def process_referent(self, begin: 'Token',
end: 'Token') -> 'ReferentToken':
pat = PersonAttrToken.try_attach(
begin, None, PersonAttrToken.PersonAttrAttachAttrs.INPROCESS)
if (pat is not None and pat.prop_ref is not None):
return ReferentToken._new2595(pat.prop_ref, pat.begin_token,
pat.end_token, pat.morph, pat)
return None
def isPersonSayOrAttrAfter(t : 'Token') -> bool:
if (t is None):
return False
tt = PersonHelper.__correctTailAttributes(None, t)
if (tt is not None and tt != t):
return True
if (t.is_comma and t.next0_ is not None):
t = t.next0_
if (t.chars.is_latin_letter):
if (t.isValue("SAY", None) or t.isValue("ASK", None) or t.isValue("WHO", None)):
return True
if (t.isChar('.') and (isinstance(t.next0_, TextToken)) and ((t.next0_.morph.class0_.is_pronoun or t.next0_.morph.class0_.is_personal_pronoun))):
if (t.next0_.morph.gender == MorphGender.FEMINIE or t.next0_.morph.gender == MorphGender.MASCULINE):
return True
if (t.is_comma and t.next0_ is not None):
t = t.next0_
if (PersonAttrToken.tryAttach(t, None, PersonAttrToken.PersonAttrAttachAttrs.NO) is not None):
return True
return False
def _createReferentToken(p : 'PersonReferent', begin : 'Token', end : 'Token', morph_ : 'MorphCollection', attrs : typing.List['PersonAttrToken'], ad : 'PersonAnalyzerData', for_attribute : bool, after_be_predicate : bool) -> 'ReferentToken':
from pullenti.ner.person.internal.PersonIdentityToken import PersonIdentityToken
if (p is None):
return None
has_prefix = False
if (attrs is not None):
for a in attrs:
if (a.typ == PersonAttrTerminType.BESTREGARDS):
has_prefix = True
else:
if (a.begin_char < begin.begin_char):
begin = a.begin_token
if (a.typ != PersonAttrTerminType.PREFIX):
if (a.age is not None):
p.addSlot(PersonReferent.ATTR_AGE, a.age, False, 0)
if (a.prop_ref is None):
p.addSlot(PersonReferent.ATTR_ATTR, a.value, False, 0)
else:
p.addSlot(PersonReferent.ATTR_ATTR, a, False, 0)
elif (a.gender == MorphGender.FEMINIE and not p.is_female):
p.is_female = True
elif (a.gender == MorphGender.MASCULINE and not p.is_male):
p.is_male = True
elif ((isinstance(begin.previous, TextToken)) and (begin.whitespaces_before_count < 3)):
if ((begin.previous).term == "ИП"):
a = PersonAttrToken(begin.previous, begin.previous)
a.prop_ref = PersonPropertyReferent()
a.prop_ref.name = "индивидуальный предприниматель"
p.addSlot(PersonReferent.ATTR_ATTR, a, False, 0)
begin = begin.previous
m0 = MorphCollection()
for it in morph_.items:
bi = MorphBaseInfo(it)
bi.number = MorphNumber.SINGULAR
if (bi.gender == MorphGender.UNDEFINED):
if (p.is_male and not p.is_female):
bi.gender = MorphGender.MASCULINE
if (not p.is_male and p.is_female):
bi.gender = MorphGender.FEMINIE
m0.addItem(bi)
morph_ = m0
if ((attrs is not None and len(attrs) > 0 and not attrs[0].morph.case_.is_undefined) and morph_.case_.is_undefined):
morph_.case_ = attrs[0].morph.case_
if (attrs[0].morph.number == MorphNumber.SINGULAR):
morph_.number = MorphNumber.SINGULAR
if (p.is_male and not p.is_female):
morph_.gender = MorphGender.MASCULINE
elif (p.is_female):
morph_.gender = MorphGender.FEMINIE
if (begin.previous is not None):
ttt = begin.previous
if (ttt.isValue("ИМЕНИ", "ІМЕНІ")):
for_attribute = True
else:
if (ttt.isChar('.') and ttt.previous is not None):
ttt = ttt.previous
if (ttt.whitespaces_after_count < 3):
if (ttt.isValue("ИМ", "ІМ")):
for_attribute = True
if (for_attribute):
