michael@0: #!/usr/bin/python
michael@0: 
michael@0: import sys
michael@0: 
michael@0: if len (sys.argv) != 4:
michael@0: 	print >>sys.stderr, "usage: ./gen-indic-table.py IndicSyllabicCategory.txt IndicMatraCategory.txt Blocks.txt"
michael@0: 	sys.exit (1)
michael@0: 
michael@0: files = [file (x) for x in sys.argv[1:]]
michael@0: 
michael@0: headers = [[f.readline () for i in range (2)] for f in files]
michael@0: 
michael@0: blocks = {}
michael@0: data = [{} for f in files]
michael@0: values = [{} for f in files]
michael@0: for i, f in enumerate (files):
michael@0: 	for line in f:
michael@0: 
michael@0: 		j = line.find ('#')
michael@0: 		if j >= 0:
michael@0: 			line = line[:j]
michael@0: 
michael@0: 		fields = [x.strip () for x in line.split (';')]
michael@0: 		if len (fields) == 1:
michael@0: 			continue
michael@0: 
michael@0: 		uu = fields[0].split ('..')
michael@0: 		start = int (uu[0], 16)
michael@0: 		if len (uu) == 1:
michael@0: 			end = start
michael@0: 		else:
michael@0: 			end = int (uu[1], 16)
michael@0: 
michael@0: 		t = fields[1]
michael@0: 
michael@0: 		for u in range (start, end + 1):
michael@0: 			data[i][u] = t
michael@0: 		values[i][t] = values[i].get (t, 0) + 1
michael@0: 
michael@0: 		if i == 2:
michael@0: 			blocks[t] = (start, end)
michael@0: 
michael@0: # Merge data into one dict:
michael@0: defaults = ('Other', 'Not_Applicable', 'No_Block')
michael@0: for i,v in enumerate (defaults):
michael@0: 	values[i][v] = values[i].get (v, 0) + 1
michael@0: combined = {}
michael@0: for i,d in enumerate (data):
michael@0: 	for u,v in d.items ():
michael@0: 		if i == 2 and not u in combined:
michael@0: 			continue
michael@0: 		if not u in combined:
michael@0: 			combined[u] = list (defaults)
michael@0: 		combined[u][i] = v
michael@0: data = combined
michael@0: del combined
michael@0: num = len (data)
michael@0: 
michael@0: # Move the outliers NO-BREAK SPACE and DOTTED CIRCLE out
michael@0: singles = {}
michael@0: for u in [0x00A0, 0x25CC]:
michael@0: 	singles[u] = data[u]
michael@0: 	del data[u]
michael@0: 
michael@0: print "/* == Start of generated table == */"
michael@0: print "/*"
michael@0: print " * The following table is generated by running:"
michael@0: print " *"
michael@0: print " *   ./gen-indic-table.py IndicSyllabicCategory.txt IndicMatraCategory.txt Blocks.txt"
michael@0: print " *"
michael@0: print " * on files with these headers:"
michael@0: print " *"
michael@0: for h in headers:
michael@0: 	for l in h:
michael@0: 		print " * %s" % (l.strip())
michael@0: print " */"
michael@0: print
michael@0: print '#include "hb-ot-shape-complex-indic-private.hh"'
michael@0: print
michael@0: 
michael@0: # Shorten values
michael@0: short = [{
michael@0: 	"Bindu":		'Bi',
michael@0: 	"Visarga":		'Vs',
michael@0: 	"Vowel":		'Vo',
michael@0: 	"Vowel_Dependent":	'M',
michael@0: 	"Other":		'x',
michael@0: },{
michael@0: 	"Not_Applicable":	'x',
michael@0: }]
michael@0: all_shorts = [[],[]]
michael@0: 
michael@0: # Add some of the values, to make them more readable, and to avoid duplicates
michael@0: 
michael@0: 
michael@0: for i in range (2):
michael@0: 	for v,s in short[i].items ():
michael@0: 		all_shorts[i].append (s)
michael@0: 
michael@0: what = ["INDIC_SYLLABIC_CATEGORY", "INDIC_MATRA_CATEGORY"]
michael@0: what_short = ["ISC", "IMC"]
michael@0: for i in range (2):
michael@0: 	print
michael@0: 	vv = values[i].keys ()
michael@0: 	vv.sort ()
michael@0: 	for v in vv:
michael@0: 		v_no_and = v.replace ('_And_', '_')
michael@0: 		if v in short[i]:
michael@0: 			s = short[i][v]
michael@0: 		else:
