def extract_tags(text):
text_clean = clean(text)
return sum(
[
[dict(tag) for tag in tagger.tag(tag_group, text_clean)]
for tag_group in pattern.tag_groups.itervalues()
],
[]
)
def check_taggers(taggers, input, expected):
tag_group = TagGroup(taggers, '')
tags = tag(tag_group, input)
if expected is not None:
assert tags is not None
if isinstance(tags, list):
assert len(tags) == 1
tags = tags[0]
for key, value in expected.iteritems():
assert tags[key] == value
else:
assert (tags is None or tags == [])
def test():
tags = tag(taggers, input)
if output is not None:
assert_true(tags is not None)
if isinstance(tags, list):
assert_equal(len(tags), 1)
tags = tags[0]
for key, value in output.iteritems():
assert_equal(
tags[key], value
)
else:
assert_true(tags is None or len(tags) == 0)
def tag(self, tag_groups=None, overwrite=False, save=True):
"""Add tags to article.
:param list tag_groups: List of TagGroup objects
:param bool overwrite: Overwrite existing tags
:param bool save: Save record after update
:return list: New or modified extracted tags
"""
tag_groups = tag_groups or pattern.tag_groups.values()
if overwrite:
self.tags = []
existing_tags = []
else:
existing_tags = [
tagger.Tag(tag)
for tag in self.tags
]
new_tags = []
self.verify(save=False)
for document_type in self.verified:
document_field = DOCUMENT_TYPES_TO_FIELDS[document_type]
document = getattr(self, document_field)
# Quit if document not set
if document is None:
continue
doc = document.read()
# Quit if document empty or fails verification
if not doc:
continue
# Clean document text
doc = clean(doc)
for tag_group in tag_groups:
# Extract tags
tags = tagger.tag(tag_group, doc)
for tag in tags:
# Build context documents
context_data = {document_type: tag['context']}
group_data = {document_type: tag['group']}
span_data = {document_type: tag['span']}
# Update existing tag with context
if tag in existing_tags:
idx = existing_tags.index(tag)
if document_type not in existing_tags[idx]['context']:
existing_tags[idx]['context'].update(context_data)
existing_tags[idx]['group'].update(group_data)
existing_tags[idx]['span'].update(span_data)
new_tags.append(existing_tags[idx])
# Create new tag in database
else:
tag['context'] = context_data
tag['group'] = group_data
tag['span'] = span_data
existing_tags.append(tag)
new_tags.append(tag)
# Cast tags to dictionaries for ODM compatibility
self.tags = [
dict(tag)
for tag in existing_tags
]
# Update tagged date
self.date_last_tagged = datetime.datetime.utcnow()
if save:
self.save()
return new_tags
def tag(self, tag_groups=None, overwrite=False, save=True):
"""Add tags to article.
:param list tag_groups: List of TagGroup objects
:param bool overwrite: Overwrite existing tags
:param bool save: Save record after update
:return list: New or modified extracted tags
"""
tag_groups = tag_groups or pattern.tag_groups.values()
if overwrite:
self.tags = []
existing_tags = []
else:
existing_tags = [tagger.Tag(tag) for tag in self.tags]
new_tags = []
self.verify(save=False)
for document_type in self.verified:
document_field = DOCUMENT_TYPES_TO_FIELDS[document_type]
document = getattr(self, document_field)
# Quit if document not set
if document is None:
continue
doc = document.read()
# Quit if document empty or fails verification
if not doc:
continue
# Clean document text
doc = clean(doc)
for tag_group in tag_groups:
# Extract tags
tags = tagger.tag(tag_group, doc)
for tag in tags:
# Build context documents
context_data = {document_type: tag['context']}
group_data = {document_type: tag['group']}
span_data = {document_type: tag['span']}
# Update existing tag with context
if tag in existing_tags:
idx = existing_tags.index(tag)
if document_type not in existing_tags[idx]['context']:
existing_tags[idx]['context'].update(context_data)
existing_tags[idx]['group'].update(group_data)
existing_tags[idx]['span'].update(span_data)
new_tags.append(existing_tags[idx])
# Create new tag in database
else:
tag['context'] = context_data
tag['group'] = group_data
tag['span'] = span_data
existing_tags.append(tag)
new_tags.append(tag)
# Cast tags to dictionaries for ODM compatibility
self.tags = [dict(tag) for tag in existing_tags]
# Update tagged date
self.date_last_tagged = datetime.datetime.utcnow()
if save:
self.save()
return new_tags
def extract_tags(text):
text_clean = clean(text)
return sum([[dict(tag) for tag in tagger.tag(tag_group, text_clean)]
for tag_group in pattern.tag_groups.itervalues()], [])
