Code cleanup: conv_cmap.py
parent
1455f134c6
commit
774827b4ce
|
@ -4,9 +4,36 @@ import os.path
|
|||
import gzip
|
||||
import cPickle as pickle
|
||||
|
||||
def process_cid2code(fp, check_codecs=[]):
|
||||
|
||||
def get_canonicals(name):
|
||||
## CMapConverter
|
||||
##
|
||||
class CMapConverter(object):
|
||||
|
||||
def __init__(self, check_codecs=[]):
|
||||
self.check_codecs = check_codecs
|
||||
self.code2cid = {} # {'cmapname': ...}
|
||||
self.is_vertical = {}
|
||||
self.cid2unichr_h = {} # {cid: unichr}
|
||||
self.cid2unichr_v = {} # {cid: unichr}
|
||||
return
|
||||
|
||||
def load(self, fp):
|
||||
names = []
|
||||
for line in fp:
|
||||
(line,_,_) = line.strip().partition('#')
|
||||
if not line: continue
|
||||
values = line.split('\t')
|
||||
if not names:
|
||||
names = values
|
||||
continue
|
||||
d = dict( (k,v) for (k,v) in zip(names, values) if v != '*' )
|
||||
cid = int(d['CID'])
|
||||
for (key,value) in d.iteritems():
|
||||
if key == 'CID': continue
|
||||
self._register(cid, key, value)
|
||||
return
|
||||
|
||||
def get_canonicals(self, name):
|
||||
if name.endswith('-H'):
|
||||
return (name, None)
|
||||
elif name == 'H':
|
||||
|
@ -14,14 +41,14 @@ def process_cid2code(fp, check_codecs=[]):
|
|||
else:
|
||||
return (name+'-H', name+'-V')
|
||||
|
||||
def get_unichr(codes):
|
||||
def get_unichr(self, codes):
|
||||
# determine the "most popular" candidate.
|
||||
d = {}
|
||||
for code in codes:
|
||||
char = unicode(code, 'utf-8')
|
||||
if char not in d:
|
||||
d[char] = 0
|
||||
for codec in check_codecs:
|
||||
for codec in self.check_codecs:
|
||||
try:
|
||||
char.encode(codec, 'strict')
|
||||
d[char] += 1
|
||||
|
@ -30,89 +57,74 @@ def process_cid2code(fp, check_codecs=[]):
|
|||
chars = sorted(d.keys(), key=lambda char:d[char], reverse=True)
|
||||
return chars[0]
|
||||
|
||||
def put(dmap, code, cid, force=False):
|
||||
for b in code[:-1]:
|
||||
b = ord(b)
|
||||
if b in dmap:
|
||||
dmap = dmap[b]
|
||||
def _register(self, cid, key, value):
|
||||
def put(dmap, code, cid, force=False):
|
||||
for b in code[:-1]:
|
||||
b = ord(b)
|
||||
if b in dmap:
|
||||
dmap = dmap[b]
|
||||
else:
|
||||
d = {}
|
||||
dmap[b] = d
|
||||
dmap = d
|
||||
b = ord(code[-1])
|
||||
if force or ((b not in dmap) or dmap[b] == cid):
|
||||
dmap[b] = cid
|
||||
return
|
||||
|
||||
(hmapname, vmapname) = self.get_canonicals(key)
|
||||
if hmapname in self.code2cid:
|
||||
hmap = self.code2cid[hmapname]
|
||||
else:
|
||||
hmap = {}
|
||||
self.code2cid[hmapname] = hmap
|
||||
vmap = None
|
||||
if vmapname:
|
||||
self.is_vertical[vmapname] = True
|
||||
if vmapname in self.code2cid:
|
||||
vmap = self.code2cid[vmapname]
|
||||
else:
|
||||
d = {}
|
||||
dmap[b] = d
|
||||
dmap = d
|
||||
b = ord(code[-1])
|
||||
if force or ((b not in dmap) or dmap[b] == cid):
|
||||
dmap[b] = cid
|
||||
vmap = {}
|
||||
self.code2cid[vmapname] = vmap
|
||||
|
||||
hcodes = []
|
||||
vcodes = []
|
||||
for code in value.split(','):
|
||||
vertical = code.endswith('v')
|
||||
if vertical:
|
||||
code = code[:-1]
|
||||
try:
|
||||
code = code.decode('hex')
|
||||
except:
|
||||
code = chr(int(code, 16))
|
||||
if vertical:
|
||||
vcodes.append(code)
|
||||
else:
|
||||
hcodes.append(code)
|
||||
|
||||
if vcodes:
|
||||
assert vmap is not None
|
||||
for code in vcodes:
|
||||
put(vmap, code, cid, True)
|
||||
for code in hcodes:
|
||||
put(hmap, code, cid, True)
|
||||
if key.endswith('-UTF8'):
|
||||
if hcodes:
|
||||
self.cid2unichr_h[cid] = self.get_unichr(hcodes)
|
||||
if vcodes:
|
||||
self.cid2unichr_v[cid] = self.get_unichr(vcodes)
|
||||
else:
|