return ReferentToken._new2329(p, begin, end, morph_, p._m_person_identity_typ)
if ((begin.previous is not None and begin.previous.is_comma_and and (isinstance(begin.previous.previous, ReferentToken))) and (isinstance(begin.previous.previous.getReferent(), PersonReferent))):
rt00 = Utils.asObjectOrNull(begin.previous.previous, ReferentToken)
ttt = rt00
while ttt is not None:
if (ttt.previous is None or not ((isinstance(ttt.previous.previous, ReferentToken)))):
break
if (not ttt.previous.is_comma_and or not ((isinstance(ttt.previous.previous.getReferent(), PersonReferent)))):
break
rt00 = (Utils.asObjectOrNull(ttt.previous.previous, ReferentToken))
ttt = (rt00)
if (isinstance(rt00.begin_token.getReferent(), PersonPropertyReferent)):
ok = False
if ((rt00.begin_token).end_token.next0_ is not None and (rt00.begin_token).end_token.next0_.isChar(':')):
ok = True
elif (rt00.begin_token.morph.number == MorphNumber.PLURAL):
ok = True
if (ok):
p.addSlot(PersonReferent.ATTR_ATTR, rt00.begin_token.getReferent(), False, 0)
if (ad is not None):
if (ad.overflow_level > 10):
return ReferentToken._new2329(p, begin, end, morph_, p._m_person_identity_typ)
ad.overflow_level += 1
attrs1 = None
has_position = False
open_br = False
t = end.next0_
first_pass3095 = True
while True:
if first_pass3095: first_pass3095 = False
else: t = t.next0_
if (not (t is not None)): break
if (t.is_table_control_char):
break
if (t.is_newline_before):
if (t.newlines_before_count > 2):
break
if (attrs1 is not None and len(attrs1) > 0):
break
ml = MailLine.parse(t, 0)
if (ml is not None and ml.typ == MailLine.Types.FROM):
break
if (t.chars.is_capital_upper):
attr1 = PersonAttrToken.tryAttach(t, (None if ad is None else ad.local_ontology), PersonAttrToken.PersonAttrAttachAttrs.NO)
ok1 = False
if (attr1 is not None):
if (has_prefix or attr1.is_newline_after or ((attr1.end_token.next0_ is not None and attr1.end_token.next0_.is_table_control_char))):
ok1 = True
else:
tt2 = t.next0_
while tt2 is not None and tt2.end_char <= attr1.end_char:
if (tt2.is_whitespace_before):
ok1 = True
tt2 = tt2.next0_
else:
ttt = PersonHelper.__correctTailAttributes(p, t)
if (ttt is not None and ttt != t):
t = ttt
end = t
continue
if (not ok1):
break
if (t.is_hiphen or t.isCharOf("_>|")):
continue
if (t.isValue("МОДЕЛЬ", None)):
break
tt = PersonHelper.__correctTailAttributes(p, t)
if (tt != t and tt is not None):
t = tt
end = t
continue
is_be = False
if (t.isChar('(') and t == end.next0_):
open_br = True
t = t.next0_
if (t is None):
break
pit1 = PersonItemToken.tryAttach(t, None, PersonItemToken.ParseAttr.NO, None)
if ((pit1 is not None and t.chars.is_capital_upper and pit1.end_token.next0_ is not None) and (isinstance(t, TextToken)) and pit1.end_token.next0_.isChar(')')):
if (pit1.lastname is not None):
inf = MorphBaseInfo._new2321(MorphCase.NOMINATIVE)
if (p.is_male):
inf.gender = Utils.valToEnum((inf.gender) | (MorphGender.MASCULINE), MorphGender)
if (p.is_female):
inf.gender = Utils.valToEnum((inf.gender) | (MorphGender.FEMINIE), MorphGender)
sur = PersonIdentityToken.createLastname(pit1, inf)
if (sur is not None):
p._addFioIdentity(sur, None, None)
t = pit1.end_token.next0_
end = t
continue
elif (t.is_comma):
t = t.next0_
if ((isinstance(t, TextToken)) and (t).isValue("WHO", None)):
continue
elif ((isinstance(t, TextToken)) and (t).is_verb_be):
t = t.next0_
elif (t.is_and and t.is_whitespace_after and not t.is_newline_after):
if (t == end.next0_):
break
