Two additions to the metadata sidecars, following up on the 2026-09-01
collab/dedup investigation already documented here:
- coauthors added to the post-level JSON's include list -- a direct,
native signal for collab detection instead of inferring it from the
filename/directory identity mismatch also documented that day.
- width, height, width_original, height_original and tagged_users get
their own per-file sidecar (event: "file", one JSON per downloaded
media item, named "<filename>.json"), since a carousel's items can
each have different dimensions and tags -- data the existing
post-level JSON has no way to represent. owner is deliberately left
out, same reasoning as audio_user's existing exclusion.
Verified against a live re-fetch of an already-archived carousel:
correct per-item data came back, zero media re-downloaded (skip-archive
still applies; only the new sidecars are new files). 3 new tests.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011qAds5qr7nZRq5R4yAuxUk
gallery-dl's dedicated reels extractor POSTs to /api/v1/clips/user/,
which now 302-redirects for this account -- confirmed across multiple
profiles, hours apart, with a freshly-warmed session and a correct
X-IG-WWW-Claim header (ruled out as the cause). The reels tab itself
loads fine in a real, already-signed-in browser, so reels-scrape.py
drives that same Chrome via its loopback CDP port, scrolls the reels
tab like a person would, and scrapes /reel/<code>/ links out of the
rendered page instead of calling the blocked endpoint at all.
It only finds shortcodes -- deduped against the archive via the same
--index gdl-sync.py already uses -- and prints new post URLs. Feeding
many of those into gdl-sync.py needed two small additions: a
--post-urls-file so the list doesn't have to become a giant argv, and
inter-item pacing in run_post_urls (each --post-url was its own
subprocess with nothing pacing the gap between them).
Verified end to end against zindoriyam: 26 reels found, 16 already
archived, 10 new ones fetched and published cleanly.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011qAds5qr7nZRq5R4yAuxUk
Lets an out-of-band link (shared by someone, not one of the tracked
profiles) be pulled in directly by URL, filed under its owner's account
like any other post. Bypasses profile planning, archive-db seeding, and
the --min-interval floor entirely, since it's a single request rather
than a recurring surface to budget against.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011qAds5qr7nZRq5R4yAuxUk
The scripts and docs for fetching from Instagram now live here rather than on
main, which is the branch published to GitHub. They carry things that do not
belong in a public repo: the fetch host's public IP, the browser profile path
the cookie is read from, the NAS archive path, and the list of accounts being
archived.
This branch is a superset of main — the viewer plus the tooling — so it can
take main's changes by merging, and the npm script and CLAUDE.md entries that
reference the tooling live here where the files actually exist.
Restored with the sync work from the 2026-08-20 run already in place: the
--abort flag, the corrected yt-dlp install advice, and the measurements behind
both.
Note that main's history was rewritten to strip these paths, so the tooling's
own per-file history does not exist on this branch. It is preserved on gitea
as pre-rewrite-20260820 and pre-rewrite-tooling-20260820.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UXfdJu7QhSJLr47K7koTDF
ergosteurandClaude Sonnet 5