[utils] Align `parse_duration()` behaviour with yt-dlp

* handle comma-separated long-form durations
* support : as millisecond separator.

1 files changed, 25 insertions, 21 deletions

diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py
index edac2456d..b93f4be5c 100644
--- a/youtube_dl/utils.py
+++ b/youtube_dl/utils.py
@@ -3931,38 +3931,46 @@ def parse_duration(s):
         return None
 
     s = s.strip()
+    if not s:
+        return None
 
     days, hours, mins, secs, ms = [None] * 5
-    m = re.match(r'(?:(?:(?:(?P<days>[0-9]+):)?(?P<hours>[0-9]+):)?(?P<mins>[0-9]+):)?(?P<secs>[0-9]+)(?P<ms>\.[0-9]+)?Z?$', s)
+    m = re.match(r'''(?x)
+            (?P<before_secs>
+                (?:(?:(?P<days>[0-9]+):)?(?P<hours>[0-9]+):)?
+                (?P<mins>[0-9]+):)?
+                (?P<secs>(?(before_secs)[0-9]{1,2}|[0-9]+))
+            (?:[.:](?P<ms>[0-9]+))?Z?$
+        ''', s)
     if m:
-        days, hours, mins, secs, ms = m.groups()
+        days, hours, mins, secs, ms = m.group('days', 'hours', 'mins', 'secs', 'ms')
     else:
         m = re.match(
             r'''(?ix)(?:P?
                 (?:
-                    [0-9]+\s*y(?:ears?)?\s*
+                    [0-9]+\s*y(?:ears?)?,?\s*
                 )?
                 (?:
-                    [0-9]+\s*m(?:onths?)?\s*
+                    [0-9]+\s*m(?:onths?)?,?\s*
                 )?
                 (?:
-                    [0-9]+\s*w(?:eeks?)?\s*
+                    [0-9]+\s*w(?:eeks?)?,?\s*
                 )?
                 (?:
-                    (?P<days>[0-9]+)\s*d(?:ays?)?\s*
+                    (?P<days>[0-9]+)\s*d(?:ays?)?,?\s*
                 )?
                 T)?
                 (?:
-                    (?P<hours>[0-9]+)\s*h(?:ours?)?\s*
+                    (?P<hours>[0-9]+)\s*h(?:(?:ou)?rs?)?,?\s*
                 )?
                 (?:
-                    (?P<mins>[0-9]+)\s*m(?:in(?:ute)?s?)?\s*
+                    (?P<mins>[0-9]+)\s*m(?:in(?:ute)?s?)?,?\s*
                 )?
                 (?:
-                    (?P<secs>[0-9]+)(?P<ms>\.[0-9]+)?\s*s(?:ec(?:ond)?s?)?\s*
+                    (?P<secs>[0-9]+)(?:\.(?P<ms>[0-9]+))?\s*s(?:ec(?:ond)?s?)?\s*
                 )?Z?$''', s)
         if m:
-            days, hours, mins, secs, ms = m.groups()
+            days, hours, mins, secs, ms = m.group('days', 'hours', 'mins', 'secs', 'ms')
         else:
             m = re.match(r'(?i)(?:(?P<hours>[0-9.]+)\s*(?:hours?)|(?P<mins>[0-9.]+)\s*(?:mins?\.?|minutes?)\s*)Z?$', s)
             if m:
@@ -3970,17 +3978,13 @@ def parse_duration(s):
             else:
                 return None
 
-    duration = 0
-    if secs:
-        duration += float(secs)
-    if mins:
-        duration += float(mins) * 60
-    if hours:
-        duration += float(hours) * 60 * 60
-    if days:
-        duration += float(days) * 24 * 60 * 60
-    if ms:
-        duration += float(ms)
+    duration = (
+        ((((float(days) * 24) if days else 0)
+          + (float(hours) if hours else 0)) * 60
+         + (float(mins) if mins else 0)) * 60
+        + (float(secs) if secs else 0)
+        + (float(ms) / 10 ** len(ms) if ms else 0))
+
     return duration