Fix clean up tokenization error

Author: OlivieFranklova

column2Vec/research/column2Vec_re.py

@@ -48,7 +48,8 @@
 ]
 MODEL = "paraphrase-multilingual-mpnet-base-v2"  # 'bert-base-nli-mean-tokens'
 THIS_DIR = os.path.dirname(os.path.abspath(__file__))
-model = SentenceTransformer(MODEL)
+model = SentenceTransformer(MODEL, tokenizer_kwargs={
+        'clean_up_tokenization_spaces': True})
 
 
 def count_embedding(column1: pd.Series, function, key: str) -> pd.Series:

constants.py

@@ -51,7 +51,9 @@ class TrainedModel:
     """
 
     configure()
-    __model = SentenceTransformer("paraphrase-multilingual-mpnet-base-v2")
+    __model = SentenceTransformer("paraphrase-multilingual-mpnet-base-v2", tokenizer_kwargs={
+        'clean_up_tokenization_spaces': True,
+    })
 
     def set_module(self, model: SentenceTransformer):
         """

similarity/DataFrameMetadataCreator.py

@@ -63,7 +63,8 @@ def __init__(self, dataframe: pd.DataFrame):
         """
         self.dataframe = dataframe
         self.metadata = DataFrameMetadata()
-        self.model: Optional[SentenceTransformer] = SentenceTransformer("bert-base-nli-mean-tokens")
+        self.model: Optional[SentenceTransformer] = SentenceTransformer("bert-base-nli-mean-tokens", tokenizer_kwargs={
+        'clean_up_tokenization_spaces': True})
         self.metadata.size = dataframe.shape[0]
         self.metadata.column_names = list(dataframe.columns)
         self.metadata.column_names_clean = {i: re.sub("[^(0-9 |a-z).]", " ", i.lower()) for i in self.metadata.column_names}
@@ -152,7 +153,8 @@ def __get_model(self) -> SentenceTransformer:
         :return: embedding model if exists or creates new one
         """
         if not self.model:
-            self.model = SentenceTransformer("bert-base-nli-mean-tokens")
+            self.model = SentenceTransformer("bert-base-nli-mean-tokens", tokenizer_kwargs={
+        'clean_up_tokenization_spaces': True})
         return self.model
 
     # Setting Creator

test/test_column2Vec.py

@@ -33,7 +33,8 @@ def get_vectors(function, data):
     count = 1
     for key in data:
         # print("Processing column: " + key + " " + str(round((count / len(data)) * 100, 2)) + "%")
-        result[key] = function(data[key], SentenceTransformer(MODEL), key)
+        result[key] = function(data[key], SentenceTransformer(MODEL, tokenizer_kwargs={
+        'clean_up_tokenization_spaces': True}), key)
         count += 1
     end = time.time()
     print(f"ELAPSED TIME :{end - start}")