t = t.next0_
elif (t.is_hiphen and t == end.next0_):
t = t.next0_
elif (t.isChar('.') and t == end.next0_ and has_prefix):
t = t.next0_
ttt2 = PersonHelper.createNickname(p, t)
if (ttt2 is not None):
end = ttt2
t = end
continue
if (t is None):
break
attr = None
attr = PersonAttrToken.tryAttach(t, (None if ad is None else ad.local_ontology), PersonAttrToken.PersonAttrAttachAttrs.NO)
if (attr is None):
if ((t is not None and t.getReferent() is not None and t.getReferent().type_name == "GEO") and attrs1 is not None and open_br):
continue
if ((t.chars.is_capital_upper and open_br and t.next0_ is not None) and t.next0_.isChar(')')):
if (p.findSlot(PersonReferent.ATTR_LASTNAME, None, True) is None):
p.addSlot(PersonReferent.ATTR_LASTNAME, t.getSourceText().upper(), False, 0)
t = t.next0_
end = t
if (t is not None and t.isValue("КОТОРЫЙ", None) and t.morph.number == MorphNumber.SINGULAR):
if (not p.is_female and t.morph.gender == MorphGender.FEMINIE):
p.is_female = True
p._correctData()
elif (not p.is_male and t.morph.gender == MorphGender.MASCULINE):
p.is_male = True
p._correctData()
break
if (attr.morph.number == MorphNumber.PLURAL):
break
if (attr.typ == PersonAttrTerminType.BESTREGARDS):
break
if (attr.is_doubt):
if (has_prefix):
pass
elif (t.is_newline_before and attr.is_newline_after):
pass
elif (t.previous is not None and ((t.previous.is_hiphen or t.previous.isChar(':')))):
pass
else:
break
if (not morph_.case_.is_undefined and not attr.morph.case_.is_undefined):
if (((morph_.case_) & attr.morph.case_).is_undefined and not is_be):
break
if (open_br):
if (PersonAnalyzer._tryAttachPerson(t, ad, False, 0, True) is not None):
break
if (attrs1 is None):
if (t.previous.is_comma and t.previous == end.next0_):
ttt = attr.end_token.next0_
if (ttt is not None):
if (ttt.morph.class0_.is_verb):
if (MiscHelper.canBeStartOfSentence(begin)):
pass
else:
break
attrs1 = list()
attrs1.append(attr)
if (attr.typ == PersonAttrTerminType.POSITION or attr.typ == PersonAttrTerminType.KING):
if (not is_be):
has_position = True
elif (attr.typ != PersonAttrTerminType.PREFIX):
if (attr.typ == PersonAttrTerminType.OTHER and attr.age is not None):
pass
else:
attrs1 = (None)
break
t = attr.end_token
if (attrs1 is not None and has_position and attrs is not None):
te1 = attrs[len(attrs) - 1].end_token.next0_
te2 = attrs1[0].begin_token
if (te1.whitespaces_after_count > te2.whitespaces_before_count and (te2.whitespaces_before_count < 2)):
pass
elif (attrs1[0].age is not None):
pass
elif (((te1.is_hiphen or te1.isChar(':'))) and not attrs1[0].is_newline_before and ((te2.previous.is_comma or te2.previous == end))):
pass
else:
for a in attrs:
if (a.typ == PersonAttrTerminType.POSITION):
te = attrs1[len(attrs1) - 1].end_token
if (te.next0_ is not None):
if (not te.next0_.isChar('.')):
attrs1 = (None)
break
if (attrs1 is not None and not has_prefix):
attr = attrs1[len(attrs1) - 1]
ok = False
if (attr.end_token.next0_ is not None and attr.end_token.next0_.chars.is_capital_upper):
ok = True
else:
rt = PersonAnalyzer._tryAttachPerson(attr.begin_token, ad, False, -1, False)
if (rt is not None and (isinstance(rt.referent, PersonReferent))):
ok = True
if (ok):
if (attr.begin_token.whitespaces_before_count > attr.end_token.whitespaces_after_count):
attrs1 = (None)
elif (attr.begin_token.whitespaces_before_count == attr.end_token.whitespaces_after_count):
rt1 = PersonAnalyzer._tryAttachPerson(attr.begin_token, ad, False, -1, False)
