def answer_question(input_question):
en_doc = nlp(u'' + input_question)
question_class = classify_question(en_doc)
print("Class:", question_class)
question_keywords = extract_features(question_class, en_doc)
print("Question Features:", question_keywords)
question_query = construct_query(question_keywords, en_doc)
print("Question Query:", question_query)
print("Fetching Knowledge source...")
wiki_pages = fetch_wiki(question_keywords, number_of_search=3)
print("Pages Fetched:", len(wiki_pages))
# Anaphora Resolution
ranked_wiki_docs = rank_docs(question_keywords)
print("Ranked Pages:", ranked_wiki_docs)
candidate_answers, split_keywords = get_candidate_answers(
question_query, ranked_wiki_docs, nlp)
print("Candidate Answer:", "(" + str(len(candidate_answers)) + ")",
candidate_answers)
print("Answer:", " ".join(candidate_answers))
answer = " ".join(candidate_answers)
return answer
def answer_question(q, num_sentences):
q = nlp(u'' + q)
question_class = classify_question(q)
_logger.info("Question Class: {}".format(question_class))
question_keywords = extract_features(question_class, q)
_logger.debug("Question Features: {}".format(question_keywords))
query = construct_query(question_keywords, q)
_logger.debug("Query: {}".format(query))
_logger.info("Retrieving {} wikipedia pages...".format(num_sentences))
wiki_pages = fetch_wiki(question_keywords, number_of_search=num_sentences)
_logger.debug("Pages retrieved: {}".format(len(wiki_pages)))
# Anaphora Resolution
ranked_wiki_docs = rank_docs(question_keywords)
_logger.debug("Ranked pages: {}".format(ranked_wiki_docs))
candidate_answers, keywords = get_candidate_answers(query, ranked_wiki_docs, nlp)
_logger.info("Candidate answers ({}):\n{}".format(len(candidate_answers), '\n'.join(candidate_answers)))
return " ".join(candidate_answers)
def answer_question(input_question):
warnings.warn("This method is now deprecated.", DeprecationWarning)
en_nlp = spacy.load('en_core_web_md')
en_doc = en_nlp(u'' + input_question)
question_class = classify_question(en_doc)
print("Class:", question_class)
question_keywords = extract_features(question_class, en_doc)
print("Question Features:", question_keywords)
question_query = construct_query(question_keywords, en_doc)
print("Question Query:", question_query)
print("Fetching Knowledge source...")
wiki_pages = fetch_wiki(question_keywords, number_of_search=3)
print("Pages Fetched:", len(wiki_pages))
# Anaphora Resolution
ranked_wiki_docs = rank_docs(question_keywords)
print("Ranked Pages:", ranked_wiki_docs)
candidate_answers, split_keywords = get_candidate_answers(
question_query, ranked_wiki_docs, en_nlp)
print("Candidate Answer:", "(" + str(len(candidate_answers)) + ")",
candidate_answers)
print("Answer:", " ".join(candidate_answers))
answer = " ".join(candidate_answers)
return answer
def process_question(self):
self.question_class = classify_question(self.question_doc)
_logger.info("Question Class: {}".format(self.question_class))
self.question_keywords = extract_features(self.question_class,
self.question_doc)
_logger.info("Question Features: {}".format(self.question_keywords))
self.query = construct_query(self.question_keywords, self.question_doc)
_logger.info("Query: {}".format(self.query))
def process_question(self, dfOut):
self.question_class = classify_question(self.question_doc)
_logger.info("Question Class: {}".format(self.question_class))
temp = self.question_class.tostring()
dfentry[0] = "" + self.question_class
self.question_keywords = extract_features(self.question_class,
self.question_doc)
_logger.info("Question Features: {}".format(self.question_keywords))
dfentry[1] = ','.join(self.question_keywords)
self.query = construct_query(self.question_keywords, self.question_doc)
_logger.info("Query: {}".format(self.query))
dfentry[2] = "{}".format(self.query)
insert(dfOut, dfentry)
def test_extract_features(self):
sql_man = SqLiteManager()
en_nlp_l = spacy.load(EN_MODEL_MD)
result = sql_man.get_questions_between(5, 7)
for row in result:
qid = row[0]
with self.subTest(qid):
question = row[1]
question_type = row[2]
en_doc = en_nlp_l(u'' + question)
features = extract_features(question_type, en_doc, True)
print("{0}){1} :\nExtracted: {2}".format(qid, question, features))
js_feat = json.dumps(features)
sql_man.update_feature(qid, js_feat)
assert features is not None