def __init__(self, spec, splitter="><", ignore_first_character=True):
assert isinstance(spec, six.text_type)
self.spec = spec.rstrip()
self.rels = None
self._tag = None
if len(spec) == 0:
return
tag_start = 0
if ignore_first_character:
tag_start = 1
tag_end = None
while tag_end != -1:
tag_end = self.spec.find(splitter, tag_start)
kv_splitter = self.spec.find(':', tag_start, tag_end)
if self.spec[tag_start:].startswith('rel '):
rel = Rel(self.spec[tag_start:tag_end])
if rel.ignore == False:
if self.rels is None:
self.rels = []
self.rels.append(rel)
elif kv_splitter == -1:
key = self.spec[tag_start:tag_end]
val = True # Dummy value
self[key] = val
else:
key = self.spec[tag_start:kv_splitter]
val = self.spec[kv_splitter + 1:tag_end]
self[key] = val
tag_start = tag_end + len(splitter)
def _extract_rel_tags(tag: Tag) -> List[Rel]:
"""parse tag.fstring to extract <rel> tags"""
splitter = "><"
rels = []
spec = tag.fstring
tag_start = 1
tag_end = None
while tag_end != -1:
tag_end = spec.find(splitter, tag_start)
if spec[tag_start:].startswith('rel '):
rel = Rel(spec[tag_start:tag_end])
if rel.target:
rel.target = jaconv.h2z(rel.target, digit=True) # 不特定:人1 -> 不特定:人１
if rel.atype is not None:
rels.append(rel)
tag_start = tag_end + len(splitter)
return rels
def __init__(self, spec, splitter="><", ignore_first_character=True):
dict.__init__(self)
assert isinstance(spec, str)
self.spec = spec.rstrip()
self.pas = None
self.rels = None
if not spec:
return
tag_start = 0
if ignore_first_character:
tag_start = 1
tag_end = None
while tag_end != -1:
tag_end = self.spec.find(splitter, tag_start)
kv_splitter = self.spec.find(':', tag_start, tag_end)
if self.spec[tag_start:].startswith('rel '):
rel = Rel(self.spec[tag_start:tag_end])
if not rel.ignore:
if self.rels is None:
self.rels = []
self.rels.append(rel)
elif kv_splitter == -1:
key = self.spec[tag_start:tag_end]
val = True # Dummy value
self[key] = val
else:
key = self.spec[tag_start:kv_splitter]
val = self.spec[kv_splitter + 1:tag_end]
self[key] = val
if key == '格解析結果':
self.pas = Pas(val, knpstyle=True)
tag_start = tag_end + len(splitter)