From f5980781ef1ce0cd06ffaa31ba632d2bbf15b952 Mon Sep 17 00:00:00 2001
From: Reuben Morais <reuben.morais@gmail.com>
Date: Tue, 18 Apr 2017 08:15:53 -0700
Subject: [PATCH] Open STM files as UTF-8

---
 util/stm.py | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/util/stm.py b/util/stm.py
index 55e77002..23383341 100644
--- a/util/stm.py
+++ b/util/stm.py
@@ -1,3 +1,4 @@
+import codecs
 import unicodedata
 
 class STMSegment(object):
@@ -55,7 +56,7 @@ def parse_stm_file(stm_file):
     Parses an STM file at ``stm_file`` into a list of :class:`STMSegment`.
     """
     stm_segments = []
-    with open(stm_file) as stm_lines:
+    with codecs.open(stm_file, encoding="utf-8") as stm_lines:
         for stm_line in stm_lines:
             stmSegment = STMSegment(stm_line)
             if not "ignore_time_segment_in_scoring" == stmSegment.transcript: