legacy-arrflix/playbooks/import-media/runs/the-guardian-snowden-2013-20130709.md

the-guardian-snowden-2013-20130709

Single-video YouTube import into the STOCK Jellyfin at tv.s8n.ru (container jellyfin-stock), Education library (collectionType=movies, internet providers disabled).

First import into a brand-new "The Guardian" channel folder. Exposed a single-file folder caveat in Jellyfin movie parsing — see Notes.

Provenance

• Source: YouTube — https://www.youtube.com/watch?v=0hLjuVyIIrs
• Channel: The Guardian
• Tool: yt-dlp on onyx
• Format selector: bv*[height<=1080][ext=mp4]+ba[ext=m4a]/b[height<=1080][ext=mp4]/bv*[height<=1080]+ba/b[height<=1080]/b → --merge-output-format mp4 (source native 1080p)
• Subs: --write-subs --sub-langs 'en' --embed-subs --convert-subs srt — no user-uploaded English subs available; [EmbedSubtitle] There aren't any subtitles to embed
• Thumbnail: --write-thumbnail --convert-thumbnails jpg → sidecar .jpg used as Primary
• Staging path on onyx: /home/admin/staging-jelly/The Guardian/

Filename normalisation

Raw YouTube title:

NSA whistleblower Edward Snowden: 'I don't want to live in a society that does these sort of things'

yt-dlp's safe-name pass replaced the ASCII : with the fullwidth ： (U+FF1A). Per playbook §1f the canonical replacement is ASCII -, so the file was renamed before rsync to:

NSA whistleblower Edward Snowden - 'I don't want to live in a society that does these sort of things' — 20130709.mp4

Apostrophes preserved (playbook §1f: smart quotes and apostrophes pass).

Target

• Server: jellyfin-stock on nullstone, public URL https://tv.s8n.ru
• Library: Education (collectionType=movies, EnableInternetProviders=false)
• Path on host: /home/user/media/education/The Guardian/NSA whistleblower Edward Snowden - 'I don't want to live in a society that does these sort of things' — 20130709.mp4
• Container view: same under /media/education/
• Item ID: 578da493fdcff4e8fde5137adbcaebdb

Sidecar files

Kind	File
Media	… — 20130709.mp4 (46,324,047 B, ~44 MiB)
Subtitle	none
Thumbnail	… — 20130709.jpg (34,895 B) — Primary via Local Posters

Counts

	Before	After	Delta
Education / The Guardian items	0	1	+1
Education / The Guardian channel folders	0	1	+1

Stream summary

Duration: 00:12:34.30, bitrate: 491 kb/s
Stream #0:0[0x1](und): Video: av1 (libdav1d) (Main), yuv420p(tv, bt709), 1920x1080, 358 kb/s, 25 fps
Stream #0:1[0x2](eng): Audio: aac (LC), 44100 Hz, stereo, fltp, 127 kb/s

AV1 1080p25 + stereo AAC. No subtitle streams.

Subtitle status

• Embedded: no (channel does not publish user-uploaded en captions on this video; auto-CC not requested by playbook).
• External sidecar: no.
• Action: deferred. If subs become required, follow playbooks/subtitles/ WhisperX pipeline (do not fall back to YouTube auto-CC; see feedback_subtitle_accuracy_priority).

Verification checks

• Folder/filename canonical (<Channel>/<Title> — <YYYYMMDD>.mp4).
• No forbidden chars in path (fullwidth ： → - rename in staging).
• Permissions user:user 644 / 755.
• Scan Media Library triggered, State=Idle, ran twice (initial scan did not pull file in until folder was visible to library monitor — see Notes).
• /Items?searchTerm=NSA+whistleblower&IncludeItemTypes=Movie returns the expected single item after Name lock.
• ImageTags.Primary present.
• Direct-play in mobile / Smart-TV client not exercised.

Notes / surprises

Single-file channel folder → Jellyfin parses folder name as title

Jellyfin's MovieResolver applies the "movie-in-own-folder" heuristic when a folder under a collectionType=movies library contains exactly one media file. The resulting BaseItem.Name is taken from the folder name, not the filename — so the Snowden mp4 initially indexed with Name="The Guardian" instead of the actual video title.

Workaround used (no playbook change required for this run):

# Pull current item, set Name explicitly, LockData=true, POST back
ID=578da493fdcff4e8fde5137adbcaebdb
curl -sf -H "X-Emby-Token: $TOK" \
  "$SERVER_URL/Users/$USER_ID/Items/$ID" \
  | jq '.Name = "NSA whistleblower Edward Snowden - ..."  | .LockData = true' \
  | curl -sf -X POST -H "X-Emby-Token: $TOK" -H 'Content-Type: application/json' \
      --data @- "$SERVER_URL/Items/$ID"
# expect HTTP 204; Name now locked so future refresh cannot revert it.

Generalises to any first-video import into a brand-new channel folder. Recommend documenting in playbooks/import-media/README.md §1d as a known caveat with the LockData fix; will propose a playbooks/ PR after the second occurrence to confirm reproducibility.

Two scan passes required

First POST /ScheduledTasks/Running/<scan-task-id> returned to Idle while the file was already on disk but reported no new Education item. A second identical invocation indexed the file. Root cause unclear — possible LibraryMonitor race during inotify pickup of the freshly created folder. Documenting per playbook §5d / known-broken-trigger guidance.