if (rt1 is not None):
attrs1 = (None)
if (attrs1 is not None):
for a in attrs1:
if (a.typ != PersonAttrTerminType.PREFIX):
if (a.age is not None):
p.addSlot(PersonReferent.ATTR_AGE, a.age, True, 0)
elif (a.prop_ref is None):
p.addSlot(PersonReferent.ATTR_ATTR, a.value, False, 0)
else:
p.addSlot(PersonReferent.ATTR_ATTR, a, False, 0)
end = a.end_token
if (a.gender != MorphGender.UNDEFINED and not p.is_female and not p.is_male):
if (a.gender == MorphGender.MASCULINE and not p.is_male):
p.is_male = True
p._correctData()
elif (a.gender == MorphGender.FEMINIE and not p.is_female):
p.is_female = True
p._correctData()
if (open_br):
if (end.next0_ is not None and end.next0_.isChar(')')):
end = end.next0_
crlf_cou = 0
t = end.next0_
first_pass3096 = True
while True:
if first_pass3096: first_pass3096 = False
else: t = t.next0_
if (not (t is not None)): break
if (t.is_table_control_char):
break
if (t.is_newline_before):
ml = MailLine.parse(t, 0)
if (ml is not None and ml.typ == MailLine.Types.FROM):
break
crlf_cou += 1
if (t.isCharOf(":,(") or t.is_hiphen):
continue
if (t.isChar('.') and t == end.next0_):
continue
r = t.getReferent()
if (r is not None):
if (r.type_name == "PHONE" or r.type_name == "URI" or r.type_name == "ADDRESS"):
ty = r.getStringValue("SCHEME")
if (r.type_name == "URI"):
if ((ty != "mailto" and ty != "skype" and ty != "ICQ") and ty != "http"):
break
p._addContact(r)
end = t
crlf_cou = 0
continue
if (isinstance(r, PersonIdentityReferent)):
p.addSlot(PersonReferent.ATTR_IDDOC, r, False, 0)
end = t
crlf_cou = 0
continue
if (r is not None and r.type_name == "ORGANIZATION"):
if (t.next0_ is not None and t.next0_.morph.class0_.is_verb):
break
if (begin.previous is not None and begin.previous.morph.class0_.is_verb):
break
if (t.whitespaces_after_count == 1):
break
exist = False
for s in p.slots:
if (s.type_name == PersonReferent.ATTR_ATTR and (isinstance(s.value, PersonPropertyReferent))):
pr = Utils.asObjectOrNull(s.value, PersonPropertyReferent)
if (pr.findSlot(PersonPropertyReferent.ATTR_REF, r, True) is not None):
exist = True
break
elif (s.type_name == PersonReferent.ATTR_ATTR and (isinstance(s.value, PersonAttrToken))):
pr = Utils.asObjectOrNull(s.value, PersonAttrToken)
if (pr.referent.findSlot(PersonPropertyReferent.ATTR_REF, r, True) is not None):
exist = True
break
if (not exist):
pat = PersonAttrToken(t, t)
pat.prop_ref = PersonPropertyReferent._new2291("сотрудник")
pat.prop_ref.addSlot(PersonPropertyReferent.ATTR_REF, r, False, 0)
p.addSlot(PersonReferent.ATTR_ATTR, pat, False, 0)
continue
if (r is not None):
break
if (not has_prefix or crlf_cou >= 2):
break
rt = t.kit.processReferent("PERSON", t)
if (rt is not None):
break
if (ad is not None):
ad.overflow_level -= 1
return ReferentToken._new2329(p, begin, end, morph_, p._m_person_identity_typ)
def __try_parse(t: 'Token', prev: 'PersonIdToken') -> 'PersonIdToken':
if (t.is_value("СВИДЕТЕЛЬСТВО", None)):
tt1 = t
ip = False
reg = False
tt = t.next0_
first_pass3372 = True
while True:
if first_pass3372: first_pass3372 = False
else: tt = tt.next0_
if (not (tt is not None)): break
if (tt.is_comma_and or tt.morph.class0_.is_preposition):
continue
if (tt.is_value("РЕГИСТРАЦИЯ", None)
or tt.is_value("РЕЕСТР", None)
or tt.is_value("ЗАРЕГИСТРИРОВАТЬ", None)):
reg = True
tt1 = tt
elif (tt.is_value("ИНДИВИДУАЛЬНЫЙ", None)
or tt.is_value("ИП", None)):
ip = True
tt1 = tt
