1 #!/usr/bin/python 2 3 import sys 4 5 if len (sys.argv) != 4: 6 print >>sys.stderr, "usage: ./gen-indic-table.py IndicSyllabicCategory.txt IndicPositionalCategory.txt Blocks.txt" 7 sys.exit (1) 8 9 ALLOWED_SINGLES = [0x00A0, 0x25CC] 10 ALLOWED_BLOCKS = [ 11 'Basic Latin', 12 'Latin-1 Supplement', 13 'Devanagari', 14 'Bengali', 15 'Gurmukhi', 16 'Gujarati', 17 'Oriya', 18 'Tamil', 19 'Telugu', 20 'Kannada', 21 'Malayalam', 22 'Sinhala', 23 'Myanmar', 24 'Khmer', 25 'Vedic Extensions', 26 'General Punctuation', 27 'Superscripts and Subscripts', 28 'Devanagari Extended', 29 'Myanmar Extended-B', 30 'Myanmar Extended-A', 31 ] 32 33 files = [file (x) for x in sys.argv[1:]] 34 35 headers = [[f.readline () for i in range (2)] for f in files] 36 37 data = [{} for f in files] 38 values = [{} for f in files] 39 for i, f in enumerate (files): 40 for line in f: 41 42 j = line.find ('#') 43 if j >= 0: 44 line = line[:j] 45 46 fields = [x.strip () for x in line.split (';')] 47 if len (fields) == 1: 48 continue 49 50 uu = fields[0].split ('..') 51 start = int (uu[0], 16) 52 if len (uu) == 1: 53 end = start 54 else: 55 end = int (uu[1], 16) 56 57 t = fields[1] 58 59 for u in range (start, end + 1): 60 data[i][u] = t 61 values[i][t] = values[i].get (t, 0) + end - start + 1 62 63 # Merge data into one dict: 64 defaults = ('Other', 'Not_Applicable', 'No_Block') 65 for i,v in enumerate (defaults): 66 values[i][v] = values[i].get (v, 0) + 1 67 combined = {} 68 for i,d in enumerate (data): 69 for u,v in d.items (): 70 if i == 2 and not u in combined: 71 continue 72 if not u in combined: 73 combined[u] = list (defaults) 74 combined[u][i] = v 75 combined = {k:v for k,v in combined.items() if k in ALLOWED_SINGLES or v[2] in ALLOWED_BLOCKS} 76 data = combined 77 del combined 78 num = len (data) 79 80 for u in [0x17CD, 0x17CE, 0x17CF, 0x17D0, 0x17D3]: 81 if data[u][0] == 'Other': 82 data[u][0] = "Vowel_Dependent" 83 84 # Move the outliers NO-BREAK SPACE and DOTTED CIRCLE out 85 singles = {} 86 for u in ALLOWED_SINGLES: 87 singles[u] = data[u] 88 del data[u] 89 90 print "/* == Start of generated table == */" 91 print "/*" 92 print " * The following table is generated by running:" 93 print " *" 94 print " * ./gen-indic-table.py IndicSyllabicCategory.txt IndicPositionalCategory.txt Blocks.txt" 95 print " *" 96 print " * on files with these headers:" 97 print " *" 98 for h in headers: 99 for l in h: 100 print " * %s" % (l.strip()) 101 print " */" 102 print 103 print '#include "hb-ot-shape-complex-indic-private.hh"' 104 print 105 106 # Shorten values 107 short = [{ 108 "Bindu": 'Bi', 109 "Cantillation_Mark": 'Ca', 110 "Joiner": 'ZWJ', 111 "Non_Joiner": 'ZWNJ', 112 "Number": 'Nd', 113 "Visarga": 'Vs', 114 "Vowel": 'Vo', 115 "Vowel_Dependent": 'M', 116 "Consonant_Prefixed": 'CPrf', 117 "Other": 'x', 118 },{ 119 "Not_Applicable": 'x', 120 }] 121 all_shorts = [{},{}] 122 123 # Add some of the values, to make them more readable, and to avoid duplicates 124 125 126 for i in range (2): 127 for v,s in short[i].items (): 128 all_shorts[i][s] = v 129 130 what = ["INDIC_SYLLABIC_CATEGORY", "INDIC_MATRA_CATEGORY"] 131 what_short = ["ISC", "IMC"] 132 for i in range (2): 133 print 134 vv = values[i].keys () 135 vv.sort () 136 for v in vv: 137 v_no_and = v.replace ('_And_', '_') 138 if v in short[i]: 139 s = short[i][v] 140 else: 141 s = ''.join ([c for c in v_no_and if ord ('A') <= ord (c) <= ord ('Z')]) 142 if s in all_shorts[i]: 143 raise Exception ("Duplicate short value alias", v, all_shorts[i][s]) 144 all_shorts[i][s] = v 145 short[i][v] = s 146 print "#define %s_%s %s_%s %s/* %3d chars; %s */" % \ 147 (what_short[i], s, what[i], v.upper (), \ 148 ' '* ((48-1 - len (what[i]) - 1 - len (v)) / 8), \ 149 values[i][v], v) 150 print 151 print "#define _(S,M) INDIC_COMBINE_CATEGORIES (ISC_##S, IMC_##M)" 152 print 153 print 154 155 total = 0 156 used = 0 157 last_block = None 158 def print_block (block, start, end, data): 159 global total, used, last_block 160 if block and block != last_block: 161 print 162 print 163 print " /* %s */" % block 164 num = 0 165 assert start % 8 == 0 166 assert (end+1) % 8 == 0 167 for u in range (start, end+1): 168 if u % 8 == 0: 169 print 170 print " /* %04X */" % u, 171 if u in data: 172 num += 1 173 d = data.get (u, defaults) 174 sys.stdout.write ("%9s" % ("_(%s,%s)," % (short[0][d[0]], short[1][d[1]]))) 175 176 total += end - start + 1 177 used += num 178 if block: 179 last_block = block 180 181 uu = data.keys () 182 uu.sort () 183 184 last = -100000 185 num = 0 186 offset = 0 187 starts = [] 188 ends = [] 189 print "static const INDIC_TABLE_ELEMENT_TYPE indic_table[] = {" 190 for u in uu: 191 if u <= last: 192 continue 193 block = data[u][2] 194 195 start = u//8*8 196 end = start+1 197 while end in uu and block == data[end][2]: 198 end += 1 199 end = (end-1)//8*8 + 7 200 201 if start != last + 1: 202 if start - last <= 1+16*3: 203 print_block (None, last+1, start-1, data) 204 last = start-1 205 else: 206 if last >= 0: 207 ends.append (last + 1) 208 offset += ends[-1] - starts[-1] 209 print 210 print 211 print "#define indic_offset_0x%04xu %d" % (start, offset) 212 starts.append (start) 213 214 print_block (block, start, end, data) 215 last = end 216 ends.append (last + 1) 217 offset += ends[-1] - starts[-1] 218 print 219 print 220 occupancy = used * 100. / total 221 page_bits = 12 222 print "}; /* Table items: %d; occupancy: %d%% */" % (offset, occupancy) 223 print 224 print "INDIC_TABLE_ELEMENT_TYPE" 225 print "hb_indic_get_categories (hb_codepoint_t u)" 226 print "{" 227 print " switch (u >> %d)" % page_bits 228 print " {" 229 pages = set([u>>page_bits for u in starts+ends+singles.keys()]) 230 for p in sorted(pages): 231 print " case 0x%0Xu:" % p 232 for (start,end) in zip (starts, ends): 233 if p not in [start>>page_bits, end>>page_bits]: continue 234 offset = "indic_offset_0x%04xu" % start 235 print " if (hb_in_range (u, 0x%04Xu, 0x%04Xu)) return indic_table[u - 0x%04Xu + %s];" % (start, end-1, start, offset) 236 for u,d in singles.items (): 237 if p != u>>page_bits: continue 238 print " if (unlikely (u == 0x%04Xu)) return _(%s,%s);" % (u, short[0][d[0]], short[1][d[1]]) 239 print " break;" 240 print "" 241 print " default:" 242 print " break;" 243 print " }" 244 print " return _(x,x);" 245 print "}" 246 print 247 print "#undef _" 248 for i in range (2): 249 print 250 vv = values[i].keys () 251 vv.sort () 252 for v in vv: 253 print "#undef %s_%s" % \ 254 (what_short[i], short[i][v]) 255 print 256 print "/* == End of generated table == */" 257 258 # Maintain at least 30% occupancy in the table */ 259 if occupancy < 30: 260 raise Exception ("Table too sparse, please investigate: ", occupancy) 261
