From 61770e0fc93e5305e30d9e2c260ef507cba741e7 Mon Sep 17 00:00:00 2001
From: Jordi Mas <jmas@softcatala.org>
Date: Sat, 21 Dec 2024 10:56:36 +0100
Subject: [PATCH] Do not hash _

---
 open_dubbing/utterance.py | 8 +++++++-
 tests/utterance_test.py   | 2 ++
 2 files changed, 9 insertions(+), 1 deletion(-)

diff --git a/open_dubbing/utterance.py b/open_dubbing/utterance.py
index c3ab7cf..cf570d3 100644
--- a/open_dubbing/utterance.py
+++ b/open_dubbing/utterance.py
@@ -104,7 +104,13 @@ def save_utterances(
 
     def _hash_utterances(self, utterance_metadata):
         for utterance in utterance_metadata:
-            dict_str = json.dumps(utterance, sort_keys=True)
+            filtered_dict = {
+                key: value
+                for key, value in utterance.items()
+                if not key.startswith("_")
+            }
+
+            dict_str = json.dumps(filtered_dict, sort_keys=True)
             _hash = hashlib.sha256(dict_str.encode()).hexdigest()
             utterance["_hash"] = _hash
 
diff --git a/tests/utterance_test.py b/tests/utterance_test.py
index f283094..b6ffa02 100644
--- a/tests/utterance_test.py
+++ b/tests/utterance_test.py
@@ -70,6 +70,7 @@ def test_hash_utterances(self):
             {
                 "start": 1.26,
                 "end": 3.94,
+                "_private_not_hashed": "0",
             },
             {
                 "start": 5.24,
@@ -86,6 +87,7 @@ def test_hash_utterances(self):
             {
                 "start": 1.26,
                 "end": 3.94,
+                "_private_not_hashed": "0",
                 "_hash": "2fa6f80e0c81fb8e142f2dbbad0bceff7c21a031833b5752bc1cfd799f6b3bc6",
             },
             {