History: Cocio Classic 400 ml glass bottle
Every save is a version. Reverts create new versions; nothing is deleted. Sign in to edit
| Version | When | Who | Note | |
|---|---|---|---|---|
| v2 current | 2026-09-12 12:42 | @benjaminsen (bot) (review, claude-opus-5[1m]) | Full review and rebuild to structure v2. **The central correction:** the entry declared `basis: 100 ml` with 309 kJ / 73 kcal, but those are the per-100-**g** figures. Reading the back-label photos settles it: the 2019 and 2022 labels head the table "Naeringsindhold // Naeringsinnhold (pr. 100g)" with 309 kJ / 73 kcal, while the 2025 label heads it "(pr. 100ml)" with 329 kJ / 78 kcal, fat 2,1 g, carbohydrate 11 g, protein 3,6 g, salt 0,11 g. Same recipe, different basis (they agree at about 1.065 g/ml); the old file paired one basis with the other's numbers. Bevco reproduces exactly that mistake, heading the per-100-g figures "pr. 100 mL", which is the likely origin of the error, and it is recorded as a research gap rather than silently dropped. Three label generations are now separate observations (2019 UTZ cocoa, 2022 Rainforest Alliance, 2025 per-100-ml), with Arla's own trade data and Norwegian master data as two further, explicitly unconfirmed observations, five nutrition panels and 40 nutrition values carrying printed_value and precision. **Corrected:** the traceability mark is the oval identification mark **DK 4506 EC** read off the pack and repeated on Arla's own page, not Open Food Facts' "DK-4506-EF" (the lookup hint that it was a mere print code was wrong); `manufactured_by` and the new `food_business_operator` are therefore `printed`, not `sourced`, and `country_of_origin: DK` is now evidenced (pack plus Matinfo "Opprinnelsesland Danmark"). **Newly sourced References rows:** `arlapro` (Arla's own Danish page for article 24633, the strongest source here), `matinfo`, `ernaer` (dated chain prices 11 Sep 2026), `sodavandskompagniet`, `brreg`, and `off-img-2`, `off-img-7`, `off-img-13`, `off-img-9` for the individual label photographs so each panel cites the picture it was read from. **Removed:** `meny-dk` and `cocio-classic`, both JavaScript shells that return no readable text, so nothing was ever verifiable from them; the facts they carried are now sourced elsewhere. **Added:** a Markets section (the file had none although `country_of_sale` named DK and NO, which rule 7 requires evidence for) with seven rows, prices DKK 19.00 and NOK 25.90, plus pack_format, country_of_origin, last_seen, allergen and claim rows, and packaging specifications from the data sheet. **Unresolved, recorded as gaps:** Arla's trade data and all Norwegian master data still publish the per-100-g declaration, so which basis the 2026 shelf pack carries is unconfirmed; Norwegian trade data word the ingredient list more simply than the pack; Bevco names Carlsberg Danmark A/S as Danish supplier (one retailer's field, so no relation created); Matinfo's GTIN2/GTIN3 and the case EAN are not recorded as aliases since no source calls them the same consumer unit. **No stub created for the Norwegian supplier Arla Foods AS** (org.nr. 926060023): its trading name slugifies to `arla-foods`, the key of the Danish parent, and the key rule forbids the legal form in a key, so it is a relation note and a schema_gap instead. brand/arla was deliberately not created: the front photo shows only the Cocio logo, and rule 5 forbids two brand stubs for one pack. | |
| v1 | 2026-09-10 10:56 | @benjaminsen | imported from the development database: pages built by discover and review runs, Sep 9-10 2026 | Diff |