elif ((tt.is_value("ВНЕСЕНИЕ", None)
or tt.is_value("ГОСУДАРСТВЕННЫЙ", None)
or tt.is_value("ЕДИНЫЙ", None))
or tt.is_value("ЗАПИСЬ", None)
or tt.is_value("ПРЕДПРИНИМАТЕЛЬ", None)):
tt1 = tt
elif (tt.get_referent() is not None
and tt.get_referent().type_name == "DATERANGE"):
tt1 = tt
else:
break
if (reg and ip):
return PersonIdToken._new2505(
t, tt1, PersonIdToken.Typs.KEYWORD,
"СВИДЕТЕЛЬСТВО О ГОСУДАРСТВЕННОЙ РЕГИСТРАЦИИ ФИЗИЧЕСКОГО ЛИЦА В КАЧЕСТВЕ ИНДИВИДУАЛЬНОГО ПРЕДПРИНИМАТЕЛЯ"
)
tok = PersonIdToken.M_ONTOLOGY.try_parse(t, TerminParseAttr.NO)
if (tok is not None):
ty = Utils.valToEnum(tok.termin.tag, PersonIdToken.Typs)
res = PersonIdToken._new2505(tok.begin_token, tok.end_token, ty,
tok.termin.canonic_text)
if (prev is None):
if (ty != PersonIdToken.Typs.KEYWORD):
return None
t = tok.end_token.next0_
first_pass3373 = True
while True:
if first_pass3373: first_pass3373 = False
else: t = t.next0_
if (not (t is not None)): break
r = t.get_referent()
if (r is not None and (isinstance(r, GeoReferent))):
res.referent = r
res.end_token = t
continue
if (t.is_value("ГРАЖДАНИН", None) and t.next0_ is not None
and
(isinstance(t.next0_.get_referent(), GeoReferent))):
res.referent = t.next0_.get_referent()
res.end_token = t.next0_
t = res.end_token
continue
if (r is not None):
break
ait = PersonAttrToken.try_attach(
t, None, PersonAttrToken.PersonAttrAttachAttrs.NO)
if (ait is not None):
if (ait.referent is not None):
for s in ait.referent.slots:
if (s.type_name
== PersonPropertyReferent.ATTR_REF and
(isinstance(s.value, GeoReferent))):
res.referent = (Utils.asObjectOrNull(
s.value, Referent))
res.end_token = ait.end_token
break
if (t.is_value("ДАННЫЙ", None)):
res.end_token = t
continue
break
if ((isinstance(res.referent, GeoReferent))
and not res.referent.is_state):
res.referent = (None)
return res
if (ty == PersonIdToken.Typs.NUMBER):
tmp = io.StringIO()
tt = tok.end_token.next0_
if (tt is not None and tt.is_char(':')):
tt = tt.next0_
while tt is not None:
if (tt.is_newline_before):
break
if (not (isinstance(tt, NumberToken))):
break
print(tt.get_source_text(), end="", file=tmp)
res.end_token = tt
tt = tt.next0_
if (tmp.tell() < 1):
return None
res.value = Utils.toStringStringIO(tmp)
res.has_prefix = True
return res
if (ty == PersonIdToken.Typs.SERIA):
tmp = io.StringIO()
tt = tok.end_token.next0_
if (tt is not None and tt.is_char(':')):
tt = tt.next0_
next_num = False
first_pass3374 = True
while True:
if first_pass3374: first_pass3374 = False
else: tt = tt.next0_
if (not (tt is not None)): break
if (tt.is_newline_before):
break
if (MiscHelper.check_number_prefix(tt) is not None):
next_num = True
break
if (not (isinstance(tt, NumberToken))):
if (not (isinstance(tt, TextToken))):
break
if (not tt.chars.is_all_upper):
break
nu = NumberHelper.try_parse_roman(tt)
if (nu is not None):
print(nu.get_source_text(), end="", file=tmp)
tt = nu.end_token
elif (tt.length_char != 2):
break
else:
print(tt.term, end="", file=tmp)
res.end_token = tt
if (tt.next0_ is not None and tt.next0_.is_hiphen):
tt = tt.next0_
continue
if (tmp.tell() >= 4):
break
print(tt.get_source_text(), end="", file=tmp)
res.end_token = tt
if (tmp.tell() < 4):
if (tmp.tell() < 2):
return None
tt1 = res.end_token.next0_
if (tt1 is not None and tt1.is_comma):
tt1 = tt1.next0_
next0__ = PersonIdToken.__try_parse(tt1, res)
if (next0__ is not None
and next0__.typ == PersonIdToken.Typs.NUMBER):
pass
else:
return None
