goyomi/sources/base/raijinscans/raijinscans.go

// Package raijinscans implements the RaijinScans manga base.
// French scan site; CF-protected; chapter page URLs are Base64-encoded in data-src attributes.
package raijinscans

import (
	"context"
	"encoding/base64"
	"fmt"
	"net/http"
	"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 {
	c := flare.NewClient(flare.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+"/")
	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("raijinscans: HTTP %d", resp.StatusCode)
	}
	return goquery.NewDocumentFromReader(resp.Body)
}

func (s *Source) parseMangaFromUnit(el *goquery.Selection) source.SManga {
	a := el.Find("a").First()
	m := source.SManga{
		URL:   a.AttrOr("href", ""),
		Title: strings.TrimSpace(el.Find(".title, h2, h3, span.name").First().Text()),
	}
	if m.Title == "" {
		m.Title = strings.TrimSpace(a.AttrOr("title", ""))
	}
	if img := el.Find("img").First(); img.Length() > 0 {
		m.ThumbnailURL = util.AbsURL(s.cfg.BaseURL, img.AttrOr("src", img.AttrOr("data-src", "")))
	}
	return m
}

func (s *Source) GetPopularManga(page int) (source.MangasPage, error) {
	var u string
	if page == 1 {
		u = s.base()
	} else {
		u = fmt.Sprintf("%s/page/%d/", s.base(), page)
	}
	doc, err := s.get(context.Background(), u)
	if err != nil {
		return source.MangasPage{}, err
	}
	var mangas []source.SManga
	doc.Find("section#most-viewed div.swiper-slide.unit").Each(func(_ int, el *goquery.Selection) {
		m := s.parseMangaFromUnit(el)
		if m.URL != "" && m.Title != "" {
			mangas = append(mangas, m)
		}
	})
	hasNext := doc.Find(".pagination .next, a[rel=next]").Length() > 0
	return source.MangasPage{Mangas: mangas, HasNextPage: hasNext}, nil
}

func (s *Source) GetLatestUpdates(page int) (source.MangasPage, error) {
	var u string
	if page == 1 {
		u = s.base()
	} else {
		u = fmt.Sprintf("%s/page/%d/", s.base(), page)
	}
	doc, err := s.get(context.Background(), u)
	if err != nil {
		return source.MangasPage{}, err
	}
	var mangas []source.SManga
	doc.Find("div.original.card-lg div.unit, div.latest div.unit").Each(func(_ int, el *goquery.Selection) {
		m := s.parseMangaFromUnit(el)
		if m.URL != "" && m.Title != "" {
			mangas = append(mangas, m)
		}
	})
	hasNext := doc.Find(".pagination .next, a[rel=next]").Length() > 0
	return source.MangasPage{Mangas: mangas, HasNextPage: hasNext}, nil
}

func (s *Source) GetSearchManga(page int, query string, filters []source.Filter) (source.MangasPage, error) {
	u := fmt.Sprintf("%s/?s=%s&post_type=wp-manga&paged=%d", s.base(), query, page)
	doc, err := s.get(context.Background(), u)
	if err != nil {
		return source.MangasPage{}, err
	}
	var mangas []source.SManga
	doc.Find("div.original.card-lg div.unit, div.search-result div.unit").Each(func(_ int, el *goquery.Selection) {
		m := s.parseMangaFromUnit(el)
		if m.URL != "" && m.Title != "" {
			mangas = append(mangas, m)
		}
	})
	hasNext := doc.Find(".pagination .next, a[rel=next]").Length() > 0
	return source.MangasPage{Mangas: mangas, HasNextPage: hasNext}, 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("h1.entry-title, h1.manga-title, h1").First().Text())
	if result.Title == "" {
		result.Title = manga.Title
	}
	if img := doc.Find("img.cover, div.cover img, div.manga-cover img").First(); img.Length() > 0 {
		result.ThumbnailURL = util.AbsURL(s.cfg.BaseURL, img.AttrOr("src", img.AttrOr("data-src", "")))
	}
	result.Description = strings.TrimSpace(doc.Find(".summary__content, .manga-summary, .description").First().Text())

	var genres []string
	doc.Find("div.genre-list div.genre-link, a.genre-tag").Each(func(_ int, a *goquery.Selection) {
		if t := strings.TrimSpace(a.Text()); t != "" {
			genres = append(genres, t)
		}
	})
	result.Genre = strings.Join(genres, ", ")

