youtube-dl

commit 4e0cff2a50f4c297fc25dae01c460596d8f5badb
parent 0436157b95ddd6d7c415549ba48b7c0305b2c0df
Author: remitamine <remitamine@gmail.com>
Date:   Tue,  2 Feb 2016 18:36:26 +0100

Merge pull request #8348 from remitamine/dfxp2srt-text

[utils] fix dfxp2srt text extraction(fixes #8055)
Diffstat:
M
youtube_dl/utils.py
 | 
29
++++++++++++++++++-----------

1 file changed, 18 insertions(+), 11 deletions(-)
diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py
@@ -2017,20 +2017,27 @@ def dfxp2srt(dfxp_data):
         'ttaf1': 'http://www.w3.org/2006/10/ttaf1',
     })
 
-    def parse_node(node):
-        str_or_empty = functools.partial(str_or_none, default='')
+    class TTMLPElementParser:
+        out = ''
 
-        out = str_or_empty(node.text)
+        def start(self, tag, attrib):
+            if tag in (_x('ttml:br'), _x('ttaf1:br'), 'br'):
+                self.out += '\n'
 
-        for child in node:
-            if child.tag in (_x('ttml:br'), _x('ttaf1:br'), 'br'):
-                out += '\n' + str_or_empty(child.tail)
-            elif child.tag in (_x('ttml:span'), _x('ttaf1:span'), 'span'):
-                out += str_or_empty(parse_node(child))
-            else:
-                out += str_or_empty(xml.etree.ElementTree.tostring(child))
+        def end(self, tag):
+            pass
 
-        return out
+        def data(self, data):
+            self.out += data
+
+        def close(self):
+            return self.out.strip()
+
+    def parse_node(node):
+        target = TTMLPElementParser()
+        parser = xml.etree.ElementTree.XMLParser(target=target)
+        parser.feed(xml.etree.ElementTree.tostring(node))
+        return parser.close()
 
     dfxp = compat_etree_fromstring(dfxp_data.encode('utf-8'))
     out = []