feat: auto-detect encoding see #3

tympanix · Nov 10, 2018 · 39170d5 · 39170d5
1 parent 088c67f
commit 39170d5
Show file tree

Hide file tree

Showing 2 changed files with 10 additions and 2 deletions.
diff --git a/subsync/media.py b/subsync/media.py
@@ -7,6 +7,7 @@
 from pysrt import SubRipTime
 import string
 import random
+import chardet
 import re
 from datetime import timedelta
 
@@ -103,7 +104,7 @@ class Subtitle:
     def __init__(self, media, path):
         self.media = media
         self.path = path
-        self.subs = pysrt.open(self.path, encoding='utf-8')
+        self.subs = pysrt.open(self.path, encoding=self._find_encoding())
 
     def labels(self, subs=None):
         if self.media.mfcc is None:
@@ -119,6 +120,13 @@ def labels(self, subs=None):
 
         return labels
 
+    def _find_encoding(self):
+        data = None
+        with open(self.path, "rb") as f:
+            data = f.read()
+        det = chardet.detect(data)
+        return det.get("encoding")
+
 
     def offset(self):
         d = self.media.offset

diff --git a/subsync/version.py b/subsync/version.py
@@ -1 +1 @@
-__version__ = '0.1.4'
+__version__ = '0.1.5'