goyomi/sources/base/liliana/liliana.go

// Package liliana implements the Liliana manga base.
// GET {base}/ranking/week/{page} for popular; FlareSolverr required.
package liliana

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

	"github.com/PuerkitoBio/goquery"

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

type Config struct {
	Name    string
	BaseURL string
	Lang    string
}

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

func New(cfg Config) *Source {
	opts := []flare.Option{flare.WithRateLimit(1, 2)}
	c := flare.NewClient(opts...)
	return &Source{cfg: cfg, client: c, id: source.GenerateSourceID(cfg.Name, cfg.Lang)}
	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+"/")
	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("liliana: HTTP %d", resp.StatusCode)
	}
	return goquery.NewDocumentFromReader(resp.Body)
}

func (s *Source) parseMangaList(doc *goquery.Document) source.MangasPage {
	var mangas []source.SManga
	doc.Find("div#main div.grid > div, div.listmanga div.col").Each(func(_ int, el *goquery.Selection) {
		m := source.SManga{}
		el.Find(".text-center a, .manga-title a").First().Each(func(_ int, a *goquery.Selection) {
			m.URL, _ = a.Attr("href")
			m.Title = strings.TrimSpace(a.Text())
		})
		m.ThumbnailURL = imgAttr(el.Find("img").First(), s.cfg.BaseURL)
		if m.URL != "" {
			mangas = append(mangas, m)
		}
	})
	hasNext := doc.Find(".blog-pager > span.pagecurrent + span, .pagination .next").Length() > 0
	return source.MangasPage{Mangas: mangas, HasNextPage: hasNext}
}

func (s *Source) GetPopularManga(page int) (source.MangasPage, error) {
	doc, err := s.get(context.Background(), fmt.Sprintf("%s/ranking/week/%d", s.base(), page))
	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(), fmt.Sprintf("%s/all-manga/%d/?sort=last_update&status=0", s.base(), page))
	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) {
	doc, err := s.get(context.Background(), fmt.Sprintf("%s/all-manga/%d/?s=%s", s.base(), page, query))
	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(".a2 header h1").Text())
	if result.Title == "" {
		result.Title = manga.Title
	}
	result.ThumbnailURL = imgAttr(doc.Find(".a1 > figure img").First(), s.cfg.BaseURL)
	result.Description = strings.TrimSpace(doc.Find("div#syn-target").Text())
	result.Author = strings.TrimSpace(doc.Find("div.y6x11p i.fas.fa-user + span.dt").Text())
	result.Status = util.StatusFromString(doc.Find("div.y6x11p i.fas.fa-rss + span.dt").Text())
	var genres []string
	doc.Find("a[rel='tag'].label").Each(func(_ int, el *goquery.Selection) {
		if t := strings.TrimSpace(el.Text()); t != "" {
			genres = append(genres, t)
		}
	})
	result.Genre = strings.Join(genres, ", ")
	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("ul > li.chapter").Each(func(_ int, el *goquery.Selection) {
		ch := source.SChapter{}
		el.Find("a").First().Each(func(_ int, a *goquery.Selection) {
			ch.URL, _ = a.Attr("href")
			ch.Name = strings.TrimSpace(a.Text())
		})
		if dt, ok := el.Find("time[datetime]").First().Attr("datetime"); ok {
			ch.DateUpload = util.ParseAbsoluteDate(dt, "2006-01-02")
			if ch.DateUpload == 0 {
				ch.DateUpload = util.ParseRelativeDate(dt)
			}
		}
		if ch.URL != "" {
			chapters = append(chapters, ch)
		}
	})
	return chapters, nil
}

var chapterIDRe = regexp.MustCompile(`const\s+CHAPTER_ID\s*=\s*(\d+)`)

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
	}
	// extract CHAPTER_ID from inline script
	chapterID := ""
	doc.Find("script").Each(func(_ int, el *goquery.Selection) {
		if m := chapterIDRe.FindStringSubmatch(el.Text()); len(m) == 2 {
			chapterID = m[1]
		}
	})
	if chapterID == "" {
		return nil, fmt.Errorf("liliana: could not find CHAPTER_ID")
	}
	ajaxDoc, err := s.get(context.Background(), fmt.Sprintf("%s/ajax/image/list/chap/%s", s.base(), chapterID))
	if err != nil {
		return nil, err
	}
	var pages []source.Page
	// two formats: div.separator[data-index] or plain div.separator > a
	ajaxDoc.Find("div.separator[data-index]").Each(func(_ int, el *goquery.Selection) {
		idx := len(pages)
		if href, ok := el.Find("a").First().Attr("href"); ok && href != "" {
			pages = append(pages, source.Page{Index: idx, ImageURL: util.AbsURL(s.cfg.BaseURL, href)})
		}
	})
	if len(pages) == 0 {
		ajaxDoc.Find("div.separator").Each(func(_ int, el *goquery.Selection) {
			idx := len(pages)
			if href, ok := el.Find("a").First().Attr("href"); ok && href != "" {
				pages = append(pages, source.Page{Index: idx, ImageURL: util.AbsURL(s.cfg.BaseURL, href)})
			}
		})
	}
	return pages, nil
}

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

func imgAttr(img *goquery.Selection, baseURL string) string {
	for _, attr := range []string{"data-lazy-src", "data-src", "data-cfsrc", "src"} {
		if v, ok := img.Attr(attr); ok && v != "" && !strings.HasPrefix(v, "data:") {
			return util.AbsURL(baseURL, v)
		}
	}
	return ""
}