Fix browse filters having no effect due to filter-agnostic cache key

The MinIO cache key only encoded page number, so all filter combinations
hit the same cache entry and returned unfiltered results. Introduce
BrowseFilteredHTMLKey() that encodes sort/genre/status into the key
(e.g. novelfire.net/html/new-isekai-completed/page-1.html); falls back
to the original page-only key for default filters to preserve existing
cached pages.

scraper/internal/server/handlers_browse.go

@@ -80,7 +80,7 @@ func (s *Server) handleBrowse(w http.ResponseWriter, r *http.Request) {
 	defer cancel()
 
 	// ── Cache-first: try MinIO snapshot (new key layout) ─────────────────
-	cacheKey := s.store.BrowseHTMLKey(novelFireDomain, pageNum)
+	cacheKey := s.store.BrowseFilteredHTMLKey(novelFireDomain, pageNum, sortBy, genre, status)
 	if html, ok, err := s.store.GetBrowsePage(ctx, cacheKey); err == nil && ok && len(html) > 0 {
 		novels, hasNext := parseBrowsePage(strings.NewReader(html))
 		s.log.Debug("browse: served from cache", "key", cacheKey)