[Billa] Ignore app when in cyrillic

StefanBratanov · Sep 5, 2023 · 81b6eb0 · 81b6eb0
1 parent 3cd96dd
commit 81b6eb0
Show file tree

Hide file tree

Showing 2 changed files with 11 additions and 3 deletions.
diff --git a/src/main/kotlin/com/stefanbratanov/sofiasupermarketsapi/extractors/BillaProductsExtractor.kt b/src/main/kotlin/com/stefanbratanov/sofiasupermarketsapi/extractors/BillaProductsExtractor.kt
@@ -43,9 +43,7 @@ class BillaProductsExtractor : UrlProductsExtractor {
     )
 
   private val regexesToDeleteBilla =
-    listOf(
-      "с (Billa|билла|била) (card|app)".toRegex(IGNORE_CASE),
-    )
+    listOf("с (Billa|билла|била) (card|app|арр)".toRegex(IGNORE_CASE))
 
   override fun extract(url: URL): List<Product> {
     log.info("Processing Billa URL: {}", url.toString())

diff --git a/src/test/resources/extractors/billa/input.html b/src/test/resources/extractors/billa/input.html
@@ -3685,6 +3685,16 @@
       <div style="float:left; width:15%">
         <span class="price">11.995</span><span class="currency">лв.</span></div>
     </div>
+    <div class="product">
+      <div class="actualProduct" style="float:left; width:25%; margin-right:2%;">Само с BILLA Арр Ракия Троянска сливова отлежала 0.7л Произход - България</div>
+      <div class="priceText" style="float:left; width:10%;">СТАРА<br>ЦЕНА</div>
+      <div style="float:left; width:22%">
+        <span class="price">19.99</span> <span class="currency">лв.</span>	</div>
+      <div class="priceText" style="float:left; width:10%;">НОВА<br>ЦЕНА</div>
+      <div class="discount" style="float:left; width:10%;"> - -30%</div>
+      <div style="float:left; width:15%">
+        <span class="price">13.99</span><span class="currency">лв.</span>	</div>
+    </div>
   </div>
 </div>