added draft scope for review/search loop

This commit is contained in:
2026-03-21 09:48:34 -04:00
parent afadd0c0d0
commit e8e11e15b3
2 changed files with 64 additions and 1 deletions

View File

@@ -500,4 +500,66 @@ Decide whether two normalized retailer items are "the same product"; match items
** Symptoms
- `LIME` and `LIME . / .` appearing in canonical_catalog:
- names must come from review-approved names, not raw strings
* notes
** to fix
*** Done
fuji apple, apple, produce (not apple, fruit, produce)
spinach, , produce -> frozen vs fresh?
frozen chicken thighs ->
rotisserie chicken, chicken, poultry -> rotisserie chicken, chicken, meat
beef patty, hamburger, meat -> hamburger patty, beef, meat
oats > cereal
cheerios > cereal
** takeaways
- variants not caught, how to fix?
catalog_name = what you actually bought
product_type = reasonable substitute
category = store aisle
1. Using different categories maintains a direct comparison (product_type==spinach) and a distinction.
fresh spinach, spinach, produce
frozen spinach, spinach, frozen
include in catalog_name:
- form: frozen, fresh, ground, shredded
- fat level: whole, skim, 2%
- flavor when primary: vanilla yogurt vs plain yogurt
- cut: diced tomatoes vs crushed tomatoes
- species when relevant: gala apple vs fuji apple
exclude from catalog_name:
- package size / multipack count
- promo wording; adjectives like "premium"; retailer marketing fluff
** AC
1. fix internal search flow, add same menu
#+begin_src diff
Review 4/345: SHRP CHDR
5 matched items:
[1] KS SHRP CHDR EC20T9H5 W12T13H5 SL130 | costco | 2026-03-12 | 5.49 |
[2] KS SHRP CHDR EC20T9H5 W12T13H5 SL130 | costco | 2025-01-24 | 12.58 |
[3] KS SHRP CHDR EC20T9H5 W12T13H5 SL130 | costco | 2025-01-10 | 6.29 |
[4] KS SHRP CHDR EC20T9H5 W12T13H5 SL130 | costco | 2024-12-14 | 6.29 |
[5] KS SHRP CHDR EC20T9H5 W12T13H5 SL130 | costco | 2024-08-06 | 5.99 |
no catalog_name suggestions found
[f]ind [n]ew [s]kip e[x]clude [q]uit >
f
search: cheddar
1 search results found:
[1] cheddar cheese, cheese, dairy (0 items, 0 rows)
- selection: 1
+ [#] link to suggestion [f]ind [n]ew [s]kip e[x]clude [q]uit >
#+end_src
instead of
#+begin_src diff
search: banana
no matches found
- search again? [enter=yes, q=no]:
+ [f]ind [n]ew [s]kip e[x]clude [q]uit >
#+end_src
*

View File

@@ -1,4 +1,4 @@
#+title: Scrape-Giant Task Log
#+title: Scrape-Giant Task Log
#+STARTUP: overview
* [X] t1.1: harden giant receipt fetch cli (2-4 commits)
** acceptance criteria
@@ -763,6 +763,7 @@ enable fast lookup of catalog items during review via tokenized search and repla
- Search intentionally optimizes for manual speed rather than smart ranking: simple token overlap, max 10 rows, and immediate persistence on selection.
- Follow-up fix: search moved to `[f]ind` so `[s]kip` remains available at the main prompt.
* [ ] t1.16.2: catalog search refinement
* [ ] 1t.10: add optional llm-assisted suggestion workflow for unresolved normalized retailer items (2-4 commits)