def test_srl_reader_can_filter_by_domain(self):
conll_reader = SrlReader(domain_identifier="subdomain2")
instances = conll_reader.read('tests/fixtures/conll_2012/')
instances = ensure_list(instances)
# If we'd included the folder, we'd have 9 instances.
assert len(instances) == 2
def test_srl_reader_can_filter_by_domain(self):
conll_reader = SrlReader(domain_identifier="subdomain2")
instances = conll_reader.read(AllenNlpTestCase.FIXTURES_ROOT / 'conll_2012')
instances = ensure_list(instances)
# If we'd included the folder, we'd have 9 instances.
assert len(instances) == 2
def test_read_from_file(self):
conll_reader = SrlReader()
dataset = conll_reader.read('tests/fixtures/conll_2012/')
instances = dataset.instances
fields = instances[0].fields()
assert fields["tokens"].tokens() == ["Mali", "government", "officials", "say",
"the", "woman", "'s", "confession", "was", "forced", "."]
assert fields["verb_indicator"].sequence_index() == 3
assert fields["tags"].tags() == ['B-ARG0', 'I-ARG0', 'I-ARG0', 'B-V', 'B-ARG1',
'I-ARG1', 'I-ARG1', 'I-ARG1', 'I-ARG1', 'I-ARG1', 'O']
fields = instances[1].fields()
assert fields["tokens"].tokens() == ["Mali", "government", "officials", "say",
"the", "woman", "'s", "confession", "was", "forced", "."]
assert fields["verb_indicator"].sequence_index() == 8
assert fields["tags"].tags() == ['O', 'O', 'O', 'O', 'B-ARG1', 'I-ARG1',
'I-ARG1', 'I-ARG1', 'B-V', 'B-ARG2', 'O']
fields = instances[2].fields()
assert fields["tokens"].tokens() == ['The', 'prosecution', 'rested', 'its', 'case', 'last', 'month',
'after', 'four', 'months', 'of', 'hearings', '.']
assert fields["verb_indicator"].sequence_index() == 2
assert fields["tags"].tags() == ['B-ARG0', 'I-ARG0', 'B-V', 'B-ARG1', 'I-ARG1', 'B-ARGM-TMP',
'I-ARGM-TMP', 'B-ARGM-TMP', 'I-ARGM-TMP', 'I-ARGM-TMP',
'I-ARGM-TMP', 'I-ARGM-TMP', 'O']
fields = instances[3].fields()
assert fields["tokens"].tokens() == ['The', 'prosecution', 'rested', 'its', 'case', 'last', 'month',
'after', 'four', 'months', 'of', 'hearings', '.']
assert fields["verb_indicator"].sequence_index() == 11
assert fields["tags"].tags() == ['O', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'B-V', 'O']
# Tests a sentence with no verbal predicates.
fields = instances[4].fields()
assert fields["tokens"].tokens() == ["Denise", "Dillon", "Headline", "News", "."]
assert fields["verb_indicator"].sequence_index() is None
assert fields["tags"].tags() == ['O', 'O', 'O', 'O', 'O']
def test_srl_reader_can_filter_by_domain(self):
conll_reader = SrlReader(domain_identifier="subdomain2")
instances = conll_reader.read('tests/fixtures/conll_2012/')
instances = ensure_list(instances)
# If we'd included the folder, we'd have 9 instances.
assert len(instances) == 2
def test_read_from_file(self, lazy):
conll_reader = SrlReader(lazy=lazy)
instances = conll_reader.read('tests/fixtures/conll_2012/subdomain')
instances = ensure_list(instances)
fields = instances[0].fields
tokens = [t.text for t in fields['tokens'].tokens]
assert tokens == [
"Mali", "government", "officials", "say", "the", "woman", "'s",
"confession", "was", "forced", "."