michael@0: 			s = ''.join ([c for c in v_no_and if ord ('A') <= ord (c) <= ord ('Z')])
michael@0: 			if s in all_shorts[i]:
michael@0: 				raise Exception ("Duplicate short value alias", v, s)
michael@0: 			all_shorts[i].append (s)
michael@0: 			short[i][v] = s
michael@0: 		print "#define %s_%s	%s_%s	%s/* %3d chars; %s */" % \
michael@0: 			(what_short[i], s, what[i], v.upper (), \
michael@0: 			'	'* ((48-1 - len (what[i]) - 1 - len (v)) / 8), \
michael@0: 			values[i][v], v)
michael@0: print
michael@0: print "#define _(S,M) INDIC_COMBINE_CATEGORIES (ISC_##S, IMC_##M)"
michael@0: print
michael@0: print
michael@0: 
michael@0: total = 0
michael@0: used = 0
michael@0: def print_block (block, start, end, data):
michael@0: 	print
michael@0: 	print
michael@0: 	print "  /* %s  (%04X..%04X) */" % (block, start, end)
michael@0: 	num = 0
michael@0: 	for u in range (start, end+1):
michael@0: 		if u % 8 == 0:
michael@0: 			print
michael@0: 			print "  /* %04X */" % u,
michael@0: 		if u in data:
michael@0: 			num += 1
michael@0: 		d = data.get (u, defaults)
michael@0: 		sys.stdout.write ("%9s" % ("_(%s,%s)," % (short[0][d[0]], short[1][d[1]])))
michael@0: 
michael@0: 	global total, used
michael@0: 	total += end - start + 1
michael@0: 	used += num
michael@0: 
michael@0: uu = data.keys ()
michael@0: uu.sort ()
michael@0: 
michael@0: last = -1
michael@0: num = 0
michael@0: offset = 0
michael@0: starts = []
michael@0: ends = []
michael@0: print "static const INDIC_TABLE_ELEMENT_TYPE indic_table[] = {"
michael@0: for u in uu:
michael@0: 	if u <= last:
michael@0: 		continue
michael@0: 	block = data[u][2]
michael@0: 	(start, end) = blocks[block]
michael@0: 
michael@0: 	if start != last + 1:
michael@0: 		if start - last <= 33:
michael@0: 			print_block ("FILLER", last+1, start-1, data)
michael@0: 			last = start-1
michael@0: 		else:
michael@0: 			if last >= 0:
michael@0: 				ends.append (last + 1)
michael@0: 				offset += ends[-1] - starts[-1]
michael@0: 			print
michael@0: 			print
michael@0: 			print "#define indic_offset_0x%04x %d" % (start, offset)
michael@0: 			starts.append (start)
michael@0: 
michael@0: 	print_block (block, start, end, data)
michael@0: 	last = end
michael@0: ends.append (last + 1)
michael@0: offset += ends[-1] - starts[-1]
michael@0: print
michael@0: print
michael@0: print "#define indic_offset_total %d" % offset
michael@0: print
michael@0: occupancy = used * 100. / total
michael@0: print "}; /* Table occupancy: %d%% */" % occupancy
michael@0: print
michael@0: print "INDIC_TABLE_ELEMENT_TYPE"
michael@0: print "hb_indic_get_categories (hb_codepoint_t u)"
michael@0: print "{"
michael@0: for (start,end) in zip (starts, ends):
michael@0: 	offset = "indic_offset_0x%04x" % start
michael@0: 	print "  if (0x%04X <= u && u <= 0x%04X) return indic_table[u - 0x%04X + %s];" % (start, end, start, offset)
michael@0: for u,d in singles.items ():
michael@0: 	print "  if (unlikely (u == 0x%04X)) return _(%s,%s);" % (u, short[0][d[0]], short[1][d[1]])
michael@0: print "  return _(x,x);"
michael@0: print "}"
michael@0: print
michael@0: print "#undef _"
michael@0: for i in range (2):
michael@0: 	print
michael@0: 	vv = values[i].keys ()
michael@0: 	vv.sort ()
michael@0: 	for v in vv:
michael@0: 		print "#undef %s_%s" % \
michael@0: 			(what_short[i], short[i][v])
michael@0: print
michael@0: print "/* == End of generated table == */"
michael@0: 
michael@0: # Maintain at least 30% occupancy in the table */
michael@0: if occupancy < 30:
michael@0: 	raise Exception ("Table too sparse, please investigate: ", occupancy)