on the curve,
# use pen.qCurveTo's special case by specifying None
# as the on-curve point.
contour.append(None)
pen.qCurveTo(*contour)
else:
# Shuffle the points so that contour the is guaranteed
# to *end* in an on-curve point, which we'll use for
# the moveTo.
firstOnCurve = cFlags.index(1) + 1
contour = contour[firstOnCurve:] + contour[:firstOnCurve]
cFlags = cFlags[firstOnCurve:] + cFlags[:firstOnCurve]
pen.moveTo(contour[-1])
while contour:
nextOnCurve = cFlags.index(1) + 1
if nextOnCurve == 1:
pen.lineTo(contour[0])
else:
pen.qCurveTo(*contour[:nextOnCurve])
contour = contour[nextOnCurve:]
cFlags = cFlags[nextOnCurve:]
pen.closePath()
class GlyphOrder(object):
"""A pseudo table. The glyph order isn't in the font as a separate
table, but it's nice to present it as such in the TTX format.
"""
def __init__(self, tag=None):
pass
def toXML(self, writer, ttFont):
glyphOrder = ttFont.getGlyphOrder()
writer.comment("The 'id' attribute is only for humans; "
"it is ignored when parsed.")
writer.newline()
for i in range(len(glyphOrder)):
glyphName = glyphOrder[i]
writer.simpletag("GlyphID", id=i, name=glyphName)
writer.newline()
def fromXML(self, name, attrs, content, ttFont):
if not hasattr(self, "glyphOrder"):
self.glyphOrder = []
ttFont.setGlyphOrder(self.glyphOrder)
if name == "GlyphID":
self.glyphOrder.append(attrs["name"])
def getTableModule(tag):
"""Fetch the packer/unpacker module for a table.
Return None when no module is found.
"""
from . import tables
pyTag = tagToIdentifier(tag)
try:
__import__("fontTools.ttLib.tables." + pyTag)
except ImportError as err:
# If pyTag is found in the ImportError message,
# means table is not implemented. If it's not
# there, then some other module is missing, don't
# suppress the error.
if str(err).find(pyTag) >= 0:
return None
else:
raise err
else:
return getattr(tables, pyTag)
def getTableClass(tag):
"""Fetch the packer/unpacker class for a table.
Return None when no class is found.
"""
module = getTableModule(tag)
if module is None:
from .tables.DefaultTable import DefaultTable
return DefaultTable
pyTag = tagToIdentifier(tag)
tableClass = getattr(module, "table_" + pyTag)
return tableClass
def getClassTag(klass):
"""Fetch the table tag for a class object."""
name = klass.__name__
assert name[:6] == 'table_'
name = name[6:] # Chop 'table_'
return identifierToTag(name)
def newTable(tag):
"""Return a new instance of a table."""
tableClass = getTableClass(tag)
return tableClass(tag)
def _escapechar(c):
"""Helper function for tagToIdentifier()"""
import re
if re.match("[a-z0-9]", c):
return "_" + c
elif re.match("[A-Z]", c):
return c + "_"
else:
return hex(byteord(c))[2:]
def tagToIdentifier(tag):
"""Convert a table tag to a valid (but UGLY) python identifier,
as well as a filename that's guaranteed to be unique even on a
caseless file system. Each character is mapped to two characters.
Lowercase letters get an underscore before the letter, uppercase
letters get an underscore after the letter. Trailing spaces are
trimmed. Illegal characters are escaped as two hex bytes. If the
result starts with a number (as the result of a hex escape), an
extra underscore is prepended. Examples:
'glyf' -> '_g_l_y_f'
'cvt ' -> '_c_v_t'
'OS/2' -> 'O_S_2f_2'
"""
import re
tag = Tag(tag)
if tag == "GlyphOrder":
return tag
assert len(tag) == 4, "tag should be 4 characters long"
while len(tag) > 1 and tag[-1] == ' ':
tag = tag[:-1]
ident = ""
for c in tag:
ident = ident + _escapechar(c)
if re.match("[0-9]", ident):
ident = "_" + ident
return ident
def identifierToTag(ident):
"""the opposite of tagToIdentifier()"""
if ident == "GlyphOrder":
return ident
if len(ident) % 2 and ident[0] == "_":
ident = ident[1:]
assert not (len(ident) % 2)
tag = ""
for i in range(0, len(ident), 2):
if ident[i] == "_":
tag = tag + ident[i+1]
elif ident[i+1] == "_":
tag = tag + ident[i]
else:
# assume hex
tag = tag + chr(int(ident[i:i+2], 16))
# append trailing spaces
tag = tag + (4 - len(tag)) * ' '
return Tag(tag)
def tagToXML(tag):
"""Similarly to tagToIdentifier(), this converts a TT tag
to a valid XML element name. Since XML element names are
case sensitive, this is a fairly simple/readable translation.
"""
import re
tag = Tag(tag)
if tag == "OS/2":
return "OS_2"
elif tag == "GlyphOrder":
return tag
if re.match("[A-Za-z_][A-Za-z_0-9]* *$", tag):
return tag.strip()
else:
return tagToIdentifier(tag)
def xmlToTag(tag):
"""The opposite of tagToXML()"""
if tag == "OS_2":
return Tag("OS/2")
if len(tag) == 8:
return identifierToTag(tag)
else:
return Tag(tag + " " * (4 - len(tag)))
def debugmsg(msg):
import time
print(msg + time.strftime(" (%H:%M:%S)", time.localtime(time.time())))
# Table order as recommended in the OpenType specification 1.4
TTFTableOrder = ["head", "hhea", "maxp", "OS/2", "hmtx", "LTSH", "VDMX",
"hdmx", "cmap", "fpgm", "prep", "cvt ", "loca", "glyf",
"kern", "name", "post", "gasp", "PCLT"]
OTFTableOrder = ["head", "hhea", "maxp", "OS/2", "name", "cmap", "post",
"CFF "]
def sortedTagList(tagList, tableOrder=None):
"""Return a sorted copy of tagList, sorted according to the OpenType
specification, or according to a custom tableOrder. If given and not
None, tableOrder needs to be a list of tag names.
"""
tagList = sorted(tagList)
if tableOrder is None:
if "DSIG" in tagList:
# DSIG should be last (XXX spec reference?)
tagList.remove("DSIG")
tagList.append("DSIG")
if "CFF " in tagList:
tableOrder = OTFTableOrder
else:
tableOrder = TTFTableOrder
orderedTables = []
for tag in tableOrder:
if tag in tagList:
orderedTables.append(tag)
tagList.remove(tag)
orderedTables.extend(tagList)
return orderedTables
def reorderFontTables(inFile, outFile, tableOrder=None, checkChecksums=False):
"""Rewrite a font file, ordering the tables as recommended by the
OpenType specification 1.4.
"""
from fontTools.ttLib.sfnt import SFNTReader, SFNTWriter
reader = SFNTReader(inFile, checkChecksums=checkChecksums)
writer = SFNTWriter(outFile, len(reader.tables), reader.sfntVersion, reader.flavor, reader.flavorData)
tables = list(reader.keys())
for tag in sortedTagList(tables, tableOrder):
writer[tag] = reader[tag]
writer.close()
|