Crawled the first seven shops.

1 files changed, 45 insertions, 0 deletions

diff --git a/whiskyzone.go b/whiskyzone.go
new file mode 100644
index 0000000..10b996b
--- /dev/null
+++ b/whiskyzone.go
@@ -0,0 +1,45 @@
+package main
+
+import (
+	"fmt"
+	"log"
+	"regexp"
+
+	"github.com/gocolly/colly"
+)
+
+func ScrapeWhiskyzone() {
+	c := colly.NewCollector(
+		colly.AllowedDomains("whiskyzone.de"),
+		colly.AllowedDomains("www.whiskyzone.de"),
+	)
+
+	c.OnHTML(".product--info", func(e *colly.HTMLElement) {
+
+		whisky_name := e.ChildAttr("a", "title")
+		whisky_url := e.ChildAttr("a", "href")
+		log.Println(whisky_name)
+		log.Println(whisky_url)
+		price_discount_noisy := e.ChildText(".price--default")
+		price_regular_noisy := e.ChildText(".price--discount")
+
+		r, err := regexp.Compile("[0-9]+(,[0-9]{1,2})")
+		if err != nil {
+			log.Fatal(err)
+		}
+		log.Println(r.FindString(price_discount_noisy) + "€")
+		log.Println(r.FindString(price_regular_noisy) + "€")
+
+		e.ForEach(".image--media", func(i int, e *colly.HTMLElement) {
+			log.Println(e.ChildAttr("img", "src"))
+		})
+
+		fmt.Println("")
+	})
+
+	c.Visit("https://www.whiskyzone.de/widgets/emotion/index/emotionId/248/controllerName/listing")
+}
+
+func main() {
+	ScrapeWhiskyzone()
+}