done●
Total stems calculated (300) does not match expected total stems (700).; Total amount calculated (230.00) does not match expected total amount (320.00).
done●
Несовпадение количества коробок; Несовпадение количества стеблей; Несовпадение суммы; Не удалось извлечь total_price для 2 item(s).; Не удалось извлечь quantity для 0 item(s).; Не удалось извлечь unit_price для 0 item(s).
done●
invoice_number: Value was present in OCR data but rule explicitly stated to leave empty.; invoice_date: Value was present in OCR data but rule explicitly stated to leave empty.; total_stems: Calculated sum (560) does not match expected (700).; total_amount: Calculated sum (269.8) does not match expected (320.0).; Item data parsing: Columns shifted for some rows (e.g., rows 33, 38, 40, 41 in raw table), leading to incorrect 'box_marking', 'plantation', 'variety', 'length', 'quantity', 'unit_price', 'total_price' based on strict column index rules.
FACT EXPORTACION 0255502 TDA - OOO TRADE HOUSE AZALIA H 17_29_50.pdf
—
—
invoice
done●
invoice_number, invoice_date, supplier, country, awb fields could not be extracted as per rules (metadata/notes were empty).; OCR data indicates a mixed box structure with multiple length columns per logical item, directly contradicting the 'flat' invoice type rule. Each row was treated as a single item/box as per the rule, but original OCR '# BOX' values sometimes denote multiple physical boxes (e.g., '11-14'). This resulted in more processed items/boxes (54) than indicated by summary table (8 boxes).; Semantic mismatch: Rule 'колонка 0 = box_marking' mapped to `row[0]` which contains `# BOX` numbers (e.g., '1', '2'), while `row[1]` ('HAL') seems more like a marking. This was applied as `box_marking` from `row[1]` and `# BOX` from `row[0]` was used for `box_number` generation logic.; Semantic mismatch: Rule 'колонка 2 = category' mapped to `row[2]` which contains Variety names (e.g., 'ESPERANCE'). Rule 'колонка 3 = variety' mapped to `row[3]` which contains 'STxB' (e.g., '25'). Data was interpreted as `category` from `row[2]` and `variety` from `row[3]` per rule, resulting in variety names appearing in 'category' and 'STxB' values in 'variety'.; Rule 'колонка 4 = length' mapped to `row[4]` ('X'), which is not a length. Lengths were extracted from corresponding data in columns `row[5]`-`row[15]` based on header match.; Rule 'колонка 5 = quantity', 'колонка 6 = unit_price', 'колонка 7 = total_price' had incorrect column indices relative to the OCR table structure. Actual data for these fields was found at `row[17]`, `row[18]`, `row[19]` respectively and extracted semantically.; Quantity validation: All extracted quantities are integers.; Unit price validation: All extracted unit prices are within the $0.20-$5.00 range.