Fix subscription feed crash on unparseable time_published

Failed timestamp parsing left a raw string in the integer
time_published column, crashing posix_to_dumbed_down (float - str).
Store None on parse failure and coerce string values defensively on
read so the feed renders and corrupt rows self-heal.

Fixes ~heckyel/yt-local#26

1 files changed, 30 insertions, 1 deletions

diff --git a/youtube/subscriptions.py b/youtube/subscriptions.py
index 7d3efab..50a7942 100644
--- a/youtube/subscriptions.py
+++ b/youtube/subscriptions.py
@@ -291,6 +291,11 @@ def youtube_timestamp_to_posix(dumb_timestamp):
 
 def posix_to_dumbed_down(posix_time):
     '''Inverse of youtube_timestamp_to_posix.'''
+    # Legacy/corrupt DB rows may store time_published as a non-numeric string
+    # (e.g. an unparsed "2 days ago"). Coerce defensively so the feed renders
+    # instead of crashing with: float - str.
+    posix_time = _coerce_posix_time(posix_time)
+
     delta = int(time.time() - posix_time)
     # Guard against future timestamps (clock drift) without relying on
     # `assert` (which is stripped under `python -O`).
@@ -312,12 +317,31 @@ def posix_to_dumbed_down(posix_time):
 
 
 def exact_timestamp(posix_time):
+    posix_time = _coerce_posix_time(posix_time)
     result = time.strftime('%I:%M %p %m/%d/%y', time.localtime(posix_time))
     if result[0] == '0':    # remove 0 infront of hour (like 01:00 PM)
         return result[1:]
     return result
 
 
+def _coerce_posix_time(posix_time):
+    '''Best-effort conversion of a stored time_published value to a float.
+
+    Older versions could persist an unparsed timestamp string into the
+    integer time_published column. Recover gracefully: try numeric
+    conversion, then the dumbed-down parser, and fall back to "now".'''
+    if isinstance(posix_time, (int, float)):
+        return posix_time
+    try:
+        return float(posix_time)
+    except (TypeError, ValueError):
+        pass
+    try:
+        return youtube_timestamp_to_posix(posix_time)
+    except Exception:
+        return time.time()
+
+
 try:
     existing_thumbnails = set(os.path.splitext(name)[0] for name in os.listdir(thumbnails_directory))
 except FileNotFoundError:
@@ -575,7 +599,12 @@ def _get_upstream_videos(channel_id):
             try:
                 video_item['time_published'] = youtube_timestamp_to_posix(video_item['time_published']) - i  # subtract a few seconds off the videos so they will be in the right order
             except Exception:
-                print(video_item)
+                print('Failed to parse time_published:', video_item)
+                # Don't leave the unparsed string in place: it would be stored
+                # in the integer time_published column and later crash
+                # posix_to_dumbed_down (float - str). Treat as unknown instead;
+                # the fill-in passes below will approximate it.
+                video_item['time_published'] = None
         else:
             video_item['is_time_published_exact'] = False
             video_item['time_published'] = None