fix: do not cache invalid html

* chg: fix: add bludv url decoding support

* chg: feat: add safeguards against future algorithim changes

api/bludv.go

@@ -152,6 +152,30 @@ func getTorrentsBluDV(ctx context.Context, i *Indexer, link string) ([]schema.In
 		magnetLinks = append(magnetLinks, magnetLink)
 	})
 
+	adwareLinks := textContent.Find("a[href^=\"https://www.seuvideo.xyz\"]")
+	adwareLinks.Each(func(_ int, s *goquery.Selection) {
+		href, _ := s.Attr("href")
+		// extract querysting "id" from url
+		parsedUrl, err := url.Parse(href)
+		if err != nil {
+			fmt.Println(err)
+			return
+		}
+		magnetLink := parsedUrl.Query().Get("id")
+		magnetLinkDecoded, err := utils.DecodeAdLink(magnetLink)
+		if err != nil {
+			fmt.Printf("failed to decode ad link \"%s\": %v\n", href, err)
+			return
+		}
+
+		// if decoded magnet link is indeed a magnet link, append it
+		if strings.HasPrefix(magnetLinkDecoded, "magnet:") {
+			magnetLinks = append(magnetLinks, magnetLinkDecoded)
+		} else {
+			fmt.Printf("WARN: link \"%s\" decoding resulted in non-magnet link: %s\n", href, magnetLinkDecoded)
+		}
+	})
+
 	var audio []schema.Audio
 	var year string
 	var size []string

requester/requester.go

@@ -10,6 +10,7 @@ import (
 	"time"
 
 	"github.com/felipemarinho97/torrent-indexer/cache"
+	"github.com/felipemarinho97/torrent-indexer/utils"
 )
 
 const (
@@ -77,8 +78,8 @@ func (i *Requster) GetDocument(ctx context.Context, url string) (io.ReadCloser,
 		fmt.Printf("request served from plain client: %s\n", url)
 	}
 
-	// save response to cache if it's not a challange and body is not empty
-	if !hasChallange(bodyByte) && len(bodyByte) > 0 {
+	// save response to cache if it's not a challange, body is not empty and is valid HTML
+	if !hasChallange(bodyByte) && len(bodyByte) > 0 && utils.IsValidHTML(string(bodyByte)) {
 		err = i.c.SetWithExpiration(ctx, key, bodyByte, i.shortLivedCacheExpiration)
 		if err != nil {
 			fmt.Printf("failed to save response to cache: %v\n", err)

utils/decoder.go

@@ -0,0 +1,28 @@
+package utils
+
+import (
+	"encoding/base64"
+	"html"
+)
+
+func DecodeAdLink(encodedStr string) (string, error) {
+	reversed := reverseString(encodedStr)
+
+	decodedBytes, err := base64.StdEncoding.DecodeString(reversed)
+	if err != nil {
+		return "", err
+	}
+
+	htmlUnescaped := html.UnescapeString(string(decodedBytes))
+
+	return htmlUnescaped, nil
+}
+
+// Helper function to reverse a string
+func reverseString(s string) string {
+	runes := []rune(s)
+	for i, j := 0, len(runes)-1; i < j; i, j = i+1, j-1 {
+		runes[i], runes[j] = runes[j], runes[i]
+	}
+	return string(runes)
+}