def test_lemmatize_unicode():
_check_corenlp()
lines = parse(u"\u0540\u0578\u057e\u056b\u056f hit C\xe9sar",
annotators=['tokenize', 'ssplit', 'pos', 'lemma'])
saf = stanford_to_saf(lines)
assert_equal({t['lemma'] for t in saf['tokens']},
{'Cesar', 'hit', 'Hovik'})
def test_lemmatize_unicode():
_check_corenlp()
lines = parse(u"H\xf6v\xedk hit C\xe9sar",
annotators=['tokenize', 'ssplit', 'pos', 'lemma'])
saf = stanford_to_saf(lines)
assert_equal({t['lemma'] for t in saf['tokens']},
{'Cesar', 'hit', 'Hovik'})
def test_ner():
_check_corenlp()
annotators = ['tokenize', 'ssplit', 'pos', 'lemma', 'ner']
saf = stanford_to_saf(parse("John lives in Amsterdam",
annotators=annotators))
lemmata = {t['id']: t['lemma'] for t in saf['tokens']}
entities = {lemmata[e['tokens'][0]]: e['type'] for e in saf['entities']}
assert_equal(entities, {'John': 'PERSON', 'Amsterdam': 'LOCATION'})
def test_ner():
_check_corenlp()
annotators = ['tokenize', 'ssplit', 'pos', 'lemma', 'ner']
saf = stanford_to_saf(parse("John lives in Amsterdam",
annotators=annotators))
lemmata = {t['id']: t['lemma'] for t in saf['tokens']}
entities = {lemmata[e['tokens'][0]]: e['type'] for e in saf['entities']}
assert_equal(entities, {'John': 'PERSON', 'Amsterdam': 'LOCATION'})
def test_raw():
_check_corenlp()
lines = parse("It. Works\n", annotators=['tokenize', 'ssplit'])
expected = ['Sentence #1 (2 tokens):', 'It.',
'[Text=It CharacterOffsetBegin=0 CharacterOffsetEnd=2]'
' [Text=. CharacterOffsetBegin=2 CharacterOffsetEnd=3]',
'Sentence #2 (1 tokens):', 'Works',
'[Text=Works CharacterOffsetBegin=4 CharacterOffsetEnd=9]']
assert_equal(lines, expected)
def test_lemmatize():
_check_corenlp()
lines = parse("He jumped. \n\n Cool!",
annotators=['tokenize', 'ssplit', 'pos', 'lemma'])
saf = stanford_to_saf(lines)
assert_equal(set(saf.keys()), {'tokens', 'header'})
assert_equal({t['lemma']
for t in saf['tokens']}, {'he', 'jump', 'cool', '!', '.'})
assert_equal({t['sentence'] for t in saf['tokens']}, {1, 2})
def test_lemmatize():
_check_corenlp()
lines = parse("He jumped. \n\n Cool!",
annotators=['tokenize', 'ssplit', 'pos', 'lemma'])
saf = stanford_to_saf(lines)
assert_equal(set(saf.keys()), {'tokens', 'header'})
assert_equal({t['lemma'] for t in saf['tokens']},
{'he', 'jump', 'cool', '!', '.'})
assert_equal({t['sentence'] for t in saf['tokens']},
{1, 2})
def test_raw():
_check_corenlp()
lines = parse("It. Works\n", annotators=['tokenize', 'ssplit'])
expected = [
'Sentence #1 (2 tokens):', 'It.',
'[Text=It CharacterOffsetBegin=0 CharacterOffsetEnd=2]'
' [Text=. CharacterOffsetBegin=2 CharacterOffsetEnd=3]',
'Sentence #2 (1 tokens):', 'Works',
'[Text=Works CharacterOffsetBegin=4 CharacterOffsetEnd=9]'
]
assert_equal(lines, expected)
def test_parse():
_check_corenlp()
saf = stanford_to_saf(parse("John loves himself"))
lemmata = {t['id']: t['lemma'] for t in saf['tokens']}
assert_equal(saf['trees'], [{
"tree": "(ROOT (S (NP (NNP John)) (VP (VBZ loves) "
"(NP (PRP himself)))))",
"sentence": 1
}])
deps = {(lemmata[d['child']], lemmata[d['parent']], d['relation'])
for d in saf['dependencies']}
assert_equal(deps, {('John', 'love', 'nsubj'),
('himself', 'love', 'dobj')})
corefs = {tuple(sorted([lemmata[c[0][0]], lemmata[c[1][0]]]))
for c in saf['coreferences']}
assert_equal(corefs, {tuple(sorted(['John', 'himself']))})
def test_parse():
_check_corenlp()
saf = stanford_to_saf(parse("John loves himself"))
lemmata = {t['id']: t['lemma'] for t in saf['tokens']}
assert_equal(saf['trees'], [{
"tree": "(ROOT (S (NP (NNP John)) (VP (VBZ loves) "
"(NP (PRP himself)))))",
"sentence": 1
}])
deps = {(lemmata[d['child']], lemmata[d['parent']], d['relation'])
for d in saf['dependencies']}
assert_equal(deps, {('John', 'love', 'nsubj'),
('himself', 'love', 'dobj')})
corefs = {tuple(sorted([lemmata[c[0][0]], lemmata[c[1][0]]]))
for c in saf['coreferences']}
assert_equal(corefs, {tuple(sorted(['John', 'himself']))})
def test_multiple_sentences():
_check_corenlp()
p = parse("John lives in Amsterdam. He works in London")
saf = stanford_to_saf(p)
tokens = {t['id'] : t for t in saf['tokens']}
# are token ids unique?
assert_equal(len(tokens), len(saf['tokens']))
# is location in second sentence correct?
entities = {tokens[e['tokens'][0]]['lemma'] : e['type']
for e in saf['entities']}
assert_in(('London', 'LOCATION'), entities.items())
# is dependency in second sentence correct?
rels = [(tokens[rel['child']]['lemma'], rel['relation'], tokens[rel['parent']]['lemma'])
for rel in saf['dependencies']]
assert_in(("he", "nsubj", "work"), rels)
assert_in(("John", "nsubj", "live"), rels)
# is coref parsed correctly?
coref = {(tokens[x[0][0]]['lemma'], tokens[x[1][0]]['lemma']) for x in saf['coreferences']}
assert_equal(coref, {("John", "he")})
def test_multiple_sentences():
_check_corenlp()
p = parse("John lives in Amsterdam. He works in London")
saf = stanford_to_saf(p)
tokens = {t['id']: t for t in saf['tokens']}
# are token ids unique?
assert_equal(len(tokens), len(saf['tokens']))
# is location in second sentence correct?
entities = {tokens[e['tokens'][0]]['lemma']: e['type']
for e in saf['entities']}
assert_in(('London', 'LOCATION'), entities.items())
# is dependency in second sentence correct?
rels = [(tokens[rel['child']]['lemma'], rel['relation'],
tokens[rel['parent']]['lemma'])
for rel in saf['dependencies']]
assert_in(("he", "nsubj", "work"), rels)
assert_in(("John", "nsubj", "live"), rels)
# is coref parsed correctly?
coref = {(tokens[x[0][0]]['lemma'], tokens[x[1][0]]['lemma'])
for x in saf['coreferences']}
assert_equal(coref, {("John", "he")})