tools/parse_kradfile.py

   1 #!/usr/bin/env python3
   2
   3 # Script for converting kradfile-u to SQL script.
   4 # Uses custom list of strokes count (FILE_STROKES).
   5
   6 # Check for components withou strokes (aoi will treat them as with strokes=0)
   7 #   select * from components where ifnull(strokes ,0)=0
   8
   9 import time
  10
  11 FILE_KRAD     = 'kradfile-u'
  12 FILE_STROKES  = 'components.strokes'
  13 SCRIPT_NAME   = 'script.components.sql'
  14 VERSION       = 'build: ' + time.strftime("%Y-%m-%d")
  15
  16 data = {}           # kanji:components
  17 set_comps = set()   # all used components
  18 components = {}     # component:strokes
  19
  20
  21 def load_kradfile():
  22   with open( FILE_KRAD ) as f:
  23     for line in f:
  24       line = line.strip()
  25       if ( line[0] == '#' ):
  26         continue
  27       line = line.replace( ' ', '' )
  28       kanji, comps = line.split(':')
  29       for c in comps:
  30         set_comps.add(c)
  31       data[kanji] = comps
  32   print("%s kanji, %s radicals" %(len(data),len(set_comps)));
  33
  34
  35 def load_strokes_file():
  36   with open( FILE_STROKES ) as f:
  37     for line in f:
  38       line = line.strip()
  39       component, strokes = line.split(':')
  40       components[component] = strokes
  41   print("%s components" %len(components))
  42
  43
  44 if __name__ == '__main__':
  45
  46   res = input("Kradfile-u version [%s]: " %VERSION)
  47   if res:
  48     VERSION = res
  49
  50   load_kradfile()
  51 #  load_strokes_file()
  52
  53   with open( SCRIPT_NAME, 'w' ) as f:
  54     f.write( "begin transaction;\n" )
  55     f.write("drop table if exists components;\n")
  56     f.write("create table components ( component TEXT, strokes INT );\n")
  57     f.write( "REPLACE INTO aoi (key,val) VALUES('kradfile_version','%s');\n" %VERSION)
  58     for c in set_comps:
  59       f.write("insert into components (component,strokes) values ('%s',NULL);\n" %(c))
  60     for d in data:
  61       f.write("update k_kanji set components='%s' where kanji='%s';\n" %(data[d],d));
  62     f.write( "end transaction;\n" )
  63     # add stroke count
  64     f.write( "begin transaction;\n" )
  65     f.write( "update components set strokes=(select strokes from k_kanji where kanji=component);\n" )
  66     # manual fix for kradfile-u
  67     f.write("update components set strokes=4 where component='⺹';\n")
  68     f.write("update components set strokes=3 where component='⺾';\n")
  69     f.write("update components set strokes=3 where component='⻖';\n")
  70     f.write("update components set strokes=2 where component='マ';\n")
  71     f.write("update components set strokes=1 where component='ノ';\n")
  72     f.write("update components set strokes=2 where component='ユ';\n")
  73     f.write("update components set strokes=2 where component='⺅';\n")
  74     f.write("update components set strokes=1 where component='｜';\n")
  75     f.write("update components set strokes=3 where component='⺌';\n")
  76     f.write("update components set strokes=3 where component='⻏';\n")
  77     f.write("update components set strokes=3 where component='ヨ';\n")
  78     f.write("update components set strokes=2 where component='ハ';\n")
  79     f.write( "end transaction;\n" )