Rundown Generation (scripts/regenerate-rundown.js)

The rundown is a JSON snapshot of the current library organized by category — used for monitoring, inventory, and by the studio/injection tooling to know what is available right now.

How it works

1. Walks each category directory.
2. Probes every audio file with fluent-ffmpeg ffprobe to read duration and embedded metadata.
3. Emits a rundown JSON document with per-category track listings.

Live directory map

The script uses a liveDirs map tying each on-air role to a category directory:

Category	Role
arabic_hits	Main pop deck
arab_techno	Techno/electronic deck
egyptian_trends	Trending/underground deck
beds	Underscore beds (talk-over)
ads	Commercials
jingles	Station jingles
sfx	Sound effects
sweepers	Sweepers / short station IDs
mashups	Mashup content
influencers	Social-influencer clips

Record defaults

When a file has no readable tags, the rundown falls back to:

• artist → LOKLOK System
• energy → 5 (default energy rating)

The same fallback philosophy applies throughout the pipeline: unreadable or missing metadata never crashes a run — it degrades to safe defaults.

Typical output shape

{
  "arabic_hits": [
    { "file": "music/arabic_hits/amr_diab_x.mp3", "duration": 271.3, "artist": "Amr Diab", "title": "…", "energy": 5 }
  ],
  "sweepers": [
    { "file": "imaging/sweepers/loklok_id_01.mp3", "duration": 4.2, "artist": "LOKLOK System", "title": "…", "energy": 5 }
  ]
}

Run it after a pipeline pass to produce an up-to-date inventory of everything the station can play.