chore: update flyer import features and resources

- Remove outdated Willys flyer PDF (0001-0008_WIL_V21_ED1pdf.pdf) - Add new Willys flyer PDF (willys_reklamblad.pdf) - Improve offer detection logic in backend flyer-import service - Add offer limit text extraction and sanitization in Flutter UI - Fix Swedish character encoding issues in UI text
2026-05-18 23:40:05 +02:00
parent c720f611ea
commit 0ce1db5471
3 changed files with 159 additions and 51 deletions
@@ -28,12 +28,20 @@ type FlyerParseItem = {
  reasonCodes: string[];
 };

-type FlyerParseResponse = {
-  retailer: 'willys';
-  parserVersion: 'v1';
-  items: FlyerParseItem[];
-  warnings: string[];
-};
+type FlyerParseResponse = {
+  retailer: 'willys';
+  parserVersion: 'v1';
+  items: FlyerParseItem[];
+  warnings: string[];
+};
+
+type ExtractedOfferSignals = {
+  price: number | null;
+  priceUnit: string | null;
+  comparisonPrice: number | null;
+  comparisonUnit: string | null;
+  hasCampaignPattern: boolean;
+};

 type ProductLite = {
  id: number;
@@ -77,23 +85,28 @@ export class FlyerImportService {
      productById.set(product.id, product);
    }

-    const items: FlyerImportItem[] = parsed.items.map((item) => {
-      const match = this.matchItem(item, products, aliasToProduct, productById);
-      const offerLimitText = this.extractOfferLimitText(item.offerText);
-      return {
-        flyerItemId: null,
-        rawName: item.rawName,
-        normalizedName: item.normalizedName,
-        category: item.category,
-        price: item.price,
-        priceUnit: item.priceUnit,
-        comparisonPrice: item.comparisonPrice,
-        comparisonUnit: item.comparisonUnit,
-        offerText: item.offerText,
-        isOffer: this.isOfferItem(item),
-        offerLimitText,
-        parseConfidence: item.confidence,
-        parseReasons: item.reasonCodes,
+    const items: FlyerImportItem[] = parsed.items.map((item) => {
+      const match = this.matchItem(item, products, aliasToProduct, productById);
+      const signals = this.extractOfferSignals(item.offerText);
+      const price = item.price ?? signals.price;
+      const priceUnit = this.normalizeUnit(item.priceUnit) ?? signals.priceUnit;
+      const comparisonPrice = item.comparisonPrice ?? signals.comparisonPrice;
+      const comparisonUnit = this.normalizeUnit(item.comparisonUnit) ?? signals.comparisonUnit;
+      const offerLimitText = this.extractOfferLimitText(item.offerText);
+      return {
+        flyerItemId: null,
+        rawName: item.rawName,
+        normalizedName: item.normalizedName,
+        category: item.category,
+        price,
+        priceUnit,
+        comparisonPrice,
+        comparisonUnit,
+        offerText: item.offerText,
+        isOffer: this.isOfferItem(item, signals.hasCampaignPattern),
+        offerLimitText,
+        parseConfidence: item.confidence,
+        parseReasons: item.reasonCodes,
        matchedProductId: match.product?.id ?? null,
        matchedProductName: match.product?.name ?? null,
        matchedVia: match.via,
@@ -260,28 +273,103 @@ export class FlyerImportService {
    return intersection / union;
  }

-  private isOfferItem(item: FlyerParseItem): boolean {
-    return item.price != null || item.comparisonPrice != null || !!item.offerText?.trim();
-  }
+  private isOfferItem(item: FlyerParseItem, hasCampaignPattern: boolean): boolean {
+    return (
+      item.price != null
+      || item.comparisonPrice != null
+      || !!item.offerText?.trim()
+      || hasCampaignPattern
+    );
+  }

  private extractOfferLimitText(offerText: string | null): string | null {
    if (!offerText) return null;

