goyomi/sources/base/hotcomics/hotcomics.go

// Package hotcomics implements the HotComics manga base.
// HTML scraping; popular: GET {base}/en; requires cookie hc_vfs=Y.
package hotcomics

import (
	"context"
	"fmt"
	"net/http"
	"strings"
	"time"

	"github.com/PuerkitoBio/goquery"

	"goyomi/internal/httpclient"
	"goyomi/internal/source"
	"goyomi/sources/base/util"
)

type Config struct {
	Name    string
	BaseURL string
	Lang    string
}

type Source struct {
	cfg    Config
	client *httpclient.Client
	id     int64
}

func New(cfg Config) *Source {
	c := httpclient.NewClient(httpclient.WithRateLimit(1, 2))
	return &Source{cfg: cfg, client: c, id: source.GenerateSourceID(cfg.Name, cfg.Lang)}
}

func (s *Source) ID() int64            { return s.id }
func (s *Source) Name() string         { return s.cfg.Name }
func (s *Source) Lang() string         { return s.cfg.Lang }
func (s *Source) SupportsLatest() bool { return true }

func (s *Source) base() string { return strings.TrimRight(s.cfg.BaseURL, "/") }

func (s *Source) get(ctx context.Context, rawURL string) (*goquery.Document, error) {
	req, err := http.NewRequestWithContext(ctx, http.MethodGet, rawURL, nil)
	if err != nil {
		return nil, err
	}
	req.Header.Set("Referer", s.cfg.BaseURL+"/")
	// Inject required cookie
	req.AddCookie(&http.Cookie{Name: "hc_vfs", Value: "Y"})
	resp, err := s.client.Do(req)
	if err != nil {
		return nil, err
	}
	defer resp.Body.Close()
	if resp.StatusCode != http.StatusOK {
		return nil, fmt.Errorf("hotcomics: HTTP %d", resp.StatusCode)
	}
	return goquery.NewDocumentFromReader(resp.Body)
}

func imgAttr(img *goquery.Selection) string {
	if v, ok := img.Attr("data-src"); ok && v != "" {
		return v
	}
	v, _ := img.Attr("src")
	return v
}

func (s *Source) parseMangaList(doc *goquery.Document) source.MangasPage {
	var mangas []source.SManga
	seen := map[string]bool{}
	doc.Find("li[itemtype*=ComicSeries]:not(.no-comic) > a").Each(func(_ int, a *goquery.Selection) {
		m := source.SManga{}
		m.URL, _ = a.Attr("href")
		if m.URL == "" || seen[m.URL] {
			return
		}
		seen[m.URL] = true
		a.Find("div.visual img").First().Each(func(_ int, img *goquery.Selection) {
			m.ThumbnailURL = imgAttr(img)
		})
		m.Title = strings.TrimSpace(a.Find("div.main-text > h4.title").Text())
		if m.Title != "" {
			mangas = append(mangas, m)
		}
	})
	hasNext := doc.Find("div.pagination a.vnext:not(.disabled)").Length() > 0
	return source.MangasPage{Mangas: mangas, HasNextPage: hasNext}
}

func (s *Source) GetPopularManga(page int) (source.MangasPage, error) {
	doc, err := s.get(context.Background(), s.base()+"/en")
	if err != nil {
		return source.MangasPage{}, err
	}
	return s.parseMangaList(doc), nil
}

func (s *Source) GetLatestUpdates(page int) (source.MangasPage, error) {
	doc, err := s.get(context.Background(), s.base()+"/en/new")
	if err != nil {
		return source.MangasPage{}, err
	}
	return s.parseMangaList(doc), nil
}

func (s *Source) GetSearchManga(page int, query string, filters []source.Filter) (source.MangasPage, error) {
	var u string
	if query != "" {
		u = fmt.Sprintf("%s/en/search?keyword=%s", s.base(), query)
	} else {
		u = fmt.Sprintf("%s/en?page=%d", s.base(), page)
	}
	doc, err := s.get(context.Background(), u)
	if err != nil {
		return source.MangasPage{}, err
	}
	return s.parseMangaList(doc), nil
}

