def extract_verb_case(dg):
# extract the list of verb(= yogen) and cases
# list up all the verb nodes
for node in dg.nodelist:
util.log( jpdep.node2str(node) )
if is_yogen(node):
util.log( " <-- detected yogen: %s" % normalize_yogen(node) )
# if yogen, list up all its dependees
for case_cand in [dg.nodelist[i] for i in node['deps']]:
if is_case(case_cand):
util.log( " <-- detected case: %s" % jpdep.node2str(case_cand) )
util.log("")
def is_case(node):
bhead_tag = node['tag'][node['bhead']]
bform_tag = node['tag'][node['bform']]
bform_surface = bform_tag[-1]
if bform_tag[0] == u"助詞" and bform_tag[1] == u"格助詞" and (bform_surface in [u"ガ", u"ヲ", u"ニ", u"ト", u"デ", u"カラ", u"ヨリ", u"ヘ", u"マデ"]):
return True
elif bhead_tag[0] == u"名詞" and bform_tag[0:2] == [u"名詞", u"接尾"]:
print "=== detected noun+suffix: %s" % jpdep.node2str(node)
return True
else:
return False