papers/whistlegraph-platter: dropbox manifest (12.9k files / 421 GB) + score↔video index + desktop archive viewer
- Expand README with concrete per-folder content from dropbox-manifest.md and the locally-synced DO Spaces per-piece dirs (10 score↔video pairs already organized by piece-name).
- tools/dropbox-manifest.mjs: no-dep Node script that walks /Whistlegraph/ via Dropbox HTTP API using the vault OAuth token; writes dropbox-manifest.{md,json} + dropbox-files.json.
- tools/build-html.mjs: bundles the JSONs into ~/Desktop/whistlegraph-archive.html — a self-contained, terminal-themed browser with sortable folder summary, full file search/filter, preset chips (video/audio/image/scores/procreate), and pagination.
- Manifest reveals: Scores/Score Rev1 PDF/ filenames match paper figures (rosetta confirmed across paper, Spaces, Dropbox); Research/Whistlegraphs for GAN/ is an already-curated 217 MB ML training set; 40 .procreate files carry native per-stroke timing.
- Sized "analysis core" (Scores + Procreate + Research/GAN + TikTok + Manuscripts + Pixaki + Audio Drafts) is ~17.7 GB.
Note: dropbox-{manifest,files}.json contain filenames-only listings (no file contents) including business-doc paths (LLC/Taxes/Invoices). Safe for knot-only push; revisit if the GitHub mirror is reactivated.