• file
• revisions
• annotate
• diff
• comparison
• raw

gfx/harfbuzz/src/gen-indic-table.py@6474c204b198 (annotated)

gfx/harfbuzz/src/gen-indic-table.py

Wed, 31 Dec 2014 06:09:35 +0100

author

Michael Schloh von Bennewitz <michael@schloh.com>

date

Wed, 31 Dec 2014 06:09:35 +0100

changeset 0

6474c204b198

permissions

-rwxr-xr-x

Cloned upstream origin tor-browser at tor-browser-31.3.0esr-4.5-1-build1
revision ID fc1c9ff7c1b2defdbc039f12214767608f46423f for hacking purpose.

michael@0	1	#!/usr/bin/python
michael@0	2
michael@0	3	import sys
michael@0	4
michael@0	5	if len (sys.argv) != 4:
michael@0	6	print >>sys.stderr, "usage: ./gen-indic-table.py IndicSyllabicCategory.txt IndicMatraCategory.txt Blocks.txt"
michael@0	7	sys.exit (1)
michael@0	8
michael@0	9	files = [file (x) for x in sys.argv[1:]]
michael@0	10
michael@0	11	headers = [[f.readline () for i in range (2)] for f in files]
michael@0	12
michael@0	13	blocks = {}
michael@0	14	data = [{} for f in files]
michael@0	15	values = [{} for f in files]
michael@0	16	for i, f in enumerate (files):
michael@0	17	for line in f:
michael@0	18
michael@0	19	j = line.find ('#')
michael@0	20	if j >= 0:
michael@0	21	line = line[:j]
michael@0	22
michael@0	23	fields = [x.strip () for x in line.split (';')]
michael@0	24	if len (fields) == 1:
michael@0	25	continue
michael@0	26
michael@0	27	uu = fields[0].split ('..')
michael@0	28	start = int (uu[0], 16)
michael@0	29	if len (uu) == 1:
michael@0	30	end = start
michael@0	31	else:
michael@0	32	end = int (uu[1], 16)
michael@0	33
michael@0	34	t = fields[1]
michael@0	35
michael@0	36	for u in range (start, end + 1):
michael@0	37	data[i][u] = t
michael@0	38	values[i][t] = values[i].get (t, 0) + 1
michael@0	39
michael@0	40	if i == 2:
michael@0	41	blocks[t] = (start, end)
michael@0	42
michael@0	43	# Merge data into one dict:
michael@0	44	defaults = ('Other', 'Not_Applicable', 'No_Block')
michael@0	45	for i,v in enumerate (defaults):
michael@0	46	values[i][v] = values[i].get (v, 0) + 1
michael@0	47	combined = {}
michael@0	48	for i,d in enumerate (data):
michael@0	49	for u,v in d.items ():
michael@0	50	if i == 2 and not u in combined:
michael@0	51	continue
michael@0	52	if not u in combined:
michael@0	53	combined[u] = list (defaults)
michael@0	54	combined[u][i] = v
michael@0	55	data = combined
michael@0	56	del combined
michael@0	57	num = len (data)
michael@0	58
michael@0	59	# Move the outliers NO-BREAK SPACE and DOTTED CIRCLE out
michael@0	60	singles = {}
michael@0	61	for u in [0x00A0, 0x25CC]:
michael@0	62	singles[u] = data[u]
michael@0	63	del data[u]
michael@0	64
michael@0	65	print "/* == Start of generated table == */"
michael@0	66	print "/*"
michael@0	67	print " * The following table is generated by running:"
michael@0	68	print " *"
michael@0	69	print " * ./gen-indic-table.py IndicSyllabicCategory.txt IndicMatraCategory.txt Blocks.txt"
michael@0	70	print " *"
michael@0	71	print " * on files with these headers:"
michael@0	72	print " *"
michael@0	73	for h in headers:
michael@0	74	for l in h:
michael@0	75	print " * %s" % (l.strip())
michael@0	76	print " */"
michael@0	77	print
michael@0	78	print '#include "hb-ot-shape-complex-indic-private.hh"'
michael@0	79	print
michael@0	80
michael@0	81	# Shorten values
michael@0	82	short = [{
michael@0	83	"Bindu": 'Bi',
michael@0	84	"Visarga": 'Vs',
michael@0	85	"Vowel": 'Vo',
michael@0	86	"Vowel_Dependent": 'M',
michael@0	87	"Other": 'x',
michael@0	88	},{
michael@0	89	"Not_Applicable": 'x',
michael@0	90	}]
michael@0	91	all_shorts = [[],[]]
michael@0	92
michael@0	93	# Add some of the values, to make them more readable, and to avoid duplicates
michael@0	94
michael@0	95
michael@0	96	for i in range (2):
michael@0	97	for v,s in short[i].items ():
michael@0	98	all_shorts[i].append (s)
michael@0	99
michael@0	100	what = ["INDIC_SYLLABIC_CATEGORY", "INDIC_MATRA_CATEGORY"]
michael@0	101	what_short = ["ISC", "IMC"]
michael@0	102	for i in range (2):
michael@0	103	print
