def testCapitalization(self):
tests = []
tests.append( TestCase.formVerb(SKey.REG, 'DISmiss', 'Dismisses', 'Dismissed','Dismissing') )
tests[-1].forms[SKey.INFINATIVE] = 'Dismiss' # override InflTestHelper's so test is correct
tests.append( TestCase.formAdj(SKey.REG, 'Brainy', 'Brainier', 'Brainiest') )
tests.append( TestCase.formNoun(SKey.REG, 'FLY', 'FLIES',) )
for word in tests:
infl_dict = lemminflect.getAllInflectionsOOV(word.base, word.upos)
msg = '\n%s\nCorrect : %s\nFunction: %s' % (word.base, word.forms, infl_dict)
self.assertTrue(word.inflectionsInDict(infl_dict), msg)
def sample_verb(tag_list, source_tag, source):
tag_list = [tag for tag in tag_list if tag != source_tag]
tag = rd.choice(tag_list)
cand_list = getInflection(source, tag)
if cand_list == []:
cand_list = getAllInflectionsOOV(source, upos='VERB').values()
if len(cand_list) > 0:
cand = rd.choice(cand_list)
else:
cand = None
return cand
def testUPOSLog(self):
with self.assertLogs():
infl = lemminflect.getInflection('WORD', 'X')
self.assertEqual(infl, ())
with self.assertLogs():
infls = lemminflect.getAllInflections('WORD', 'X')
self.assertEqual(infls, {})
with self.assertLogs():
infls = lemminflect.getAllInflectionsOOV('WORD', 'X')
self.assertEqual(infls, {})
token = self.nlp('testing')[0]
self.assertEqual(token._.inflect('X'), 'testing')
def get_lemmas(word: str, pos: PartOfSpeech):
word = word.lower()
if (" " in word or "." in word):
return JSONResponse (status_code = 200, content = {"message": "Input must contain only a single word without spaces or punctuation."})
# Get the basic lemma version of the word first
lemmas = getLemma(word, pos)
if len(lemmas) > 0:
lemma = getLemma(word, pos)[0]
else:
lemma = word
inflections = merge_inflections(getAllInflections(lemma, upos=pos), getAllInflectionsOOV(lemma, upos=pos))
return {"lemma": lemma, "inflections": inflections}
def __call__(self, sent, index):
# get word
if sent[index].org is not None:
word = sent[index].org
else:
word = sent[index].lemma
# get cand
cand = None
source = word.lower()
if source != '':
source_tag = sent[index].tag
tag_list = [tag for tag in self.tag_list if tag != source_tag]
tag = rd.choice(tag_list)
cand_list = getInflection(source, tag)
if cand_list == []:
cand_list = getAllInflectionsOOV(source, upos='VERB').values()
if len(cand_list) > 0:
cand = rd.choice(cand_list)
# replace to cand
if cand is not None:
if word.istitle():
cand = cand.title()
sent[index].org = cand
if ((index >= 1 and sent[index - 1].pos != 'AUX')
and (index >= 2 and sent[index - 2].pos != 'AUX')
and self.sampler() < self.aux_ratio
): # 直前にAUXがなくVBG, VBNなら"have (been)"の変化を直前に挿入する
if tag == 'VBG':
sent[index].addition.append(
EnToken(index=sent[index].index - 0.25,
org=self.vbg_sampler()))
elif tag == 'VBN':
sent[index].addition.append(
EnToken(index=sent[index].index - 0.25,
org=self.vbn_sampler()))
sent[index] = self.add_history(sent[index])
return sent
def testUPosException(self):
self.assertEqual(lemminflect.getAllInflectionsOOV('test', 'X'), {})
def api_getAllInflectionsOOV():
content = request.json
result = getAllInflectionsOOV(content['lemma'], content['upos'])
return jsonify(result)