-    const normalized = offerText.replace(/\s+/g, ' ' ).trim();
-    if (!normalized) return null;
-
-    const limitMatch = normalized.match(/(?:max|högst)\s+[^,.;]+(?:hushåll|kund)?/i);
-    if (limitMatch?.[0]) {
-      return limitMatch[0].trim();
-    }
-
-    const householdMatch = normalized.match(/[^,.;]*(?:hushåll|kund)[^,.;]*/i);
-    if (householdMatch?.[0]) {
-      return householdMatch[0].trim();
-    }
-
-    return null;
-  }
+    const normalized = offerText.replace(/\s+/g, ' ').trim();
+    if (!normalized) return null;
+
+    const limitMatch = normalized.match(
+      /(?:max|högst|begränsat\s+antal)\s+[^,.;]*(?:hushåll|kund|köp)?(?:\s*\/\s*(?:hushåll|kund))?/i,
+    );
+    if (limitMatch?.[0]) {
+      return limitMatch[0].trim();
+    }
+
+    const perCustomerMatch = normalized.match(
+      /[^,.;]*(?:per\s+(?:hushåll|kund)|\/\s*(?:hushåll|kund))[^,.;]*/i,
+    );
+    if (perCustomerMatch?.[0]) {
+      return perCustomerMatch[0].trim();
+    }
+
+    const householdMatch = normalized.match(/[^,.;]*(?:hushåll|kund)[^,.;]*/i);
+    if (householdMatch?.[0]) {
+      return householdMatch[0].trim();
+    }
+
+    return null;
+  }
+
+  private extractOfferSignals(offerText: string | null): ExtractedOfferSignals {
+    const empty: ExtractedOfferSignals = {
+      price: null,
+      priceUnit: null,
+      comparisonPrice: null,
+      comparisonUnit: null,
+      hasCampaignPattern: false,
+    };
+
+    if (!offerText?.trim()) return empty;
+
+    const normalized = offerText.replace(/\s+/g, ' ').trim().toLowerCase();
+    const campaignPattern = /(\b\d+\s*för\s*\d+[,.:]?\d*\b)|(ta\s*\d+\s*betala\s*för\s*\d+)/i;
+    const priceWithUnit = normalized.match(/(\d{1,3}[:.,]\d{2}|\d{1,3})\s*(?:kr)?\s*\/?\s*(kg|hg|g|l|dl|cl|ml|st|styck|pkt|förp|fp)/i);
+    const priceOnly = normalized.match(/(\d{1,3}[:.,]\d{2}|\d{1,3})\s*kr\b/i);
+    const comparison = normalized.match(
+      /(?:jfr\s*pris|jamforpris|jämförpris|jfr)\s*[:]?\s*(\d{1,3}[:.,]\d{2}|\d{1,3})\s*(?:kr)?\s*\/?\s*(kg|hg|g|l|dl|cl|ml|st|styck|pkt|förp|fp)/i,
+    );
+
+    const signals: ExtractedOfferSignals = {
+      ...empty,
+      hasCampaignPattern: campaignPattern.test(normalized),
+    };
+
+    if (priceWithUnit) {
+      signals.price = this.parseSwedishPrice(priceWithUnit[1]);
+      signals.priceUnit = this.normalizeUnit(priceWithUnit[2]);
+    } else if (priceOnly) {
+      signals.price = this.parseSwedishPrice(priceOnly[1]);
+    }
+
+    if (comparison) {
+      signals.comparisonPrice = this.parseSwedishPrice(comparison[1]);
+      signals.comparisonUnit = this.normalizeUnit(comparison[2]);
+    }
+
+    return signals;
+  }
+
+  private parseSwedishPrice(value: string | null | undefined): number | null {
+    if (!value) return null;
+    const normalized = value.trim().replace(':', '.').replace(',', '.');
+    const parsed = Number.parseFloat(normalized);
+    if (!Number.isFinite(parsed)) return null;
+    return parsed;
+  }
+
+  private normalizeUnit(unit: string | null | undefined): string | null {
+    if (!unit) return null;
+    const cleaned = unit.trim().toLowerCase().replace(/\./g, '');
+    if (!cleaned) return null;
+
+    if (cleaned === 'styck') return 'st';
+    if (cleaned === 'fp' || cleaned === 'forp' || cleaned === 'förp' || cleaned === 'pkt') {
+      return 'pkt';
+    }
+
+    const allowed = new Set(['kg', 'hg', 'g', 'l', 'dl', 'cl', 'ml', 'st', 'pkt']);
+    return allowed.has(cleaned) ? cleaned : cleaned;
+  }

  private async parseViaImporter(file: Express.Multer.File): Promise<FlyerParseResponse> {
    const form = new FormData();