michael@0	104	vv = values[i].keys ()
michael@0	105	vv.sort ()
michael@0	106	for v in vv:
michael@0	107	v_no_and = v.replace ('_And_', '_')
michael@0	108	if v in short[i]:
michael@0	109	s = short[i][v]
michael@0	110	else:
michael@0	111	s = ''.join ([c for c in v_no_and if ord ('A') <= ord (c) <= ord ('Z')])
michael@0	112	if s in all_shorts[i]:
michael@0	113	raise Exception ("Duplicate short value alias", v, s)
michael@0	114	all_shorts[i].append (s)
michael@0	115	short[i][v] = s
michael@0	116	print "#define %s_%s %s_%s %s/* %3d chars; %s */" % \
michael@0	117	(what_short[i], s, what[i], v.upper (), \
michael@0	118	' '* ((48-1 - len (what[i]) - 1 - len (v)) / 8), \
michael@0	119	values[i][v], v)
michael@0	120	print
michael@0	121	print "#define _(S,M) INDIC_COMBINE_CATEGORIES (ISC_##S, IMC_##M)"
michael@0	122	print
michael@0	123	print
michael@0	124
michael@0	125	total = 0
michael@0	126	used = 0
michael@0	127	def print_block (block, start, end, data):
michael@0	128	print
michael@0	129	print
michael@0	130	print " /* %s (%04X..%04X) */" % (block, start, end)
michael@0	131	num = 0
michael@0	132	for u in range (start, end+1):
michael@0	133	if u % 8 == 0:
michael@0	134	print
michael@0	135	print " /* %04X */" % u,
michael@0	136	if u in data:
michael@0	137	num += 1
michael@0	138	d = data.get (u, defaults)
michael@0	139	sys.stdout.write ("%9s" % ("_(%s,%s)," % (short[0][d[0]], short[1][d[1]])))
michael@0	140
michael@0	141	global total, used
michael@0	142	total += end - start + 1
michael@0	143	used += num
michael@0	144
michael@0	145	uu = data.keys ()
michael@0	146	uu.sort ()
michael@0	147
michael@0	148	last = -1
michael@0	149	num = 0
michael@0	150	offset = 0
michael@0	151	starts = []
michael@0	152	ends = []
michael@0	153	print "static const INDIC_TABLE_ELEMENT_TYPE indic_table[] = {"
michael@0	154	for u in uu:
michael@0	155	if u <= last:
michael@0	156	continue
michael@0	157	block = data[u][2]
michael@0	158	(start, end) = blocks[block]
michael@0	159
michael@0	160	if start != last + 1:
michael@0	161	if start - last <= 33:
michael@0	162	print_block ("FILLER", last+1, start-1, data)
michael@0	163	last = start-1
michael@0	164	else:
michael@0	165	if last >= 0:
michael@0	166	ends.append (last + 1)
michael@0	167	offset += ends[-1] - starts[-1]
michael@0	168	print
michael@0	169	print
michael@0	170	print "#define indic_offset_0x%04x %d" % (start, offset)
michael@0	171	starts.append (start)
michael@0	172
michael@0	173	print_block (block, start, end, data)
michael@0	174	last = end
michael@0	175	ends.append (last + 1)
michael@0	176	offset += ends[-1] - starts[-1]
michael@0	177	print
michael@0	178	print
michael@0	179	print "#define indic_offset_total %d" % offset
michael@0	180	print
michael@0	181	occupancy = used * 100. / total
michael@0	182	print "}; /* Table occupancy: %d%% */" % occupancy
michael@0	183	print
michael@0	184	print "INDIC_TABLE_ELEMENT_TYPE"
michael@0	185	print "hb_indic_get_categories (hb_codepoint_t u)"
michael@0	186	print "{"
michael@0	187	for (start,end) in zip (starts, ends):
michael@0	188	offset = "indic_offset_0x%04x" % start
michael@0	189	print " if (0x%04X <= u && u <= 0x%04X) return indic_table[u - 0x%04X + %s];" % (start, end, start, offset)
michael@0	190	for u,d in singles.items ():
michael@0	191	print " if (unlikely (u == 0x%04X)) return _(%s,%s);" % (u, short[0][d[0]], short[1][d[1]])
michael@0	192	print " return _(x,x);"
michael@0	193	print "}"
michael@0	194	print
michael@0	195	print "#undef _"
michael@0	196	for i in range (2):
michael@0	197	print
michael@0	198	vv = values[i].keys ()
michael@0	199	vv.sort ()
michael@0	200	for v in vv:
michael@0	201	print "#undef %s_%s" % \
michael@0	202	(what_short[i], short[i][v])
michael@0	203	print
michael@0	204	print "/* == End of generated table == */"
michael@0	205
michael@0	206	# Maintain at least 30% occupancy in the table */
michael@0	207	if occupancy < 30:
michael@0	208	raise Exception ("Table too sparse, please investigate: ", occupancy)