res.value = Utils.toStringStringIO(tmp)
res.has_prefix = True
return res
if (ty == PersonIdToken.Typs.CODE):
tt = res.end_token.next0_
first_pass3375 = True
while True:
if first_pass3375: first_pass3375 = False
else: tt = tt.next0_
if (not (tt is not None)): break
if (tt.is_char_of(":") or tt.is_hiphen):
continue
if (isinstance(tt, NumberToken)):
res.end_token = tt
continue
break
if (ty == PersonIdToken.Typs.ADDRESS):
if (isinstance(t.get_referent(), AddressReferent)):
res.referent = t.get_referent()
res.end_token = t
return res
tt = res.end_token.next0_
first_pass3376 = True
while True:
if first_pass3376: first_pass3376 = False
else: tt = tt.next0_
if (not (tt is not None)): break
if (tt.is_char_of(":") or tt.is_hiphen
or tt.morph.class0_.is_preposition):
continue
if (isinstance(tt.get_referent(), AddressReferent)):
res.referent = tt.get_referent()
res.end_token = tt
break
if (res.referent is None):
return None
return res
elif (prev is None):
return None
t0 = t
t1 = MiscHelper.check_number_prefix(t0)
if (t1 is not None):
t = t1
if (isinstance(t, NumberToken)):
tmp = io.StringIO()
res = PersonIdToken._new2507(t0, t, PersonIdToken.Typs.NUMBER)
tt = t
while tt is not None:
if (tt.is_newline_before or not (isinstance(tt, NumberToken))):
break
print(tt.get_source_text(), end="", file=tmp)
res.end_token = tt
tt = tt.next0_
if (tmp.tell() < 4):
if (tmp.tell() < 2):
return None
if (prev is None or prev.typ != PersonIdToken.Typs.KEYWORD):
return None
ne = PersonIdToken.__try_parse(res.end_token.next0_, prev)
if (ne is not None and ne.typ == PersonIdToken.Typs.NUMBER):
res.typ = PersonIdToken.Typs.SERIA
else:
return None
res.value = Utils.toStringStringIO(tmp)
if (t0 != t):
res.has_prefix = True
return res
if (isinstance(t, ReferentToken)):
r = t.get_referent()
if (r is not None):
if (r.type_name == "DATE"):
return PersonIdToken._new2508(t, t,
PersonIdToken.Typs.DATE, r)
if (r.type_name == "ORGANIZATION"):
return PersonIdToken._new2508(t, t, PersonIdToken.Typs.ORG,
r)
if (r.type_name == "ADDRESS"):
return PersonIdToken._new2508(t, t,
PersonIdToken.Typs.ADDRESS,
r)
if ((prev is not None and prev.typ == PersonIdToken.Typs.KEYWORD and
(isinstance(t, TextToken))) and not t.chars.is_all_lower
and t.chars.is_letter):
rr = PersonIdToken.__try_parse(t.next0_, prev)
if (rr is not None and rr.typ == PersonIdToken.Typs.NUMBER):
return PersonIdToken._new2505(t, t, PersonIdToken.Typs.SERIA,
t.term)
if ((t is not None and t.is_value("ОТ", "ВІД") and
(isinstance(t.next0_, ReferentToken)))
and t.next0_.get_referent().type_name == "DATE"):
return PersonIdToken._new2508(t, t.next0_, PersonIdToken.Typs.DATE,
t.next0_.get_referent())
return None
def __calc_rank_and_value(self, min_newlines_count: int) -> bool:
self.rank = 0
if (self.begin_token.chars.is_all_lower):
self.rank -= 30
words = 0
up_words = 0
notwords = 0
line_number = 0
tstart = self.begin_token
tend = self.end_token
t = self.begin_token
first_pass3396 = True
while True:
if first_pass3396: first_pass3396 = False
else: t = t.next0_
if (not (t != self.end_token.next0_ and t is not None
and t.end_char <= self.end_token.end_char)):
break
if (t.is_newline_before):
pass
tit = TitleItemToken.try_attach(t)
if (tit is not None):
if (tit.typ == TitleItemToken.Types.THEME
or tit.typ == TitleItemToken.Types.TYPANDTHEME):
if (t != self.begin_token):
if (line_number > 0):
return False
notwords = 0
up_words = notwords
words = up_words
