def test_allennlp_processor_with_existing_entries(self, overwrite_entries,
allow_parallel_entries):
config = {
'overwrite_entries': overwrite_entries,
'allow_parallel_entries': allow_parallel_entries
}
nlp = self._create_pipeline(config)
# Adding extra processor to have existing tokens and dependencies
nlp.add(component=AllenNLPProcessor(), config=config)
nlp.initialize()
if not overwrite_entries and not allow_parallel_entries:
# Processor should raise config error when both the flags are False
# and existing entries are found
with self.assertRaises(ProcessExecutionException):
nlp.process(self.document)
else:
pack = nlp.process(self.document)
processors = AllenNLPProcessor.default_configs()['processors']
tag_format = AllenNLPProcessor.default_configs()['tag_formalism']
if not overwrite_entries:
if allow_parallel_entries:
# Should raise AssertionError due to duplicate tokens
with self.assertRaises(AssertionError):
self._check_results(pack, processors, tag_format)
else:
self._check_results(pack, processors, tag_format)
def test_allennlp_processor_with_different_processors(self, processors):
nlp = self._create_pipeline({'processors': processors})
pack = nlp.process(self.document)
if processors == "":
processors = AllenNLPProcessor.default_configs()['processors']
tag_format = AllenNLPProcessor.default_configs()['tag_formalism']
self._check_results(pack, processors, tag_format)
def test_allennlp_processor_with_repeating_words(self, sentence):
processors = "tokenize"
nlp = self._create_pipeline({'processors': processors})
self.document = sentence
self.tokens = [sentence.replace('.', ' .').split()]
pack = nlp.process(self.document)
output_format = AllenNLPProcessor.default_configs()['output_format']
self._check_results(pack, processors, output_format)
def test_allennlp_processor_with_different_tag_formats(self, format):
if format == "random_dependencies":
with self.assertRaises(ProcessorConfigError):
self._create_pipeline({'tag_formalism': format})
else:
nlp = self._create_pipeline({'tag_formalism': format})
pack = nlp.process(self.document)
processors = AllenNLPProcessor.default_configs()['processors']
self._check_results(pack, processors, format)