||||
for code in hcodes:
|
||||
put(hmap, code, cid)
|
||||
put(vmap, code, cid)
|
||||
if key.endswith('-UTF8') and hcodes:
|
||||
code = self.get_unichr(hcodes)
|
||||
if cid not in self.cid2unichr_h:
|
||||
self.cid2unichr_h[cid] = code
|
||||
if cid not in self.cid2unichr_v:
|
||||
self.cid2unichr_v[cid] = code
|
||||
return
|
||||
|
||||
names = []
|
||||
code2cid = {} # {'cmapname': ...}
|
||||
is_vertical = {}
|
||||
cid2unichr_h = {} # {cid: unichr}
|
||||
cid2unichr_v = {} # {cid: unichr}
|
||||
|
||||
for line in fp:
|
||||
line = line.strip()
|
||||
if line.startswith('#'): continue
|
||||
if line.startswith('CID'):
|
||||
names = line.split('\t')[1:]
|
||||
continue
|
||||
f = line.split('\t')
|
||||
if not f: continue
|
||||
cid = int(f[0])
|
||||
for (x,name) in zip(f[1:], names):
|
||||
if x == '*': continue
|
||||
(hmapname, vmapname) = get_canonicals(name)
|
||||
if hmapname in code2cid:
|
||||
hmap = code2cid[hmapname]
|
||||
else:
|
||||
hmap = {}
|
||||
code2cid[hmapname] = hmap
|
||||
vmap = None
|
||||
if vmapname:
|
||||
is_vertical[vmapname] = True
|
||||
if vmapname in code2cid:
|
||||
vmap = code2cid[vmapname]
|
||||
else:
|
||||
vmap = {}
|
||||
code2cid[vmapname] = vmap
|
||||
hcodes = []
|
||||
vcodes = []
|
||||
for code in x.split(','):
|
||||
vertical = code.endswith('v')
|
||||
if vertical:
|
||||
code = code[:-1]
|
||||
try:
|
||||
code = code.decode('hex')
|
||||
except:
|
||||
code = chr(int(code, 16))
|
||||
if vertical:
|
||||
vcodes.append(code)
|
||||
else:
|
||||
hcodes.append(code)
|
||||
if vcodes:
|
||||
assert vmap is not None
|
||||
for code in vcodes:
|
||||
put(vmap, code, cid, True)
|
||||
for code in hcodes:
|
||||
put(hmap, code, cid, True)
|
||||
if name.endswith('-UTF8'):
|
||||
if hcodes:
|
||||
cid2unichr_h[cid] = get_unichr(hcodes)
|
||||
if vcodes:
|
||||
cid2unichr_v[cid] = get_unichr(vcodes)
|
||||
else:
|
||||
for code in hcodes:
|
||||
put(hmap, code, cid)
|
||||
put(vmap, code, cid)
|
||||
if name.endswith('-UTF8') and hcodes:
|
||||
code = get_unichr(hcodes)
|
||||
if cid not in cid2unichr_h:
|
||||
cid2unichr_h[cid] = code
|
||||
if cid not in cid2unichr_v:
|
||||
cid2unichr_v[cid] = code
|
||||
|
||||
return (code2cid, is_vertical, cid2unichr_h, cid2unichr_v)
|
||||
|
||||
# main
|
||||
def main(argv):
|
||||
|
||||
|
@ -126,16 +138,17 @@ def main(argv):
|
|||
check_codecs = args[3:]
|
||||
|
||||
print >>sys.stderr, 'reading %r...' % src
|
||||
converter = CMapConverter(check_codecs)
|
||||
fp = file(src)
|
||||
(code2cid, is_vertical, cid2unichr_h, cid2unichr_v) = process_cid2code(fp, check_codecs)
|
||||
converter.load(fp)
|
||||
fp.close()
|
||||
|
||||
for (name, cmap) in code2cid.iteritems():
|
||||
for (name, cmap) in converter.code2cid.iteritems():
|
||||
fname = '%s.pickle.gz' % name
|
||||
print >>sys.stderr, 'writing %r...' % fname
|
||||
fp = gzip.open(os.path.join(outdir, fname), 'wb')
|
||||
data = dict(
|
||||
IS_VERTICAL=is_vertical.get(name, False),
|
||||
IS_VERTICAL=converter.is_vertical.get(name, False),
|
||||
CODE2CID=cmap,
|
||||
)
|
||||
fp.write(pickle.dumps(data))
|
||||
|
@ -145,12 +158,11 @@ def main(argv):
|
|||
print >>sys.stderr, 'writing %r...' % fname
|
||||
fp = gzip.open(os.path.join(outdir, fname), 'wb')
|
||||
data = dict(
|
||||
CID2UNICHR_H=cid2unichr_h,
|
||||
CID2UNICHR_V=cid2unichr_v,
|
||||
CID2UNICHR_H=converter.cid2unichr_h,
|
||||
CID2UNICHR_V=converter.cid2unichr_v,
|
||||
)
|
||||
fp.write(pickle.dumps(data))
|
||||
fp.close()
|
||||
|
||||
return 0
|
||||
return
|
||||
|
||||
if __name__ == '__main__': sys.exit(main(sys.argv))
|
||||
|
|
Loading…
Reference in New Issue