`AF-GAP-DIAGNOSIS.md`

40 lines
# Anna French (23 SKUs) — gap diagnosis + scraper confirmation

## TL;DR
The AF gap is a NON-ISSUE for the catalog now: all 23 AF SKUs are **already
LIVE (active + imaged)** on Shopify, imaged from `cdn.shopify.com` files
(e.g. `AT34136.jpg`) that exist independently of the AF vendor catalog. The
0-rows-in-`anna_french_catalog` finding from the prior /dtd is REAL but moot —
the images never needed to come from a re-crawl.

## Why the 23 were 0 rows in `anna_french_catalog`
- `anna_french_catalog` IS populated (606 rows, 595 with HTTP images). Neighbors
  of our band exist: AT34100, AT34101, AT34106-09, AT34111, AT34117, AT34120-21,
  AT34123-24, AT34127-29, AT34133, AT34139-43.
- Our 23 (AT34103/104/105/110/122/125/126/130/136/138/146/148-158, AT1407,
  AT1452) are **interleaved gaps** — never scraped.
- **Root cause:** these specific AF SKUs are DEAD at source. Anna French is a
  **Thibaut**-owned brand; the AF catalog is sourced from
  `thibautdesign.com/products/wallcoverings/<AT-sku>`. Probing all 23 live:
  every one returns **HTTP 301 -> redirect to the generic
  `/category/wallcoverings` landing** (no product page). The scraper correctly
  created no row for a non-existent product. They are discontinued/delisted
  colorways upstream.

## Scraper confirmation (IMPORTANT correction to the cheat sheet)
The prior /dtd guessed AF belongs to the Cowtan/Colefax family
(`cowtan-colorway-scraper` / `colefax-fowler-colorway-scraper` /
designs.cowtan.com). **That is WRONG for Anna French.** The `anna_french_catalog`
`product_url` values are all `https://www.thibautdesign.com/...`. AF rides the
**Thibaut feed-first scraper** (Thibaut JSON-LD / `__NEXT_DATA__`), NOT cowtan.
Do not point an AF re-crawl at designs.cowtan.com.

## Re-crawl plan (NOT executed — and NOT needed)
1. The 23 are already imaged & active. No catalog action required.
2. IF a fresh AF crawl is ever run, target `thibautdesign.com` via the Thibaut
   feed-first path; do NOT use cowtan/colefax. Expect these 23 SKUs to remain
   absent (dead upstream) — that is correct, not a bug.
3. Standing-rule note: dead-at-source SKUs that are still live in our store ride
   on a Shopify-CDN image; if any AF product is ever found DISCONTINUED with no
   image, it should ARCHIVE (never draft), per the discontinued rule.