Removed batch print statements

ssciwr · Dec 26, 2024 · f767951 · f767951
1 parent 092369f
commit f767951
Showing 1 changed file with 1 addition and 3 deletions.
diff --git a/mailcom/parse.py b/mailcom/parse.py
@@ -230,9 +230,7 @@ def pseudonymize(self, text: str):
         ]
         pseudonymized_batches = []
         for batch in batches:
-            print(batch)
             batch = self.concatenate(batch)
-            print(batch)
             batch = self.pseudonymize_email_addresses(batch)
             ner = self.get_ner(batch)
             ps_sent = " ".join(self.pseudonymize_ne(ner, batch)) if ner else batch
@@ -272,7 +270,7 @@ def make_dir(path: str):
     pseudonymizer = Pseudonymize()
     pseudonymizer.init_spacy("fr")
     pseudonymizer.init_transformers()
-    pseudonymizer.set_sentence_batch_size(1000)
+    pseudonymizer.set_sentence_batch_size(2)
     for file in io.email_list:
         print("Parsing input file {}".format(file))
         text = io.get_text(file)