package novelfire import ( "context" "testing" "github.com/libnovel/scraper/internal/scraper" ) // rankingPage1HTML is a realistic mock of the popular genre listing page // (novelfire.net/genre-all/sort-popular/status-all/all-novel?page=1). // It uses the real novelfire.net DOM:
  • cards with //

    and a rel="next" pagination link. func rankingPage1HTML() string { return ` ` } func rankingPage2HTML() string { return ` ` } // drainRanking collects all entries from a ScrapeRanking call without // deadlocking. It uses "for A != nil || B != nil" — nil channels are never // selected, so setting one to nil effectively removes it from the select. func drainRanking(t *testing.T, entryCh <-chan scraper.BookMeta, errCh <-chan error) []scraper.BookMeta { t.Helper() var entries []scraper.BookMeta for entryCh != nil || errCh != nil { select { case meta, ok := <-entryCh: if !ok { entryCh = nil } else { entries = append(entries, meta) } case err, ok := <-errCh: if !ok { errCh = nil } else if err != nil { t.Fatalf("unexpected scrape error: %v", err) } } } return entries } // TestScrapeRanking_SinglePage verifies a single page is parsed into entries // with sequential Ranking numbers using a stub client. // ScrapeRanking uses s.client (the main client, not urlClient) because the // ranking page is fully server-rendered. func TestScrapeRanking_SinglePage(t *testing.T) { // newScraper passes the stub as s.client — exactly what ScrapeRanking uses. s := newScraper(rankingPage1HTML()) entryCh, errCh := s.ScrapeRanking(context.Background(), 1) entries := drainRanking(t, entryCh, errCh) if len(entries) != 2 { t.Fatalf("expected 2 entries, got %d", len(entries)) } if entries[0].Ranking != 1 || entries[0].Title != "The Iron Throne" { t.Errorf("entry[0]: got rank=%d title=%q, want rank=1 title=%q", entries[0].Ranking, entries[0].Title, "The Iron Throne") } if entries[1].Ranking != 2 || entries[1].Title != "Shadow Mage" { t.Errorf("entry[1]: got rank=%d title=%q, want rank=2 title=%q", entries[1].Ranking, entries[1].Title, "Shadow Mage") } } // TestScrapeRanking_MultiPage verifies pagination across two pages yields // contiguous rank numbers (1, 2, 3). func TestScrapeRanking_MultiPage(t *testing.T) { // Use pagedStubClient for s.client so each GetContent call returns the // next page. ScrapeRanking now calls s.client directly. urlClient := &pagedStubClient{pages: []string{rankingPage1HTML(), rankingPage2HTML()}} s := New(urlClient, nil, nil, nil, nil) // nil cache — no disk I/O in tests entryCh, errCh := s.ScrapeRanking(context.Background(), 0) // 0 = all pages entries := drainRanking(t, entryCh, errCh) if len(entries) != 3 { t.Fatalf("expected 3 entries across 2 pages, got %d", len(entries)) } want := []struct { rank int title string }{ {1, "The Iron Throne"}, {2, "Shadow Mage"}, {3, "Void Hunter"}, } for i, w := range want { if entries[i].Ranking != w.rank || entries[i].Title != w.title { t.Errorf("entry[%d]: got rank=%d title=%q, want rank=%d title=%q", i, entries[i].Ranking, entries[i].Title, w.rank, w.title) } } } // TestScrapeRanking_EmptyPage verifies that a page with no .novel-item // cards produces zero entries and closes channels cleanly (no deadlock). func TestScrapeRanking_EmptyPage(t *testing.T) { s := newScraper(`
    `) entryCh, errCh := s.ScrapeRanking(context.Background(), 1) entries := drainRanking(t, entryCh, errCh) if len(entries) != 0 { t.Errorf("expected 0 entries for empty page, got %d", len(entries)) } }