fix(downloads): take the duration from the file when yt-dlp gives none

A Facebook or Instagram reel extracts with no duration, so the asset kept a
NULL duration_s although the file plays for minutes, and the Library row had
no trim action. The finalize probe now returns the duration with the codecs
(the file wins, info fills a gap), and runs before the sidecars so the .nfo
runtime gets it too.

backend/app/downloads/pipeline.py

@@ -483,13 +483,19 @@ def _download(
         produced = _produced_file(info, staging)
         produced = _ensure_browser_playable(job_id, produced, spec, info)
         ext = produced.suffix.lstrip(".").lower()
+        # Codecs/dims/duration come from a PROBE of the produced file — the ground truth — NOT yt-dlp's
+        # `info` dict, whose fields are absent for some sources (Facebook among them): deriving
+        # media_kind from those nulls misclassified a real H.264 video as a download-only 'file', and a
+        # missing duration hid the trim action. Probed before the sidecars so the .nfo runtime has it
+        # too. Same probe the upload ingest uses.
+        vcodec, acodec, width, height, duration_s = probe.resolve_media_facts(produced, info)
         meta = storage.MediaMeta(
             video_id=info.get("id") or source_ref,
             title=normalize_title(info.get("title")) or source_ref,
             uploader=info.get("uploader") or info.get("channel") or "Unknown Channel",
             upload_date=_parse_upload_date(info.get("upload_date")),
             uniqueid_type=adapter.nfo_uniqueid_type(),
-            duration_s=int(info["duration"]) if info.get("duration") else None,
+            duration_s=duration_s,
             description=info.get("description"),
             thumbnail_url=info.get("thumbnail"),
         )
@@ -515,12 +521,11 @@ def _download(
                 asset.rel_path = rel
                 asset.size_bytes = size
                 asset.duration_s = meta.duration_s
-                # Codecs/dims come from a PROBE of the finished file — the ground truth — NOT yt-dlp's
-                # `info` dict, whose codec fields are absent for some sources (Facebook among them):
-                # deriving media_kind from those nulls misclassified a real H.264 video as a
-                # download-only 'file' with no in-app play/edit. Same probe the upload ingest uses.
-                asset.vcodec, asset.acodec, asset.width, asset.height = probe.resolve_media_facts(
-                    final, info
+                asset.vcodec, asset.acodec, asset.width, asset.height = (
+                    vcodec,
+                    acodec,
+                    width,
+                    height,
                 )
                 asset.container = ext
                 asset.media_kind = mediakind.derive(asset.vcodec, asset.acodec)

backend/app/downloads/probe.py

@@ -81,10 +81,12 @@ def probe_media(path: Path) -> tuple[str | None, str | None, int | None, int | N
 
 def resolve_media_facts(
     path: Path, info: dict
-) -> tuple[str | None, str | None, int | None, int | None]:
-    """(vcodec, acodec, width, height) for the stored download asset: PROBE the finished file on disk
-    (the ground truth). yt-dlp's `info` omits codec/dimension fields for some sources (Facebook among
-    them), so deriving media_kind from it there labelled a real H.264 video a download-only 'file'.
+) -> tuple[str | None, str | None, int | None, int | None, int | None]:
+    """(vcodec, acodec, width, height, duration_s) for the stored download asset: PROBE the finished
+    file on disk (the ground truth). yt-dlp's `info` omits codec/dimension fields for some sources
+    (Facebook among them), so deriving media_kind from it there labelled a real H.264 video a
+    download-only 'file'. It omits the duration too (a Facebook reel, an Instagram reel), and a NULL
+    duration hides the Library row's trim action although the file plays for minutes.
 
     Precedence is by whether the probe RAN, not per-field: if ffprobe read the file, its codec verdict
     is authoritative — INCLUDING a deliberate None (cover-art-only audio, a still image). A per-field
@@ -92,16 +94,23 @@ def resolve_media_facts(
     have (yt-dlp reports a progressive `vcodec` for an audio-mode download whose video it then strips),
     mislabelling an audio download as a broken in-app 'video'. Only when ffprobe could NOT read the
     file at all do we fall back to `info` (the pre-probe behaviour, a safety net); the cosmetic
-    dimensions still gap-fill from `info` in either case."""
+    dimensions and the duration still gap-fill from `info` in either case."""
     data = editmod.run_ffprobe_json(path, show_streams=True, show_format=True)
     info_v = info.get("vcodec") if info.get("vcodec") not in (None, "none") else None
     info_a = info.get("acodec") if info.get("acodec") not in (None, "none") else None
     info_w = int(info["width"]) if info.get("width") else None
     info_h = int(info["height"]) if info.get("height") else None
+    info_d = int(info["duration"]) if info.get("duration") else None
     if data is None:
         # ffprobe could not read the file — fall back to the extractor's info entirely.
-        return info_v, info_a, info_w, info_h
-    pv, pa, pw, ph, _dur = _facts_from_probe(data)
+        return info_v, info_a, info_w, info_h, info_d
+    pv, pa, pw, ph, pd = _facts_from_probe(data)
     # Probe ran → its codecs win outright (a None means "no such track", not "unknown"); info only
-    # gap-fills the cosmetic width/height the probe couldn't read.
-    return pv, pa, pw if pw is not None else info_w, ph if ph is not None else info_h
+    # gap-fills the width/height/duration the probe couldn't read.
+    return (
+        pv,
+        pa,
+        pw if pw is not None else info_w,
+        ph if ph is not None else info_h,
+        pd if pd is not None else info_d,
+    )