From f767951150a9d64c301f54a24dfd56ad6d3ec7cd Mon Sep 17 00:00:00 2001
From: Felix <65565033+fexfl@users.noreply.github.com>
Date: Thu, 26 Dec 2024 09:40:20 +0100
Subject: [PATCH] Removed batch print statements

---
 mailcom/parse.py | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/mailcom/parse.py b/mailcom/parse.py
index 12fc739..6f18192 100644
--- 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)