Lemmatizer honors exceptions: Fix #1387

This commit is contained in:
Orion Montoya 2017-10-05 10:49:02 -04:00
parent e81a608173
commit ffb50d21a0

View File

@ -78,15 +78,16 @@ def lemmatize(string, index, exceptions, rules):
# forms.append(string) # forms.append(string)
forms.extend(exceptions.get(string, [])) forms.extend(exceptions.get(string, []))
oov_forms = [] oov_forms = []
for old, new in rules: if not forms:
if string.endswith(old): for old, new in rules:
form = string[:len(string) - len(old)] + new if string.endswith(old):
if not form: form = string[:len(string) - len(old)] + new
pass if not form:
elif form in index or not form.isalpha(): pass
forms.append(form) elif form in index or not form.isalpha():
else: forms.append(form)
oov_forms.append(form) else:
oov_forms.append(form)
if not forms: if not forms:
forms.extend(oov_forms) forms.extend(oov_forms)
if not forms: if not forms: