goyomi/sources/all/commitstrip/commitstrip.go

// Package commitstrip implements the Commit Strip webcomic source.
// Two language instances (en, fr). Popular is a static list of per-year entries
// (2012 → current year). Chapters scraped from paginated year archives; one page per strip.
package commitstrip

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

	"github.com/PuerkitoBio/goquery"

	"goyomi/internal/httpclient"
	"goyomi/internal/registry"
	"goyomi/internal/source"
)

const siteURL = "https://www.commitstrip.com"

var (
	dateRe = regexp.MustCompile(`\d{4}/\d{2}/\d{2}`)
	pageRe = regexp.MustCompile(`\d+`)
)

type Source struct {
	name     string
	lang     string
	siteLang string
	client   *httpclient.Client
	id       int64
}

func newSource(lang, siteLang string) *Source {
	return &Source{
		name:     "Commit Strip",
		lang:     lang,
		siteLang: siteLang,
		client:   httpclient.NewClient(httpclient.WithRateLimit(2, 1)),
		id:       source.GenerateSourceID("Commit Strip", lang),
	}
}

func (s *Source) ID() int64            { return s.id }
func (s *Source) Name() string         { return s.name }
func (s *Source) Lang() string         { return s.lang }
func (s *Source) SupportsLatest() bool { return false }

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
	}
	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("commitstrip: HTTP %d", resp.StatusCode)
	}
	return goquery.NewDocumentFromReader(resp.Body)
}

func currentYear() int { return time.Now().Year() }

func (s *Source) thumbnail() string {
	if s.lang == "fr" {
		return "https://i.imgur.com/I7ps9zS.jpg"
	}
	return "https://i.imgur.com/HODJlt9.jpg"
}

func (s *Source) author() string {
	if s.lang == "fr" {
		return "Thomas Gx"
	}
	return "Mark Nightingale"
}

func (s *Source) summary(year int) string {
	note := fmt.Sprintf("\n\nNote: This entry includes all the chapters published in %d", year)
	if s.lang == "fr" {
		return "Le blog qui raconte la vie des codeurs" + note
	}
	return "The blog relating the daily life of web agency developers." + note
}

func (s *Source) makeYearManga(year int) source.SManga {
	status := source.StatusOngoing
	if year != currentYear() {
		status = source.StatusCompleted
	}
	return source.SManga{
		URL:          fmt.Sprintf("/%s/%d", s.siteLang, year),
		Title:        fmt.Sprintf("Commit Strip (%d)", year),
		ThumbnailURL: s.thumbnail(),
		Author:       s.author(),
		Artist:       "Etienne Issartial",
		Status:       status,
		Description:  s.summary(year),
	}
}

func (s *Source) GetPopularManga(page int) (source.MangasPage, error) {
	if page > 1 {
		return source.MangasPage{}, nil
	}
	cur := currentYear()
	mangas := make([]source.SManga, 0, cur-2011)
	for y := cur; y >= 2012; y-- {
		mangas = append(mangas, s.makeYearManga(y))
	}
	return source.MangasPage{Mangas: mangas, HasNextPage: false}, nil
}

func (s *Source) GetLatestUpdates(page int) (source.MangasPage, error) {
	return source.MangasPage{}, fmt.Errorf("commitstrip: latest not supported")
}

func (s *Source) GetSearchManga(page int, query string, _ []source.Filter) (source.MangasPage, error) {
	all, _ := s.GetPopularManga(1)
	if query == "" {
		return all, nil
	}
	q := strings.ToLower(query)
	var matched []source.SManga
	for _, m := range all.Mangas {
		if strings.Contains(strings.ToLower(m.Title), q) {
			matched = append(matched, m)
		}
	}
	return source.MangasPage{Mangas: matched, HasNextPage: false}, nil
}

func (s *Source) GetMangaDetails(manga source.SManga) (source.SManga, error) {
	return manga, nil
}

func (s *Source) GetChapterList(manga source.SManga) ([]source.SChapter, error) {
	yearURL := siteURL + manga.URL
	doc, err := s.get(context.Background(), yearURL)
	if err != nil {
		return nil, err
	}

	// Find total pages from ".wp-pagenavi .pages" text (e.g. "Page 1 of 12").
	totalPages := 1
	if pagesText := doc.Find(".wp-pagenavi .pages").First().Text(); pagesText != "" {
		matches := pageRe.FindAllString(pagesText, -1)
		if len(matches) >= 2 {
			fmt.Sscanf(matches[len(matches)-1], "%d", &totalPages)
		}
	}

	var chapters []source.SChapter
	collect := func(d *goquery.Document) {
		d.Find(".excerpt a").Each(func(_ int, a *goquery.Selection) {
			href := a.AttrOr("href", "")
			if href == "" {
				return
			}
			chURL := strings.TrimPrefix(href, siteURL)
			name := strings.TrimSpace(a.Find("span").Text())
			if name == "" {
				name = strings.TrimSpace(a.Text())
			}
			var date int64
			if m := dateRe.FindString(chURL); m != "" {
				if t, err := time.Parse("2006/01/02", m); err == nil {
					date = t.UnixMilli()
				}
			}
			chapters = append(chapters, source.SChapter{URL: chURL, Name: name, DateUpload: date})
		})
	}

	collect(doc)
	for pg := 2; pg <= totalPages; pg++ {
		pageDoc, err := s.get(context.Background(), fmt.Sprintf("%s/page/%d", yearURL, pg))
		if err != nil {
			break
		}
		collect(pageDoc)
	}

	// Deduplicate and assign chapter numbers.
	seen := make(map[string]bool)
	var unique []source.SChapter
	for _, ch := range chapters {
		if !seen[ch.URL] {
			seen[ch.URL] = true
			unique = append(unique, ch)
		}
	}
	total := len(unique)
	for i := range unique {
		unique[i].ChapterNumber = float32(total - i)
	}
	return unique, nil
}

func (s *Source) GetPageList(chapter source.SChapter) ([]source.Page, error) {
	doc, err := s.get(context.Background(), siteURL+chapter.URL)
	if err != nil {
		return nil, err
	}
	src := doc.Find(".entry-content p img").First().AttrOr("src", "")
	if src == "" {
		return nil, fmt.Errorf("commitstrip: image not found")
	}
	return []source.Page{{Index: 0, ImageURL: src}}, nil
}

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

func init() {
	registry.Register(newSource("en", "en"))
	registry.Register(newSource("fr", "fr"))
}