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 IndicMatraCategory.txt Blocks.txt" 7 sys.exit (1) 8 9 files = [file (x) for x in sys.argv[1:]] 10 11 headers = [[f.readline () for i in range (2)] for f in files] 12 13 blocks = {} 14 data = [{} for f in files] 15 values = [{} for f in files] 16 for i, f in enumerate (files): 17 for line in f: 18 19 j = line.find ('#') 20 if j >= 0: 21 line = line[:j] 22 23 fields = [x.strip () for x in line.split (';')] 24 if len (fields) == 1: 25 continue 26 27 uu = fields[0].split ('..') 28 start = int (uu[0], 16) 29 if len (uu) == 1: 30 end = start 31 else: 32 end = int (uu[1], 16) 33 34 t = fields[1] 35 36 for u in range (start, end + 1): 37 data[i][u] = t 38 values[i][t] = values[i].get (t, 0) + 1 39 40 if i == 2: 41 blocks[t] = (start, end) 42 43 # Merge data into one dict: 44 defaults = ('Other', 'Not_Applicable', 'No_Block') 45 for i,v in enumerate (defaults): 46 values[i][v] = values[i].get (v, 0) + 1 47 combined = {} 48 for i,d in enumerate (data): 49 for u,v in d.items (): 50 if i == 2 and not u in combined: 51 continue 52 if not u in combined: 53 combined[u] = list (defaults) 54 combined[u][i] = v 55 data = combined 56 del combined 57 num = len (data) 58 59 # Move the outliers NO-BREAK SPACE and DOTTED CIRCLE out 60 singles = {} 61 for u in [0x00A0, 0x25CC]: 62 singles[u] = data[u] 63 del data[u] 64 65 print "/* == Start of generated table == */" 66 print "/*" 67 print " * The following table is generated by running:" 68 print " *" 69 print " * ./gen-indic-table.py IndicSyllabicCategory.txt IndicMatraCategory.txt Blocks.txt" 70 print " *" 71 print " * on files with these headers:" 72 print " *" 73 for h in headers: 74 for l in h: 75 print " * %s" % (l.strip()) 76 print " */" 77 print 78 print '#include "hb-ot-shape-complex-indic-private.hh"' 79 print 80 81 # Shorten values 82 short = [{ 83 "Bindu": 'Bi', 84 "Visarga": 'Vs', 85 "Vowel": 'Vo', 86 "Vowel_Dependent": 'M', 87 "Other": 'x', 88 },{ 89 "Not_Applicable": 'x', 90 }] 91 all_shorts = [[],[]] 92 93 # Add some of the values, to make them more readable, and to avoid duplicates 94 95 96 for i in range (2): 97 for v,s in short[i].items (): 98 all_shorts[i].append (s) 99 100 what = ["INDIC_SYLLABIC_CATEGORY", "INDIC_MATRA_CATEGORY"] 101 what_short = ["ISC", "IMC"] 102 for i in range (2): 103 print 104 vv = values[i].keys () 105 vv.sort () 106 for v in vv: 107 v_no_and = v.replace ('_And_', '_') 108 if v in short[i]: 109 s = short[i][v] 110 else: 111 s = ''.join ([c for c in v_no_and if ord ('A') <= ord (c) <= ord ('Z')]) 112 if s in all_shorts[i]: 113 raise Exception ("Duplicate short value alias", v, s) 114 all_shorts[i].append (s) 115 short[i][v] = s 116 print "#define %s_%s %s_%s %s/* %3d chars; %s */" % \ 117 (what_short[i], s, what[i], v.upper (), \ 118 ' '* ((48-1 - len (what[i]) - 1 - len (v)) / 8), \ 119 values[i][v], v) 120 print 121 print "#define _(S,M) INDIC_COMBINE_CATEGORIES (ISC_##S, IMC_##M)" 122 print 123 print 124 125 total = 0 126 used = 0 127 def print_block (block, start, end, data): 128 print 129 print 130 print " /* %s (%04X..%04X) */" % (block, start, end) 131 num = 0 132 for u in range (start, end+1): 133 if u % 8 == 0: 134 print 135 print " /* %04X */" % u, 136 if u in data: 137 num += 1 138 d = data.get (u, defaults) 139 sys.stdout.write ("%9s" % ("_(%s,%s)," % (short[0][d[0]], short[1][d[1]]))) 140 141 global total, used 142 total += end - start + 1 143 used += num 144 145 uu = data.keys () 146 uu.sort () 147 148 last = -1 149 num = 0 150 offset = 0 151 starts = [] 152 ends = [] 153 print "static const INDIC_TABLE_ELEMENT_TYPE indic_table[] = {" 154 for u in uu: 155 if u <= last: 156 continue 157 block = data[u][2] 158 (start, end) = blocks[block] 159 160 if start != last + 1: 161 if start - last <= 33: 162 print_block ("FILLER", last+1, start-1, data) 163 last = start-1 164 else: 165 if last >= 0: 166 ends.append (last + 1) 167 offset += ends[-1] - starts[-1] 168 print 169 print 170 print "#define indic_offset_0x%04x %d" % (start, offset) 171 starts.append (start) 172 173 print_block (block, start, end, data) 174 last = end 175 ends.append (last + 1) 176 offset += ends[-1] - starts[-1] 177 print 178 print 179 print "#define indic_offset_total %d" % offset 180 print 181 occupancy = used * 100. / total 182 print "}; /* Table occupancy: %d%% */" % occupancy 183 print 184 print "INDIC_TABLE_ELEMENT_TYPE" 185 print "hb_indic_get_categories (hb_codepoint_t u)" 186 print "{" 187 for (start,end) in zip (starts, ends): 188 offset = "indic_offset_0x%04x" % start 189 print " if (0x%04X <= u && u <= 0x%04X) return indic_table[u - 0x%04X + %s];" % (start, end, start, offset) 190 for u,d in singles.items (): 191 print " if (unlikely (u == 0x%04X)) return _(%s,%s);" % (u, short[0][d[0]], short[1][d[1]]) 192 print " return _(x,x);" 193 print "}" 194 print 195 print "#undef _" 196 for i in range (2): 197 print 198 vv = values[i].keys () 199 vv.sort () 200 for v in vv: 201 print "#undef %s_%s" % \ 202 (what_short[i], short[i][v]) 203 print 204 print "/* == End of generated table == */" 205 206 # Maintain at least 30% occupancy in the table */ 207 if occupancy < 30: 208 raise Exception ("Table too sparse, please investigate: ", occupancy) 209