cothec / Corneferencer

Commit 094a0f2f0dd6441064cfbaebcfb6ce935a4a5c3f

Authored by Bartłomiej Nitoń
1 parent d5a75fe6

Fixes for unknown heads.

Inline Side-by-side
Showing 1 changed file with 10 additions and 7 deletions
corneferencer/resolvers/features.py
  View file @094a0f2
... ... @@ -71,7 +71,9 @@ def sentence_vec(mention):
71 71  
72 72 def mention_type(mention):
73 73 type_vec = [0] * 4
74   - if mention.head['ctag'] in constants.NOUN_TAGS:
  74 + if mention.head is None:
  75 + type_vec[3] = 1
  76 + elif mention.head['ctag'] in constants.NOUN_TAGS:
75 77 type_vec[0] = 1
76 78 elif mention.head['ctag'] in constants.PPRON_TAGS:
77 79 type_vec[1] = 1
... ... @@ -145,7 +147,8 @@ def ante_contains_rarest_from_ana(ante, ana):
145 147  
146 148 def agreement(ante, ana, tag_name):
147 149 agr_vec = [0] * 3
148   - if ante.head[tag_name] == 'unk' or ana.head[tag_name] == 'unk':
  150 + if (ante.head[tag_name] == 'unk' or ana.head[tag_name] == 'unk'
  151 + or ante.head is None or ana.head is None):
149 152 agr_vec[2] = 1
150 153 elif ante.head[tag_name] == ana.head[tag_name]:
151 154 agr_vec[0] = 1
... ... @@ -158,8 +161,8 @@ def is_acronym(ante, ana):
158 161 if ana.text.upper() == ana.text:
159 162 return check_one_way_acronym(ana.text, ante.text)
160 163 if ante.text.upper() == ante.text:
161   - return check_one_way_acronym(ante.text, ana.text);
162   - return 0;
  164 + return check_one_way_acronym(ante.text, ana.text)
  165 + return 0
163 166  
164 167  
165 168 def same_sentence(ante, ana):
... ... @@ -232,7 +235,7 @@ def check_one_way_acronym(acronym, expression):
232 235 for expr2 in expr1.split():
233 236 expr2 = expr2.strip()
234 237 if expr2:
235   - initials += unicode(expr2[0]).upper()
  238 + initials += expr2[0].upper()
236 239 if acronym == initials:
237   - return 1;
238   - return 0;
  240 + return 1
  241 + return 0
... ...