1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/gfx/harfbuzz/src/gen-indic-table.py Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,208 @@
1.4 +#!/usr/bin/python
1.5 +
1.6 +import sys
1.7 +
1.8 +if len (sys.argv) != 4:
1.9 + print >>sys.stderr, "usage: ./gen-indic-table.py IndicSyllabicCategory.txt IndicMatraCategory.txt Blocks.txt"
1.10 + sys.exit (1)
1.11 +
1.12 +files = [file (x) for x in sys.argv[1:]]
1.13 +
1.14 +headers = [[f.readline () for i in range (2)] for f in files]
1.15 +
1.16 +blocks = {}
1.17 +data = [{} for f in files]
1.18 +values = [{} for f in files]
1.19 +for i, f in enumerate (files):
1.20 + for line in f:
1.21 +
1.22 + j = line.find ('#')
1.23 + if j >= 0:
1.24 + line = line[:j]
1.25 +
1.26 + fields = [x.strip () for x in line.split (';')]
1.27 + if len (fields) == 1:
1.28 + continue
1.29 +
1.30 + uu = fields[0].split ('..')
1.31 + start = int (uu[0], 16)
1.32 + if len (uu) == 1:
1.33 + end = start
1.34 + else:
1.35 + end = int (uu[1], 16)
1.36 +
1.37 + t = fields[1]
1.38 +
1.39 + for u in range (start, end + 1):
1.40 + data[i][u] = t
1.41 + values[i][t] = values[i].get (t, 0) + 1
1.42 +
1.43 + if i == 2:
1.44 + blocks[t] = (start, end)
1.45 +
1.46 +# Merge data into one dict:
1.47 +defaults = ('Other', 'Not_Applicable', 'No_Block')
1.48 +for i,v in enumerate (defaults):
1.49 + values[i][v] = values[i].get (v, 0) + 1
1.50 +combined = {}
1.51 +for i,d in enumerate (data):
1.52 + for u,v in d.items ():
1.53 + if i == 2 and not u in combined:
1.54 + continue
1.55 + if not u in combined:
1.56 + combined[u] = list (defaults)
1.57 + combined[u][i] = v
1.58 +data = combined
1.59 +del combined
1.60 +num = len (data)
1.61 +
1.62 +# Move the outliers NO-BREAK SPACE and DOTTED CIRCLE out
1.63 +singles = {}
1.64 +for u in [0x00A0, 0x25CC]:
1.65 + singles[u] = data[u]
1.66 + del data[u]
1.67 +
1.68 +print "/* == Start of generated table == */"
1.69 +print "/*"
1.70 +print " * The following table is generated by running:"
1.71 +print " *"
1.72 +print " * ./gen-indic-table.py IndicSyllabicCategory.txt IndicMatraCategory.txt Blocks.txt"
1.73 +print " *"
1.74 +print " * on files with these headers:"
1.75 +print " *"
1.76 +for h in headers:
1.77 + for l in h:
1.78 + print " * %s" % (l.strip())
1.79 +print " */"
1.80 +print
1.81 +print '#include "hb-ot-shape-complex-indic-private.hh"'
1.82 +print
1.83 +
1.84 +# Shorten values
1.85 +short = [{
1.86 + "Bindu": 'Bi',
1.87 + "Visarga": 'Vs',
1.88 + "Vowel": 'Vo',
1.89 + "Vowel_Dependent": 'M',
1.90 + "Other": 'x',
1.91 +},{
1.92 + "Not_Applicable": 'x',
1.93 +}]
1.94 +all_shorts = [[],[]]
1.95 +
1.96 +# Add some of the values, to make them more readable, and to avoid duplicates
1.97 +
1.98 +
1.99 +for i in range (2):
1.100 + for v,s in short[i].items ():
1.101 + all_shorts[i].append (s)
1.102 +
1.103 +what = ["INDIC_SYLLABIC_CATEGORY", "INDIC_MATRA_CATEGORY"]
1.104 +what_short = ["ISC", "IMC"]
1.105 +for i in range (2):
1.106 + print
1.107 + vv = values[i].keys ()
1.108 + vv.sort ()
1.109 + for v in vv:
1.110 + v_no_and = v.replace ('_And_', '_')
1.111 + if v in short[i]:
1.112 + s = short[i][v]