func (s *Source) GetMangaDetails(manga source.SManga) (source.SManga, error) {
	doc, err := s.get(context.Background(), util.AbsURL(s.cfg.BaseURL, manga.URL))
	if err != nil {
		return manga, err
	}
	result := source.SManga{URL: manga.URL}
	result.Title = strings.TrimSpace(doc.Find("h2.episode-title").Text())
	if result.Title == "" {
		result.Title = manga.Title
	}
	typeBox := doc.Find("p.type_box")
	result.Author = strings.TrimSpace(
		strings.TrimPrefix(typeBox.Find("span.writer").Text(), "ⓒ"))
	var genres []string
	typeBox.Find("span.type").First().Each(func(_ int, el *goquery.Selection) {
		for _, g := range strings.Split(el.Text(), "/") {
			if g = strings.TrimSpace(g); g != "" {
				genres = append(genres, g)
			}
		}
	})
	result.Genre = strings.Join(genres, ", ")
	switch typeBox.Find("span.date").Text() {
	case "End", "Ende":
		result.Status = source.StatusCompleted
	case "":
		result.Status = source.StatusUnknown
	default:
		result.Status = source.StatusOngoing
	}
	var descParts []string
	doc.Find("div.episode-contents header").First().Each(func(_ int, el *goquery.Selection) {
		if t := strings.TrimSpace(el.Text()); t != "" {
			descParts = append(descParts, t)
		}
	})
	doc.Find("div.title_content > h2:not(.episode-title)").First().Each(func(_ int, el *goquery.Selection) {
		if t := strings.TrimSpace(el.Text()); t != "" {
			descParts = append(descParts, t)
		}
	})
	result.Description = strings.Join(descParts, "\n\n")
	return result, nil
}

func (s *Source) GetChapterList(manga source.SManga) ([]source.SChapter, error) {
	doc, err := s.get(context.Background(), util.AbsURL(s.cfg.BaseURL, manga.URL))
	if err != nil {
		return nil, err
	}
	var chapters []source.SChapter
	doc.Find("#tab-chapter a").Each(func(_ int, a *goquery.Selection) {
		onclick := a.AttrOr("onclick", "")
		// onclick="popupLogin('/chapter/url')"
		u := strings.TrimSuffix(strings.TrimPrefix(
			strings.TrimPrefix(onclick, "popupLogin('"), "popupLogin(\""), "'")
		u = strings.TrimSuffix(u, "\")")
		u = strings.TrimSuffix(u, "')")
		if u == "" || u == onclick {
			return
		}
		name := strings.TrimSpace(a.Find(".cell-num").Text())
		dateStr := strings.TrimSpace(a.Find(".cell-time").Text())
		chapters = append(chapters, source.SChapter{
			URL:        u,
			Name:       name,
			DateUpload: parseDate(dateStr),
		})
	})
	// reverse: oldest first
	for i, j := 0, len(chapters)-1; i < j; i, j = i+1, j-1 {
		chapters[i], chapters[j] = chapters[j], chapters[i]
	}
	return chapters, nil
}

func parseDate(s string) int64 {
	if s == "" {
		return 0
	}
	t, err := time.Parse("Jan 02, 2006", s)
	if err != nil {
		return 0
	}
	return t.UnixMilli()
}

func (s *Source) GetPageList(chapter source.SChapter) ([]source.Page, error) {
	doc, err := s.get(context.Background(), util.AbsURL(s.cfg.BaseURL, chapter.URL))
	if err != nil {
		return nil, err
	}
	var pages []source.Page
	doc.Find("#viewer-img img").Each(func(i int, img *goquery.Selection) {
		u := imgAttr(img)
		if u != "" {
			if !strings.HasPrefix(u, "http") {
				u = util.AbsURL(s.cfg.BaseURL, u)
			}
			pages = append(pages, source.Page{Index: i, ImageURL: u})
		}
	})
	return pages, nil
}

func (s *Source) GetImageURL(page source.Page) (string, error) { return page.ImageURL, nil }
func (s *Source) GetFilterList() []source.Filter               { return nil }