]
assert fields["verb_indicator"].labels[3] == 1
assert fields["tags"].labels == [
'B-ARG0', 'I-ARG0', 'I-ARG0', 'B-V', 'B-ARG1', 'I-ARG1', 'I-ARG1',
'I-ARG1', 'I-ARG1', 'I-ARG1', 'O'
]
fields = instances[1].fields
tokens = [t.text for t in fields['tokens'].tokens]
assert tokens == [
"Mali", "government", "officials", "say", "the", "woman", "'s",
"confession", "was", "forced", "."
]
assert fields["verb_indicator"].labels[8] == 1
assert fields["tags"].labels == [
'O', 'O', 'O', 'O', 'B-ARG1', 'I-ARG1', 'I-ARG1', 'I-ARG1', 'B-V',
'B-ARG2', 'O'
]
fields = instances[2].fields
tokens = [t.text for t in fields['tokens'].tokens]
assert tokens == [
'The', 'prosecution', 'rested', 'its', 'case', 'last', 'month',
'after', 'four', 'months', 'of', 'hearings', '.'
]
assert fields["verb_indicator"].labels[2] == 1
assert fields["tags"].labels == [
'B-ARG0', 'I-ARG0', 'B-V', 'B-ARG1', 'I-ARG1', 'B-ARGM-TMP',
'I-ARGM-TMP', 'B-ARGM-TMP', 'I-ARGM-TMP', 'I-ARGM-TMP',
'I-ARGM-TMP', 'I-ARGM-TMP', 'O'
]
fields = instances[3].fields
tokens = [t.text for t in fields['tokens'].tokens]
assert tokens == [
'The', 'prosecution', 'rested', 'its', 'case', 'last', 'month',
'after', 'four', 'months', 'of', 'hearings', '.'
]
assert fields["verb_indicator"].labels[11] == 1
assert fields["tags"].labels == [
'O', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'B-V', 'O'
]
# Tests a sentence with no verbal predicates.
fields = instances[4].fields
tokens = [t.text for t in fields['tokens'].tokens]
assert tokens == ["Denise", "Dillon", "Headline", "News", "."]
assert fields["verb_indicator"].labels == [0, 0, 0, 0, 0]
assert fields["tags"].labels == ['O', 'O', 'O', 'O', 'O']
def test_read_from_file(self, lazy):
conll_reader = SrlReader(lazy=lazy)
instances = conll_reader.read(AllenNlpTestCase.FIXTURES_ROOT / 'conll_2012' / 'subdomain')
instances = ensure_list(instances)
fields = instances[0].fields
tokens = [t.text for t in fields['tokens'].tokens]
assert tokens == ["Mali", "government", "officials", "say", "the", "woman", "'s",
"confession", "was", "forced", "."]
assert fields["verb_indicator"].labels[3] == 1
assert fields["tags"].labels == ['B-ARG0', 'I-ARG0', 'I-ARG0', 'B-V', 'B-ARG1',
'I-ARG1', 'I-ARG1', 'I-ARG1', 'I-ARG1', 'I-ARG1', 'O']
fields = instances[1].fields
tokens = [t.text for t in fields['tokens'].tokens]
assert tokens == ["Mali", "government", "officials", "say", "the", "woman", "'s",
"confession", "was", "forced", "."]
assert fields["verb_indicator"].labels[8] == 1
assert fields["tags"].labels == ['O', 'O', 'O', 'O', 'B-ARG1', 'I-ARG1',
'I-ARG1', 'I-ARG1', 'B-V', 'B-ARG2', 'O']
fields = instances[2].fields
tokens = [t.text for t in fields['tokens'].tokens]
assert tokens == ['The', 'prosecution', 'rested', 'its', 'case', 'last', 'month', 'after',
'four', 'months', 'of', 'hearings', '.']
