fix: resolve Google News URLs to actual source articles before fetching link previews

go-backend/internal/services/link_preview_service.go

@@ -98,6 +98,14 @@ func (s *LinkPreviewService) FetchPreview(ctx context.Context, rawURL string, tr
 		return nil, fmt.Errorf("empty URL")
 	}
 
+	// Resolve Google News URLs to the actual source article
+	if strings.Contains(rawURL, "news.google.com") {
+		resolved := ResolveGoogleNewsURL(rawURL)
+		if resolved != rawURL {
+			rawURL = resolved
+		}
+	}
+
 	parsed, err := url.Parse(rawURL)
 	if err != nil {
 		return nil, fmt.Errorf("invalid URL: %w", err)

go-backend/internal/services/official_accounts_service.go

@@ -7,6 +7,7 @@ import (
 	"fmt"
 	"io"
 	"net/http"
+	"regexp"
 	"strings"
 	"sync"
 	"time"
@@ -271,35 +272,101 @@ func (s *OfficialAccountsService) FetchRSS(ctx context.Context, rssURL string) (
 	return feed.Channel.Items, nil
 }
 
-// resolveGoogleNewsLink follows the Google News redirect to get the actual article URL.
+// resolveGoogleNewsLink follows the Google News redirect chain to get the actual article URL.
 func (s *OfficialAccountsService) resolveGoogleNewsLink(googleURL string) string {
+	return ResolveGoogleNewsURL(googleURL)
+}
+
+// ResolveGoogleNewsURL is a package-level helper that resolves a Google News URL
+// to the underlying article URL by following the full redirect chain.
+func ResolveGoogleNewsURL(googleURL string) string {
 	if googleURL == "" || !strings.Contains(googleURL, "news.google.com") {
 		return googleURL
 	}
 
-	// Use a client that does NOT follow redirects
-	noRedirectClient := &http.Client{
-		Timeout: 10 * time.Second,
+	// Track the final URL after all redirects
+	var finalURL string
+	client := &http.Client{
+		Timeout: 15 * time.Second,
 		CheckRedirect: func(req *http.Request, via []*http.Request) error {
+			if len(via) >= 10 {
+				return fmt.Errorf("too many redirects")
+			}
+			// Track every hop; stop if we've left Google domains
+			host := req.URL.Hostname()
+			if !strings.Contains(host, "google.com") && !strings.Contains(host, "google.") {
+				finalURL = req.URL.String()
 				return http.ErrUseLastResponse
+			}
+			return nil
 		},
 	}
 
-	resp, err := noRedirectClient.Get(googleURL)
+	req, err := http.NewRequest("GET", googleURL, nil)
 	if err != nil {
+		return googleURL
+	}
+	req.Header.Set("User-Agent", "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36")
+	req.Header.Set("Accept", "text/html,application/xhtml+xml")
+
+	resp, err := client.Do(req)
+	if err != nil {
+		// If we captured a non-Google URL before the error, use it
+		if finalURL != "" {
+			log.Debug().Str("resolved", finalURL).Str("original", googleURL).Msg("Resolved Google News link via redirect chain")
+			return finalURL
+		}
 		log.Debug().Err(err).Str("url", googleURL).Msg("Failed to resolve Google News link")
 		return googleURL
 	}
 	defer resp.Body.Close()
 
-	if resp.StatusCode >= 300 && resp.StatusCode < 400 {
-		loc := resp.Header.Get("Location")
-		if loc != "" {
-			return loc
+	// If we captured a non-Google URL during redirects, use that
+	if finalURL != "" {
+		log.Debug().Str("resolved", finalURL).Str("original", googleURL).Msg("Resolved Google News link via redirect chain")
+		return finalURL
+	}
+
+	// Check final response URL
+	if resp.Request != nil && resp.Request.URL != nil {
+		final := resp.Request.URL.String()
+		if !strings.Contains(final, "news.google.com") && !strings.Contains(final, "consent.google") {
+			log.Debug().Str("resolved", final).Str("original", googleURL).Msg("Resolved Google News link via final URL")
+			return final
 		}
 	}
 
-	// Some Google News links use JS-based redirect; try parsing from body as fallback
+	// Fallback: parse meta refresh or JS redirect from response body
+	body, err := io.ReadAll(io.LimitReader(resp.Body, 64*1024))
+	if err == nil {
+		html := string(body)
+		// Look for <meta http-equiv="refresh" content="0;url=...">
+		metaRe := regexp.MustCompile(`(?i)http-equiv\s*=\s*["']refresh["'][^>]*content\s*=\s*["'][^;]*;\s*url\s*=\s*([^"'\s>]+)`)
+		if m := metaRe.FindStringSubmatch(html); len(m) > 1 {
+			resolved := strings.TrimSpace(m[1])
+			if !strings.Contains(resolved, "google.com") {
+				log.Debug().Str("resolved", resolved).Str("original", googleURL).Msg("Resolved Google News link via meta refresh")
+				return resolved
+			}
+		}
+		// Look for window.location or document.location JS redirects
+		jsRe := regexp.MustCompile(`(?:window|document)\.location\s*[=]\s*["']([^"']+)["']`)
+		if m := jsRe.FindStringSubmatch(html); len(m) > 1 {
+			resolved := strings.TrimSpace(m[1])
+			if !strings.Contains(resolved, "google.com") {
+				log.Debug().Str("resolved", resolved).Str("original", googleURL).Msg("Resolved Google News link via JS redirect")
+				return resolved
+			}
+		}
+		// Look for <a href="..."> with data-n-au attribute (Google News article link)
+		auRe := regexp.MustCompile(`data-n-au\s*=\s*["']([^"']+)["']`)
+		if m := auRe.FindStringSubmatch(html); len(m) > 1 {
+			log.Debug().Str("resolved", m[1]).Str("original", googleURL).Msg("Resolved Google News link via data-n-au")
+			return m[1]
+		}
+	}
+
+	log.Warn().Str("url", googleURL).Msg("Could not resolve Google News link to source article")
 	return googleURL
 }