1.113 + else:
1.114 + s = ''.join ([c for c in v_no_and if ord ('A') <= ord (c) <= ord ('Z')])
1.115 + if s in all_shorts[i]:
1.116 + raise Exception ("Duplicate short value alias", v, s)
1.117 + all_shorts[i].append (s)
1.118 + short[i][v] = s
1.119 + print "#define %s_%s %s_%s %s/* %3d chars; %s */" % \
1.120 + (what_short[i], s, what[i], v.upper (), \
1.121 + ' '* ((48-1 - len (what[i]) - 1 - len (v)) / 8), \
1.122 + values[i][v], v)
1.123 +print
1.124 +print "#define _(S,M) INDIC_COMBINE_CATEGORIES (ISC_##S, IMC_##M)"
1.125 +print
1.126 +print
1.127 +
1.128 +total = 0
1.129 +used = 0
1.130 +def print_block (block, start, end, data):
1.131 + print
1.132 + print
1.133 + print " /* %s (%04X..%04X) */" % (block, start, end)
1.134 + num = 0
1.135 + for u in range (start, end+1):
1.136 + if u % 8 == 0:
1.137 + print
1.138 + print " /* %04X */" % u,
1.139 + if u in data:
1.140 + num += 1
1.141 + d = data.get (u, defaults)
1.142 + sys.stdout.write ("%9s" % ("_(%s,%s)," % (short[0][d[0]], short[1][d[1]])))
1.143 +
1.144 + global total, used
1.145 + total += end - start + 1
1.146 + used += num
1.147 +
1.148 +uu = data.keys ()
1.149 +uu.sort ()
1.150 +
1.151 +last = -1
1.152 +num = 0
1.153 +offset = 0
1.154 +starts = []
1.155 +ends = []
1.156 +print "static const INDIC_TABLE_ELEMENT_TYPE indic_table[] = {"
1.157 +for u in uu:
1.158 + if u <= last:
1.159 + continue
1.160 + block = data[u][2]
1.161 + (start, end) = blocks[block]
1.162 +
1.163 + if start != last + 1:
1.164 + if start - last <= 33:
1.165 + print_block ("FILLER", last+1, start-1, data)
1.166 + last = start-1
1.167 + else:
1.168 + if last >= 0:
1.169 + ends.append (last + 1)
1.170 + offset += ends[-1] - starts[-1]
1.171 + print
1.172 + print
1.173 + print "#define indic_offset_0x%04x %d" % (start, offset)
1.174 + starts.append (start)
1.175 +
1.176 + print_block (block, start, end, data)
1.177 + last = end
1.178 +ends.append (last + 1)
1.179 +offset += ends[-1] - starts[-1]
1.180 +print
1.181 +print
1.182 +print "#define indic_offset_total %d" % offset
1.183 +print
1.184 +occupancy = used * 100. / total
1.185 +print "}; /* Table occupancy: %d%% */" % occupancy
1.186 +print
1.187 +print "INDIC_TABLE_ELEMENT_TYPE"
1.188 +print "hb_indic_get_categories (hb_codepoint_t u)"
1.189 +print "{"
1.190 +for (start,end) in zip (starts, ends):
1.191 + offset = "indic_offset_0x%04x" % start
1.192 + print " if (0x%04X <= u && u <= 0x%04X) return indic_table[u - 0x%04X + %s];" % (start, end, start, offset)
1.193 +for u,d in singles.items ():
1.194 + print " if (unlikely (u == 0x%04X)) return _(%s,%s);" % (u, short[0][d[0]], short[1][d[1]])
1.195 +print " return _(x,x);"
1.196 +print "}"
1.197 +print
1.198 +print "#undef _"
1.199 +for i in range (2):
1.200 + print
1.201 + vv = values[i].keys ()
1.202 + vv.sort ()
1.203 + for v in vv:
1.204 + print "#undef %s_%s" % \
1.205 + (what_short[i], short[i][v])
1.206 +print
1.207 +print "/* == End of generated table == */"
1.208 +
1.209 +# Maintain at least 30% occupancy in the table */
1.210 +if occupancy < 30:
1.211 + raise Exception ("Table too sparse, please investigate: ", occupancy)
