Files
goyomi/sources/base/scanreader/scanreader.go
T
2026-05-11 06:48:23 +00:00

283 lines
8.5 KiB
Go
Executable File

// Package scanreader implements the ScanReader French manga base.
// GET {base}/bibliotheque/page/{n-1}/?sort=views; WordPress AJAX chapters.
package scanreader
import (
"context"
"encoding/json"
"fmt"
"io"
"net/http"
"net/url"
"strings"
"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+"/")
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("scanreader: HTTP %d", resp.StatusCode)
}
return goquery.NewDocumentFromReader(resp.Body)
}
func mangaFromCard(el *goquery.Selection, baseURL string) source.SManga {
m := source.SManga{}
el.Find("a").First().Each(func(_ int, a *goquery.Selection) {
m.URL, _ = a.Attr("href")
m.Title = strings.TrimSpace(a.AttrOr("title", a.Text()))
})
m.ThumbnailURL = imgAttr(el.Find("img").First(), baseURL)
return m
}
func (s *Source) GetPopularManga(page int) (source.MangasPage, error) {
if page == 1 {
doc, err := s.get(context.Background(), s.base())
if err != nil {
return source.MangasPage{}, err
}
var mangas []source.SManga
doc.Find("div.popular-section div.manga-card").Each(func(_ int, el *goquery.Selection) {
m := mangaFromCard(el, s.cfg.BaseURL)
if m.URL != "" {
mangas = append(mangas, m)
}
})
return source.MangasPage{Mangas: mangas, HasNextPage: true}, nil
}
doc, err := s.get(context.Background(), fmt.Sprintf("%s/bibliotheque/page/%d/?sort=views", s.base(), page-1))
if err != nil {
return source.MangasPage{}, err
}
var mangas []source.SManga
doc.Find("div.manga-card").Each(func(_ int, el *goquery.Selection) {
m := mangaFromCard(el, s.cfg.BaseURL)
if m.URL != "" {
mangas = append(mangas, m)
}
})
hasNext := doc.Find("a.pagination-next").Length() > 0
return source.MangasPage{Mangas: mangas, HasNextPage: hasNext}, nil
}
func (s *Source) GetLatestUpdates(page int) (source.MangasPage, error) {
doc, err := s.get(context.Background(), fmt.Sprintf("%s/dernieres-sorties/page/%d/", s.base(), page))
if err != nil {
return source.MangasPage{}, err
}
var mangas []source.SManga
doc.Find("div.manga-cover, div.manga-card").Each(func(_ int, el *goquery.Selection) {
m := mangaFromCard(el, s.cfg.BaseURL)
if m.URL != "" {
mangas = append(mangas, m)
}
})
hasNext := doc.Find("a.pagination-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) {
doc, err := s.get(context.Background(), fmt.Sprintf("%s/?s=%s&post_type=manga", s.base(), url.QueryEscape(query)))
if err != nil {
return source.MangasPage{}, err
}
var mangas []source.SManga
doc.Find("div.manga-card").Each(func(_ int, el *goquery.Selection) {
m := mangaFromCard(el, s.cfg.BaseURL)
if m.URL != "" {
mangas = append(mangas, m)
}
})
return source.MangasPage{Mangas: mangas, HasNextPage: false}, 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.manga-title").Text())
if result.Title == "" {
result.Title = manga.Title
}
if v, ok := doc.Find("meta[property='og:image']").First().Attr("content"); ok {
result.ThumbnailURL = v
}
result.Description = strings.TrimSpace(doc.Find("div.manga-content div[style*='background: #333'] p").Text())
doc.Find("div.manga-info-grid > div").Each(func(_ int, row *goquery.Selection) {
label := strings.ToLower(strings.TrimSpace(row.Find("div:first-child").Text()))
val := strings.TrimSpace(row.Find("div:last-child").Text())
switch {
case strings.Contains(label, "auteur"):
result.Author = val
case strings.Contains(label, "statut"):
result.Status = util.StatusFromString(val)
case strings.Contains(label, "genre"):
var genres []string
row.Find("span, a").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
}
container := doc.Find("#secure-chapters-container")
if container.Length() == 0 {
// try direct chapter list
return s.parseChapterList(doc), nil
}
mangaID := container.AttrOr("data-manga-id", "")
nonce := container.AttrOr("data-nonce", "")
if mangaID == "" || nonce == "" {
return s.parseChapterList(doc), nil
}
// POST admin-ajax.php to get chapters HTML
formData := url.Values{
"action": {"get_chapters"},
"manga_id": {mangaID},
"nonce": {nonce},
}
req, err := http.NewRequestWithContext(context.Background(), http.MethodPost,
s.base()+"/wp-admin/admin-ajax.php",
strings.NewReader(formData.Encode()))
if err != nil {
return nil, err
}
req.Header.Set("Content-Type", "application/x-www-form-urlencoded")
req.Header.Set("Referer", s.cfg.BaseURL+"/")
resp, err := s.client.Do(req)
if err != nil {
return nil, err
}
defer resp.Body.Close()
body, _ := io.ReadAll(resp.Body)
// try to unwrap AjaxResponse {"data": "<html>"}, else use raw
var ajax struct{ Data string `json:"data"` }
htmlStr := string(body)
if json.Unmarshal(body, &ajax) == nil && ajax.Data != "" {
htmlStr = ajax.Data
}
ajaxDoc, err := goquery.NewDocumentFromReader(strings.NewReader(htmlStr))
if err != nil {
return nil, err
}
return s.parseChapterList(ajaxDoc), nil
}
func (s *Source) parseChapterList(doc *goquery.Document) []source.SChapter {
var chapters []source.SChapter
doc.Find("h4").Each(func(_ int, h4 *goquery.Selection) {
var chURL string
h4.Parents().Each(func(_ int, ancestor *goquery.Selection) {
if chURL != "" {
return
}
if href, ok := ancestor.Find("a[href*='/chapitre/']").First().Attr("href"); ok {
chURL = href
}
})
if chURL == "" {
if href, ok := h4.Find("a").First().Attr("href"); ok {
chURL = href
}
}
if chURL == "" {
return
}
ch := source.SChapter{
URL: chURL,
Name: strings.TrimSpace(h4.Text()),
}
chapters = append(chapters, ch)
})
// fallback: li-based
if len(chapters) == 0 {
doc.Find("li.chapter-item, .chapter-list li").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 ch.URL != "" {
chapters = append(chapters, ch)
}
})
}
return chapters
}
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.reading-content img, .chapter-container img, #chapter-images img").Each(func(i int, img *goquery.Selection) {
if u := imgAttr(img, 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 }
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 ""
}