	// Status from stat-item containing "État du titre" or "Status"
	doc.Find("div.stat-item").Each(func(_ int, el *goquery.Selection) {
		label := strings.ToLower(el.Find("span").First().Text())
		if strings.Contains(label, "état") || strings.Contains(label, "status") {
			val := strings.TrimSpace(el.Find("span.manga, span.value, span:last-child").Last().Text())
			result.Status = util.StatusFromString(val)
		}
	})
	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.scroll-sm li.item, ul.chapter-list li, li.wp-manga-chapter").Each(func(_ int, el *goquery.Selection) {
		a := el.Find("a").First()
		href := a.AttrOr("href", "")
		if href == "" {
			return
		}
		name := strings.TrimSpace(a.Text())
		if name == "" {
			name = "Chapter"
		}
		var ts int64
		if span := el.Find("span.date, span.chapter-release-date, time").First(); span.Length() > 0 {
			ts = parseFrenchDate(strings.TrimSpace(span.Text()))
		}
		chapters = append(chapters, source.SChapter{
			URL:        href,
			Name:       name,
			DateUpload: ts,
		})
	})
	return chapters, nil
}

func parseFrenchDate(s string) int64 {
	lower := strings.ToLower(strings.TrimSpace(s))
	// "aujourd'hui" = today, "hier" = yesterday
	if strings.HasPrefix(lower, "aujourd") {
		return util.ParseRelativeDate("0 days ago")
	}
	if lower == "hier" {
		return util.ParseRelativeDate("1 day ago")
	}
	// "{n} jour(s)" / "{n} heure(s)" / "{n} semaine(s)"
	lower = strings.ReplaceAll(lower, "jour(s)", "days")
	lower = strings.ReplaceAll(lower, "jours", "days")
	lower = strings.ReplaceAll(lower, "jour", "day")
	lower = strings.ReplaceAll(lower, "heure(s)", "hours")
	lower = strings.ReplaceAll(lower, "heures", "hours")
	lower = strings.ReplaceAll(lower, "heure", "hour")
	lower = strings.ReplaceAll(lower, "semaine(s)", "weeks")
	lower = strings.ReplaceAll(lower, "semaines", "weeks")
	lower = strings.ReplaceAll(lower, "semaine", "week")
	lower = strings.ReplaceAll(lower, "mois", "months")
	lower = strings.ReplaceAll(lower, "an(s)", "years")
	lower = strings.ReplaceAll(lower, "ans", "years")
	lower = lower + " ago"
	if ts := util.ParseRelativeDate(lower); ts != 0 {
		return ts
	}
	return util.ParseAbsoluteDate(s, "2 January 2006")
}

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("div.protected-image-data").Each(func(i int, el *goquery.Selection) {
		encoded := strings.TrimSpace(el.AttrOr("data-src", ""))
		if encoded == "" {
			return
		}
		decoded, err := base64.StdEncoding.DecodeString(encoded)
		if err != nil {
			// Try RawStdEncoding (no padding)
			decoded, err = base64.RawStdEncoding.DecodeString(encoded)
			if err != nil {
				return
			}
		}
		imgURL := strings.TrimSpace(string(decoded))
		if imgURL != "" {
			pages = append(pages, source.Page{Index: i, ImageURL: imgURL})
		}
	})

	// Fallback: regular img tags
	if len(pages) == 0 {
		doc.Find("div.reading-content img, div.reader-area img").Each(func(i int, img *goquery.Selection) {
			u := img.AttrOr("src", img.AttrOr("data-src", ""))
			if u != "" {
				pages = append(pages, source.Page{Index: i, ImageURL: util.AbsURL(s.cfg.BaseURL, 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 }