tstart = tit.end_token.next0_
t = tit.end_token
if (t.next0_ is None):
return False
if (t.next0_.chars.is_letter
and t.next0_.chars.is_all_lower):
self.rank += 20
else:
self.rank += 100
tstart = t.next0_
if (tit.typ == TitleItemToken.Types.TYPANDTHEME):
self.type_value = tit.value
continue
if (tit.typ == TitleItemToken.Types.TYP):
if (t == self.begin_token):
if (tit.end_token.is_newline_after):
self.type_value = tit.value
self.rank += 5
tstart = tit.end_token.next0_
t = tit.end_token
words += 1
if (tit.begin_token != tit.end_token):
words += 1
if (tit.chars.is_all_upper):
up_words += 1
continue
if (tit.typ == TitleItemToken.Types.DUST
or tit.typ == TitleItemToken.Types.SPECIALITY):
if (t == self.begin_token):
return False
self.rank -= 20
if (tit.typ == TitleItemToken.Types.SPECIALITY):
self.speciality = tit.value
t = tit.end_token
continue
if (tit.typ == TitleItemToken.Types.CONSULTANT
or tit.typ == TitleItemToken.Types.BOSS
or tit.typ == TitleItemToken.Types.EDITOR):
t = tit.end_token
if (t.next0_ is not None and
((t.next0_.is_char_of(":") or t.next0_.is_hiphen
or t.whitespaces_after_count > 4))):
self.rank -= 10
else:
self.rank -= 2
continue
return False
blt = BookLinkToken.try_parse(t, 0)
if (blt is not None):
if (blt.typ == BookLinkTyp.MISC or blt.typ == BookLinkTyp.N
or blt.typ == BookLinkTyp.PAGES):
self.rank -= 10
elif (blt.typ == BookLinkTyp.N
or blt.typ == BookLinkTyp.PAGERANGE):
self.rank -= 20
if (t == self.begin_token and BookLinkToken.try_parse_author(
t, FioTemplateType.UNDEFINED) is not None):
self.rank -= 20
if (t.is_newline_before and t != self.begin_token):
line_number += 1
if (line_number > 4):
return False
if (t.chars.is_all_lower):
self.rank += 10
elif (t.previous.is_char('.')):
self.rank -= 10
elif (t.previous.is_char_of(",-")):
self.rank += 10
else:
npt = NounPhraseHelper.try_parse(t.previous,
NounPhraseParseAttr.NO, 0,
None)
if (npt is not None and npt.end_char >= t.end_char):
self.rank += 10
if (t != self.begin_token
and t.newlines_before_count > min_newlines_count):
self.rank -= (t.newlines_before_count - min_newlines_count)
bst = BracketHelper.try_parse(t, BracketParseAttr.NO, 100)
if (bst is not None and bst.is_quote_type
and bst.end_token.end_char <= self.end_token.end_char):
if (words == 0):
tstart = bst.begin_token
self.rank += 10
if (bst.end_token == self.end_token):
tend = self.end_token
self.rank += 10
rli = t.get_referents()
if (rli is not None):
for r in rli:
if (isinstance(r, OrganizationReferent)):
if (t.is_newline_before):
self.rank -= 10
else:
self.rank -= 4
continue
if ((isinstance(r, GeoReferent))
or (isinstance(r, PersonReferent))):
if (t.is_newline_before):
self.rank -= 5
if (t.is_newline_after or t.next0_ is None):
self.rank -= 20
elif (t.next0_.is_hiphen
or (isinstance(t.next0_, NumberToken))
or (isinstance(t.next0_.get_referent(),
DateReferent))):
self.rank -= 20
elif (t != self.begin_token):
self.rank -= 20
continue
if ((isinstance(r, GeoReferent))
or (isinstance(r, DenominationReferent))):
continue
if ((isinstance(r, UriReferent))
or (isinstance(r, PhoneReferent))):
return False
if (t.is_newline_before):
self.rank -= 4
else:
self.rank -= 2
if (t == self.begin_token and (isinstance(
self.end_token.get_referent(), PersonReferent))):
self.rank -= 10
words += 1
if (t.chars.is_all_upper):
up_words += 1
if (t == self.begin_token):
if (t.is_newline_after):
self.rank -= 10
elif (t.next0_ is not None and t.next0_.is_char('.')
