From 11f942dfb9c1af54ade05f793d48480ff3cfc4ae Mon Sep 17 00:00:00 2001
From: Adrian Petrescu <adrian@apetre.sc>
Date: Mon, 7 Nov 2022 15:34:15 -0500
Subject: [PATCH] Add support for WebVTT timeframes in MS Teams' non-compliant
 format

---
 tests/subtitles/teams_timeframe.vtt | 7 +++++++
 tests/test_webvtt_parser.py         | 7 +++++++
 webvtt/parsers.py                   | 2 +-
 webvtt/structures.py                | 2 +-
 4 files changed, 16 insertions(+), 2 deletions(-)
 create mode 100644 tests/subtitles/teams_timeframe.vtt

diff --git a/tests/subtitles/teams_timeframe.vtt b/tests/subtitles/teams_timeframe.vtt
new file mode 100644
index 0000000..03f63a6
--- /dev/null
+++ b/tests/subtitles/teams_timeframe.vtt
@@ -0,0 +1,7 @@
+WEBVTT
+
+0:0:0.0 --> 0:0:0.570
+<v Yuri Bakulin>There you go.</v>
+
+0:0:0.210 --> 0:0:1.130
+<v Adrian Petrescu>Why hello there, Yuri.</v>
diff --git a/tests/test_webvtt_parser.py b/tests/test_webvtt_parser.py
index a26741a..fac638c 100644
--- a/tests/test_webvtt_parser.py
+++ b/tests/test_webvtt_parser.py
@@ -90,6 +90,13 @@ def test_webvtt_timestamps_format(self):
         self.assertEqual(vtt.captions[2].start, '00:00:11.890')
         self.assertEqual(vtt.captions[2].end, '00:00:16.320')
 
+    def test_webvtt_microsoft_teams_timestamps_format(self):
+        vtt = webvtt.read(self._get_file('teams_timeframe.vtt'))
+        self.assertEqual(vtt.captions[0].start, '00:00:00.000')
+        self.assertEqual(vtt.captions[0].end, '00:00:00.570')
+        self.assertEqual(vtt.captions[1].start, '00:00:00.210')
+        self.assertEqual(vtt.captions[1].end, '00:00:01.130')
+
     def test_parse_timestamp(self):
         caption = Caption(start='02:03:11.890')
         self.assertEqual(
diff --git a/webvtt/parsers.py b/webvtt/parsers.py
index 3a978ca..e95a654 100644
--- a/webvtt/parsers.py
+++ b/webvtt/parsers.py
@@ -155,7 +155,7 @@ class WebVTTParser(TextBasedParser):
     WebVTT parser.
     """
 
-    TIMEFRAME_LINE_PATTERN = re.compile(r'\s*((?:\d+:)?\d{2}:\d{2}.\d{3})\s*-->\s*((?:\d+:)?\d{2}:\d{2}.\d{3})')
+    TIMEFRAME_LINE_PATTERN = re.compile(r'\s*((?:\d+:)?\d{1,2}:\d{1,2}.\d{1,3})\s*-->\s*((?:\d+:)?\d{1,2}:\d{1,2}.\d{1,3})')
     COMMENT_PATTERN = re.compile(r'NOTE(?:\s.+|$)')
     STYLE_PATTERN = re.compile(r'STYLE[ \t]*$')
 
diff --git a/webvtt/structures.py b/webvtt/structures.py
index 84f376d..4fa04d8 100644
--- a/webvtt/structures.py
+++ b/webvtt/structures.py
@@ -2,7 +2,7 @@
 
 from .errors import MalformedCaptionError
 
-TIMESTAMP_PATTERN = re.compile('(\d+)?:?(\d{2}):(\d{2})[.,](\d{3})')
+TIMESTAMP_PATTERN = re.compile('(\d+)?:?(\d{1,2}):(\d{1,2})[.,](\d{1,3})')
 
 __all__ = ['Caption']