Merge remote-tracking branch 'duncankl/airmozilla'

[youtube-dl] / youtube_dl / extractor / ceskatelevize.py

diff --git a/youtube_dl/extractor/ceskatelevize.py b/youtube_dl/extractor/ceskatelevize.py
index ba8376338c6d2046951e18c34fb778c27d3e530c..65f6be62313dfc623cf1f9aa7adc52282872aade 100644 (file)
--- a/youtube_dl/extractor/ceskatelevize.py
+++ b/youtube_dl/extractor/ceskatelevize.py
@@ -104,6 +104,11 @@ class CeskaTelevizeIE(InfoExtractor):
         duration = float_or_none(item.get('duration'))
         thumbnail = item.get('previewImageUrl')
 
+        subtitles = {}
+        subs = item.get('subtitles')
+        if subs:
+            subtitles = self.extract_subtitles(episode_id, subs)
+
         return {
             'id': episode_id,
             'title': title,
@@ -111,4 +116,40 @@ class CeskaTelevizeIE(InfoExtractor):
             'thumbnail': thumbnail,
             'duration': duration,
             'formats': formats,
+            'subtitles': subtitles,
         }
+
+    def _get_subtitles(self, episode_id, subs):
+        original_subtitles = self._download_webpage(
+            subs[0]['url'], episode_id, 'Downloading subtitles')
+        srt_subs = self._fix_subtitles(original_subtitles)
+        return {
+            'cs': [{
+                'ext': 'srt',
+                'data': srt_subs,
+            }]
+        }
+
+    @staticmethod
+    def _fix_subtitles(subtitles):
+        """ Convert millisecond-based subtitles to SRT """
+
+        def _msectotimecode(msec):
+            """ Helper utility to convert milliseconds to timecode """
+            components = []
+            for divider in [1000, 60, 60, 100]:
+                components.append(msec % divider)
+                msec //= divider
+            return "{3:02}:{2:02}:{1:02},{0:03}".format(*components)
+
+        def _fix_subtitle(subtitle):
+            for line in subtitle.splitlines():
+                m = re.match(r"^\s*([0-9]+);\s*([0-9]+)\s+([0-9]+)\s*$", line)
+                if m:
+                    yield m.group(1)
+                    start, stop = (_msectotimecode(int(t)) for t in m.groups()[1:])
+                    yield "{0} --> {1}".format(start, stop)
+                else:
+                    yield line
+
+        return "\r\n".join(_fix_subtitle(subtitles))