Adding unitest for tokenization in french (with title)

This commit is contained in:
Gregory Howard 2017-04-27 11:53:44 +02:00
parent 8ff4682255
commit 13b6957c8e

View File

@ -10,12 +10,10 @@ class Lemmatizer(Lemmatizer):
return cls(lookup or {}) return cls(lookup or {})
def __init__(self, lookup): def __init__(self, lookup):
print("Mon lemmatizer")
self.lookup = lookup self.lookup = lookup
def __call__(self, string, univ_pos, morphology=None): def __call__(self, string, univ_pos, morphology=None):
print("call")
try: try:
return self.lookup[string] return set([self.lookup[string]])
except: except:
return string return set([string])