and t.next0_.is_newline_after):
self.rank -= 10
continue
if (isinstance(t, NumberToken)):
if (t.typ == NumberSpellingType.WORDS):
words += 1
if (t.chars.is_all_upper):
up_words += 1
else:
notwords += 1
continue
pat = PersonAttrToken.try_attach(
t, None, PersonAttrToken.PersonAttrAttachAttrs.NO)
if (pat is not None):
if (t.is_newline_before):
if (not pat.morph.case_.is_undefined
and not pat.morph.case_.is_nominative):
pass
elif (pat.chars.is_all_upper):
pass
else:
self.rank -= 20
elif (t.chars.is_all_lower):
self.rank -= 1
while t is not None:
words += 1
if (t.chars.is_all_upper):
up_words += 1
if (t == pat.end_token):
break
t = t.next0_
continue
oitt = OrgItemTypeToken.try_attach(t, True, None)
if (oitt is not None):
if (oitt.morph.number != MorphNumber.PLURAL
and not oitt.is_doubt_root_word):
if (not oitt.morph.case_.is_undefined
and not oitt.morph.case_.is_nominative):
words += 1
if (t.chars.is_all_upper):
up_words += 1
else:
self.rank -= 4
if (t == self.begin_token):
self.rank -= 5
else:
words += 1
if (t.chars.is_all_upper):
up_words += 1
t = oitt.end_token
continue
tt = Utils.asObjectOrNull(t, TextToken)
if (tt is not None):
if (tt.is_char('©')):
self.rank -= 10
if (tt.is_char('_')):
self.rank -= 1
if (tt.chars.is_letter):
if (tt.length_char > 2):
words += 1
if (t.chars.is_all_upper):
up_words += 1
elif (not tt.is_char(',')):
notwords += 1
if (tt.is_pure_verb):
self.rank -= 30
words -= 1
break
if (tt == self.end_token):
if (tt.morph.class0_.is_preposition
or tt.morph.class0_.is_conjunction):
self.rank -= 10
elif (tt.is_char('.')):
self.rank += 5
elif (tt.is_char_of("._")):
self.rank -= 5
self.rank += words
self.rank -= notwords
if ((words < 1) and (self.rank < 50)):
return False
if (tstart is None or tend is None):
return False
if (tstart.end_char > tend.end_char):
return False
tit1 = TitleItemToken.try_attach(self.end_token.next0_)
if (tit1 is not None
and ((tit1.typ == TitleItemToken.Types.TYP
or tit1.typ == TitleItemToken.Types.SPECIALITY))):
if (tit1.end_token.is_newline_after):
self.rank += 15
else:
self.rank += 10
if (tit1.typ == TitleItemToken.Types.SPECIALITY):
self.speciality = tit1.value
if (up_words > 4 and up_words > (math.floor((0.8 * (words))))):
if (tstart.previous is not None and
(isinstance(tstart.previous.get_referent(), PersonReferent))):
self.rank += (5 + up_words)
self.begin_name_token = tstart
self.end_name_token = tend
return True
def kind(self) -> 'PersonPropertyKind':
""" Категория свойства """
from pullenti.ner.person.internal.PersonAttrToken import PersonAttrToken
return PersonAttrToken.check_kind(self)
def kind(self) -> 'PersonPropertyKind':
from pullenti.ner.person.internal.PersonAttrToken import PersonAttrToken
return PersonAttrToken.checkKind(self)