class LongestMatchGlobalFeature(object):
def __init__(self, lookup_data, featname):
"""
Create a global feature function that adds 3 types of features:
1) B-featname - if current token starts an entity from
the ``lookup_data``;
2) I-featname - if current token is inside an entity from
the ``lookup_data``;
3) featname - if current token belongs to an entity from the
``lookup_data``.
"""
if hasattr(lookup_data, 'find_ranges'):
self.lm = lookup_data
else:
self.lm = LongestMatch(lookup_data)
self.b_featname = 'B-' + featname
self.i_featname = 'I-' + featname
self.featname = featname
def __call__(self, doc):
token_strings = [tok.token for tok, feat in doc]
for start, end, matched_text in self.lm.find_ranges(token_strings):
self.process_range(doc, start, end, matched_text)
def process_range(self, doc, start, end, matched_text):
doc[start][1][self.b_featname] = True
doc[start][1][self.featname] = True
for idx in range(start+1, end):
doc[idx][1][self.i_featname] = True
doc[idx][1][self.featname] = True
class LongestMatchGlobalFeature(object):
def __init__(self, lookup_data, featname):
"""
Create a global feature function that adds 3 types of features:
1) B-featname - if current token starts an entity from
the ``lookup_data``;
2) I-featname - if current token is inside an entity from
the ``lookup_data``;
3) featname - if current token belongs to an entity from the
``lookup_data``.
"""
if hasattr(lookup_data, 'find_ranges'):
self.lm = lookup_data
else:
self.lm = LongestMatch(lookup_data)
self.b_featname = 'B-' + featname
self.i_featname = 'I-' + featname
self.featname = featname
def __call__(self, doc):
token_strings = [tok.token for tok, feat in doc]
for start, end, matched_text in self.lm.find_ranges(token_strings):
self.process_range(doc, start, end, matched_text)
def process_range(self, doc, start, end, matched_text):
doc[start][1][self.b_featname] = True
doc[start][1][self.featname] = True
for idx in range(start + 1, end):
doc[idx][1][self.i_featname] = True
doc[idx][1][self.featname] = True
def __init__(self, lookup_data, featname):
"""
Create a global feature function that adds 3 types of features:
1) B-featname - if current token starts an entity from
the ``lookup_data``;
2) I-featname - if current token is inside an entity from
the ``lookup_data``;
3) featname - if current token belongs to an entity from the
``lookup_data``.
"""
if hasattr(lookup_data, 'find_ranges'):
self.lm = lookup_data
else:
self.lm = LongestMatch(lookup_data)
self.b_featname = 'B-' + featname
self.i_featname = 'I-' + featname
self.featname = featname
def __init__(self, lookup_data, featname):
"""
Create a global feature function that adds 3 types of features:
1) B-featname - if current token starts an entity from
the ``lookup_data``;
2) I-featname - if current token is inside an entity from
the ``lookup_data``;
3) featname - if current token belongs to an entity from the
``lookup_data``.
"""
if hasattr(lookup_data, 'find_ranges'):
self.lm = lookup_data
else:
self.lm = LongestMatch(lookup_data)
self.b_featname = 'B-' + featname
self.i_featname = 'I-' + featname
self.featname = featname
