def sentence_has_context(sentence) -> bool:
if type(sentence) == spacy.tokens.Span:
nlp = NLPConn.get_nlp_conn()
sentence = nlp(sentence.string)
sent_subjs = [t for t in sentence if t.dep_ == 'nsubj']
return any([subj for subj in sent_subjs if subj.pos_ == "PROPN"])
def get_has_contradictatory(sentence) -> int:
if type(sentence) == spacy.tokens.Span:
nlp = NLPConn.get_nlp_conn()
sentence = nlp(sentence.string)
contradictatory_matcher = ContradictatoryMatcher.get_contradicatory_matcher(
)
return int(any(contradictatory_matcher(sentence)))
def get_contradicatory_matcher():
global matcher
if matcher is None:
nlp = NLPConn.get_nlp_conn()
matcher = PhraseMatcher(nlp.vocab)
patterns = [nlp.make_doc(phrase) for phrase in contradictatory_phrases]
matcher.add("ContradictoryTokens", None, *patterns)
return matcher
=================
"""
import os
import spacy
from spacy.tokens import Span, Token
from trivia_generator.web_scraper import Article
from trivia_generator.web_scraper.WebScraper import get_page_by_random
from trivia_generator.TUnit import TUnit
from nlp_helpers import features
from nlp_helpers import NLPConn
from nlp_helpers import ContradictatoryMatcher
nlp = NLPConn.get_nlp_conn()
contradictatory_matcher = ContradictatoryMatcher.get_contradicatory_matcher()
def create_TUnits(article: Article) -> list:
"""Creates a list of TUnits from a Wikipedia article object.
:param article: A Wikipedia article object.
:type article: Article
:returns: a list of TUnits created from article.
"""
paragraphs = ' '.join(
[para for para in article.content.splitlines() if para != ''])
tunits = []
def resolve_coreferences(content: str) -> str:
nlp = NLPConn.get_nlp_conn()
doc = nlp(content)
resolved_coref = get_resolved(doc, doc._.coref_clusters)
return resolved_coref