fix(parser): detect Co-op via tagline, BALANCE DUE as total, strip inline qty, filter CREDIT/DEBIT

ocr/parser.py

@@ -30,6 +30,10 @@ _OCR_VARIANTS = {
     "IIDL": "LIDL",
     "COOP": "CO-OP",
     "CO OP": "CO-OP",
+    "OWNED BY YOU": "CO-OP",
+    "OWNED BY YOU.": "CO-OP",
+    "RIGHT BY YOU": "CO-OP",
+    "RIGHT BY YOU.": "CO-OP",
     "M & S": "M&S",
     "MARKS & SPENCER": "M&S",
 }
@@ -53,7 +57,7 @@ _DATE_PATTERNS = [
 _PRICE_EXTRACT_RE = re.compile(r"(-?)[£$€¥]?\s*(\d{1,6}[.,\s]\d{2})")
 
 _TOTALS_KEYWORDS = re.compile(
-    r"\b(total|sub[\s-]?total|subtotal|savings|promotions|tax|gst|hst)\b",
+    r"\b(total|sub[\s-]?total|subtotal|savings|promotions|tax|gst|hst|balance)\b",
     re.IGNORECASE,
 )
 
@@ -71,7 +75,7 @@ _EACH_RE = re.compile(r"\d\s*each\b", re.IGNORECASE)
 _INLINE_QTY_RE = re.compile(r"^(\d+)\s*x\s*[£$€¥]?(\d+[.,]\d{2})$", re.IGNORECASE)
 
 # Payment / non-item rows in the totals section
-_PAYMENT_SKIP = re.compile(r"\b(cash|change|card|visa|mastercard|amex|contactless|clubcard)\b", re.IGNORECASE)
+_PAYMENT_SKIP = re.compile(r"\b(cash|change|card|visa|mastercard|amex|contactless|clubcard|credit|debit)\b", re.IGNORECASE)
 
 # Row merging tolerance in pixels
 _ROW_Y_TOLERANCE = 30
@@ -390,6 +394,11 @@ def _match_known_retailer(header_rows: list[list[dict]]) -> tuple[str | None, fl
                 return text, _top_y(block["bbox"])
             if text in _OCR_VARIANTS:
                 return _OCR_VARIANTS[text], _top_y(block["bbox"])
+        row_text = " ".join(b["text"].strip() for b in row).strip().upper()
+        if row_text in _KNOWN_RETAILERS:
+            return row_text, min(_top_y(b["bbox"]) for b in row)
+        if row_text in _OCR_VARIANTS:
+            return _OCR_VARIANTS[row_text], min(_top_y(b["bbox"]) for b in row)
     return None, 0.0
 
 
@@ -489,6 +498,13 @@ def _append_desc(item: dict, desc_blocks: list[dict], receipt_width: float) -> N
         if _is_qty_block(b, receipt_width):
             item["quantity"] = int(text)
             continue
+        # Co-op style: qty embedded in description ("1 BATCH S/NOODLE B")
+        if not item["description"] and not parts:
+            m_qty = re.match(r"^(\d{1,2})\s+([A-Za-z].+)$", text)
+            if m_qty:
+                item["quantity"] = int(m_qty.group(1))
+                parts.append(m_qty.group(2).strip())
+                continue
         if _DISCOUNT_PREFIX_RE.match(text):
             continue
         if _EACH_RE.search(text):
@@ -550,6 +566,8 @@ def _extract_totals(totals_rows: list[list[dict]]) -> dict[str, str | None]:
             result["subtotal"] = result["subtotal"] or price
         elif "total" in row_text and "sub" not in row_text and "card" not in row_text:
             result["total"] = result["total"] or price
+        elif "balance" in row_text:
+            result["total"] = result["total"] or price
         elif result["total"] is None and not price.startswith("-"):
             # Standalone positive price with no keyword — treat as total if
             # not yet set (handles receipts where OCR misses the "TOTAL" text).