VTT reader: fix handling of multiple timestamp tags

Author: palemieux

src/main/python/ttconv/vtt/reader.py

@@ -53,6 +53,7 @@ class _TextCueParser:
 
   def __init__(self, paragraph: model.P, line_number: int) -> None:
     self.line_num: int = line_number
+    self.paragraph: model.P = paragraph
     self.parent: model.ContentElement = paragraph
 
     # handle the special case of ruby elements where children cannot be added one by one
@@ -73,23 +74,25 @@ def handle_token(self, token: Token) -> None:
 
   def _handle_ts(self, token: TimestampTagToken):
 
-    span = self._make_span(self.parent)
-    self.parent.push_child(span)
-    self.parent = span
-
     ts = vtt_timestamp_to_secs(token.timestamp)
-    parent_begin = None
-    parent = self.parent
-    while parent is not None:
-      parent_begin = parent.get_begin()
-      if parent_begin is not None:
-        break
-      parent = parent.parent()
-    if ts is not None and parent_begin is not None and parent_begin <= ts:
-      span.set_begin(ts - parent_begin)
-    else:
+    if ts is None:
+      LOGGER.warning("Invalid timestamp tag %s at line %s", token.timestamp, self.line_num)
+      return
+
+    # we handle only top-level timestamp tags
+    if self.parent.get_begin() is None:
+      LOGGER.warning("Nested timestamp tag %s at line %s", token.timestamp, self.line_num)
+      return
+
+    p_begin = self.paragraph.get_begin()
+    if p_begin is None or p_begin >= ts:
       LOGGER.warning("Invalid timestamp tag %s", token.timestamp)
 
+    span = self._make_span(self.paragraph)
+    span.set_begin(ts - p_begin)
+    self.paragraph.push_child(span)
+    self.parent = span
+
   def _handle_starttag(self, token: StartTagToken):
 
     tag = token.tag.lower()

src/test/python/test_vtt_reader.py

@@ -141,6 +141,28 @@ def test_single_line_with_space(self):
     div = list(body[0])
     self.assertEqual(len(div), 2)
 
+  def test_toplevel_timestamp_tags(self):
+    # from https://github.com/sandflow/ttconv/issues/439
+    SAMPLE = """WEBVTT
+Kind: captions
+Language: en
+
+00:00:00.799 --> 00:00:02.869 align:start position:0%
+\x20
+hi<00:00:01.040><c> everyone</c><00:00:01.920><c> today</c><00:00:02.240><c> we're</c><00:00:02.399><c> going</c><00:00:02.639><c> to</c><00:00:02.720><c> be</c>
+
+00:00:02.869 --> 00:00:02.879 align:start position:0%
+hi everyone today we're going to be
+"""
+
+    doc = to_model(io.StringIO(SAMPLE))
+    body = list(doc.get_body())
+    spans_and_brs = list(body[0][0])
+    self.assertIsNone(spans_and_brs[0].get_begin()) # \x20
+    self.assertIsNone(spans_and_brs[2].get_begin()) # hi
+    self.assertEqual(spans_and_brs[3].get_begin(), 1.040 - 0.799) # everyone
+    self.assertEqual(spans_and_brs[4].get_begin(), 1.920 - 0.799) # today
+
   def test_italic(self):
     f = io.StringIO(r"""WEBVTT