assert fields["verb_indicator"].labels[2] == 1
assert fields["tags"].labels == ['B-ARG0', 'I-ARG0', 'B-V', 'B-ARG1', 'I-ARG1', 'B-ARGM-TMP',
'I-ARGM-TMP', 'B-ARGM-TMP', 'I-ARGM-TMP', 'I-ARGM-TMP',
'I-ARGM-TMP', 'I-ARGM-TMP', 'O']
fields = instances[3].fields
tokens = [t.text for t in fields['tokens'].tokens]
assert tokens == ['The', 'prosecution', 'rested', 'its', 'case', 'last', 'month', 'after',
'four', 'months', 'of', 'hearings', '.']
assert fields["verb_indicator"].labels[11] == 1
assert fields["tags"].labels == ['O', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'B-V', 'O']
# Tests a sentence with no verbal predicates.
fields = instances[4].fields
tokens = [t.text for t in fields['tokens'].tokens]
assert tokens == ["Denise", "Dillon", "Headline", "News", "."]
assert fields["verb_indicator"].labels == [0, 0, 0, 0, 0]
assert fields["tags"].labels == ['O', 'O', 'O', 'O', 'O']
def test_read_from_file(self, lazy):
conll_reader = SrlReader(lazy=lazy)
instances = conll_reader.read(AllenNlpTestCase.FIXTURES_ROOT /
u'conll_2012' / u'subdomain')
instances = ensure_list(instances)
fields = instances[0].fields
tokens = [t.text for t in fields[u'tokens'].tokens]
assert tokens == [
u"Mali", u"government", u"officials", u"say", u"the", u"woman",
u"'s", u"confession", u"was", u"forced", u"."
]
assert fields[u"verb_indicator"].labels[3] == 1
assert fields[u"tags"].labels == [
u'B-ARG0', u'I-ARG0', u'I-ARG0', u'B-V', u'B-ARG1', u'I-ARG1',
u'I-ARG1', u'I-ARG1', u'I-ARG1', u'I-ARG1', u'O'
]
fields = instances[1].fields
tokens = [t.text for t in fields[u'tokens'].tokens]
assert tokens == [
u"Mali", u"government", u"officials", u"say", u"the", u"woman",
u"'s", u"confession", u"was", u"forced", u"."
]
assert fields[u"verb_indicator"].labels[8] == 1
assert fields[u"tags"].labels == [
u'O', u'O', u'O', u'O', u'B-ARG1', u'I-ARG1', u'I-ARG1', u'I-ARG1',
u'B-V', u'B-ARG2', u'O'
]
fields = instances[2].fields
tokens = [t.text for t in fields[u'tokens'].tokens]
assert tokens == [
u'The', u'prosecution', u'rested', u'its', u'case', u'last',
u'month', u'after', u'four', u'months', u'of', u'hearings', u'.'
]
assert fields[u"verb_indicator"].labels[2] == 1
assert fields[u"tags"].labels == [
u'B-ARG0', u'I-ARG0', u'B-V', u'B-ARG1', u'I-ARG1', u'B-ARGM-TMP',
u'I-ARGM-TMP', u'B-ARGM-TMP', u'I-ARGM-TMP', u'I-ARGM-TMP',
u'I-ARGM-TMP', u'I-ARGM-TMP', u'O'
]
fields = instances[3].fields
tokens = [t.text for t in fields[u'tokens'].tokens]
assert tokens == [
u'The', u'prosecution', u'rested', u'its', u'case', u'last',
u'month', u'after', u'four', u'months', u'of', u'hearings', u'.'
]
assert fields[u"verb_indicator"].labels[11] == 1
assert fields[u"tags"].labels == [
u'O', u'O', u'O', u'O', u'O', u'O', u'O', u'O', u'O', u'O', u'O',
u'B-V', u'O'
]
# Tests a sentence with no verbal predicates.
fields = instances[4].fields
tokens = [t.text for t in fields[u'tokens'].tokens]
assert tokens == [u"Denise", u"Dillon", u"Headline", u"News", u"."]
