feat: Implement site-specific recipe parsers for ICA and generic fallback

2026-04-12 09:39:32 +02:00
parent 2c92e07d39
commit 4e2616fe2e
5 changed files with 414 additions and 86 deletions
@@ -0,0 +1,124 @@
+import { RecipeParser, ParsedRecipe } from './base.parser';
+
+/**
+ * Parser för ica.se receptsidor
+ * Använder JSON-LD structured data som primär källa
+ */
+export class IcaRecipeParser extends RecipeParser {
+  canHandle(url: string): boolean {
+    return /ica\.se\/recept/i.test(url);
+  }
+
+  parse(html: string): ParsedRecipe {
+    console.log('[IcaParser] Parsing ICA recipe...');
+
+    // Försöka extrahera JSON-LD recipe data (ICA använder detta)
+    const jsonLdMatch = html.match(
+      /<script[^>]*type="application\/ld\+json"[^>]*>([\s\S]*?)<\/script>/i
+    );
+
+    if (jsonLdMatch) {
+      try {
+        const jsonData = JSON.parse(jsonLdMatch[1]);
+
+        // Hitta recipe-objektet
+        const recipe =
+          jsonData['@type'] === 'Recipe'
+            ? jsonData
+            : jsonData['@graph']?.find((item) => item['@type'] === 'Recipe');
+
+        if (recipe) {
+          console.log('[IcaParser] ✓ JSON-LD recipe found');
+          return this.extractFromJsonLd(recipe);
+        }
+      } catch (err) {
+        console.log('[IcaParser] JSON-LD parsing failed:', err);
+      }
+    }
+
+    // Fallback: HTML parsing (sällan nödvändigt för ICA)
+    console.log('[IcaParser] Falling back to HTML parsing');
+    return this.parseFromHtml(html);
+  }
+
+  private extractFromJsonLd(recipe: any): ParsedRecipe {
+    // Extrahera titel
+    const name = recipe.name || '';
+
+    // Extrahera ingredienser
+    const ingredients: Array<{ quantity: number; unit: string; name: string }> = [];
+    if (recipe.recipeIngredient && Array.isArray(recipe.recipeIngredient)) {
+      for (const ing of recipe.recipeIngredient) {
+        const parsed = this.parseIngredientLine(ing);
+        if (parsed) {
+          ingredients.push(parsed);
+        }
+      }
+    }
+
+    // Extrahera instruktioner
+    let instructions = '';
+    if (recipe.recipeInstructions) {
+      if (typeof recipe.recipeInstructions === 'string') {
+        instructions = recipe.recipeInstructions;
+      } else if (Array.isArray(recipe.recipeInstructions)) {
+        instructions = recipe.recipeInstructions
+          .map((step) => {
+            if (typeof step === 'string') return step;
+            if (step.text) return step.text;
+            return '';
+          })
+          .filter((s) => s)
+          .join('\n\n');
+      }
+    }
+
+    return {
+      name,
+      ingredients,
+      instructions,
+    };
+  }
+
+  private parseFromHtml(html: string): ParsedRecipe {
+    let name = '';
+    const titleMatch = html.match(/<h1[^>]*>([^<]+)<\/h1>/i);
+    if (titleMatch) {
+      name = titleMatch[1].trim();
+    }
+
+    if (!name) {
+      const ogTitleMatch = html.match(
+        /<meta\s+property="og:title"\s+content="([^"]+)"/i
+      );
+      if (ogTitleMatch) {
+        name = ogTitleMatch[1].trim();
+      }
+    }
+
+    const ingredients: Array<{ quantity: number; unit: string; name: string }> = [];
+    const ingredientRegex =
+      /<li[^>]*class="[^"]*ingredient[^"]*"[^>]*>([^<]+)<\/li>/gi;
+    let match;
+    while ((match = ingredientRegex.exec(html)) !== null) {
+      const parsed = this.parseIngredientLine(match[1]);
+      if (parsed) {
+        ingredients.push(parsed);
+      }
+    }
+
+    let instructions = '';
+    const instructionsMatch = html.match(
+      /<(?:div|section)[^>]*class="[^"]*(?:instruction|howto)[^"]*"[^>]*>([^<]*)<\/(?:div|section)>/is
+    );
+    if (instructionsMatch) {
+      instructions = instructionsMatch[1].replace(/<[^>]+>/g, '').trim();
+    }
+
+    return {
+      name,
+      ingredients,
+      instructions,
+    };
+  }
+}