assert fields[u"verb_indicator"].labels == [0, 0, 0, 0, 0]
assert fields[u"tags"].labels == [u'O', u'O', u'O', u'O', u'O']
def test_read_from_file(self, lazy):
conll_reader = SrlReader(lazy=lazy)
instances = conll_reader.read(AllenNlpTestCase.FIXTURES_ROOT /
"conll_2012" / "subdomain")
instances = ensure_list(instances)
fields = instances[0].fields
tokens = [t.text for t in fields["tokens"].tokens]
assert tokens == [
"Mali",
"government",
"officials",
"say",
"the",
"woman",
"'s",
"confession",
"was",
"forced",
".",
]
assert fields["verb_indicator"].labels[3] == 1
assert fields["tags"].labels == [
"B-ARG0",
"I-ARG0",
"I-ARG0",
"B-V",
"B-ARG1",
"I-ARG1",
"I-ARG1",
"I-ARG1",
"I-ARG1",
"I-ARG1",
"O",
]
assert fields["metadata"].metadata["words"] == tokens
assert fields["metadata"].metadata["verb"] == tokens[3]
assert fields["metadata"].metadata["gold_tags"] == fields[
"tags"].labels
fields = instances[1].fields
tokens = [t.text for t in fields["tokens"].tokens]
assert tokens == [
"Mali",
"government",
"officials",
"say",
"the",
"woman",
"'s",
"confession",
"was",
"forced",
".",
]
assert fields["verb_indicator"].labels[8] == 1
assert fields["tags"].labels == [
"O",
"O",
"O",
"O",
"B-ARG1",
"I-ARG1",
"I-ARG1",
"I-ARG1",
"B-V",
"B-ARG2",
"O",
]
assert fields["metadata"].metadata["words"] == tokens
assert fields["metadata"].metadata["verb"] == tokens[8]
assert fields["metadata"].metadata["gold_tags"] == fields[
"tags"].labels
fields = instances[2].fields
tokens = [t.text for t in fields["tokens"].tokens]
assert tokens == [
"The",
"prosecution",
"rested",
"its",
"case",
"last",
"month",
"after",
"four",
"months",
"of",
"hearings",
".",
]
assert fields["verb_indicator"].labels[2] == 1
assert fields["tags"].labels == [
"B-ARG0",
"I-ARG0",
"B-V",
"B-ARG1",
"I-ARG1",
"B-ARGM-TMP",
"I-ARGM-TMP",
"B-ARGM-TMP",
"I-ARGM-TMP",
"I-ARGM-TMP",
"I-ARGM-TMP",
"I-ARGM-TMP",
"O",
]
assert fields["metadata"].metadata["words"] == tokens
assert fields["metadata"].metadata["verb"] == tokens[2]
assert fields["metadata"].metadata["gold_tags"] == fields[
"tags"].labels
fields = instances[3].fields
tokens = [t.text for t in fields["tokens"].tokens]
assert tokens == [
"The",
"prosecution",
"rested",
"its",
"case",
"last",
"month",
"after",
"four",
"months",
"of",
"hearings",
".",
]
assert fields["verb_indicator"].labels[11] == 1
assert fields["tags"].labels == [
"O",
"O",
"O",
"O",
"O",
"O",
"O",
"O",
"O",
"O",
"O",
"B-V",
"O",
]
assert fields["metadata"].metadata["words"] == tokens
assert fields["metadata"].metadata["verb"] == tokens[11]
assert fields["metadata"].metadata["gold_tags"] == fields[
"tags"].labels
# Tests a sentence with no verbal predicates.
fields = instances[4].fields
tokens = [t.text for t in fields["tokens"].tokens]
assert tokens == ["Denise", "Dillon", "Headline", "News", "."]
assert fields["verb_indicator"].labels == [0, 0, 0, 0, 0]
assert fields["tags"].labels == ["O", "O", "O", "O", "O"]
assert fields["metadata"].metadata["words"] == tokens
assert fields["metadata"].metadata["verb"] is None
assert fields["metadata"].metadata["gold_tags"] == fields[
"tags"].labels