eh-downloader/internal/eh/e2e_test.go

package eh

// 端到端协议链路测试：用 httptest 模拟 EH 服务器，
// 验证「画廊页分页 → /s/ → hath 直链 → 下载」整条链路，
// 以及断点续传时「文件已存在就跳过」的行为。
//
// 这些 HTML 样本按真实 EH 页面结构构造（字段名、缩略图 title 属性、
// Showing X - Y of N images 等都照抄线上结构）。

import (
	"context"
	"fmt"
	"net/http"
	"net/http/httptest"
	"os"
	"path/filepath"
	"strings"
	"testing"
	"time"
)

// 模拟画廊：45 张图 = 3 屏（20 + 20 + 5），用来验证 ?p=N 分页
func mockGalleryHTML(t *testing.T, gid, token string, totalImages int) map[int]string {
	pages := map[int]string{} // p 序号 -> html

	for p := 0; p < 3; p++ {
		from := p*20 + 1
		to := from + 19
		if to > totalImages {
			to = totalImages
		}
		var sb strings.Builder

		if p == 0 {
			sb.WriteString(`<!DOCTYPE html><html><head>`)
			sb.WriteString(`<title>[Noxe] Test Gallery - E-Hentai Galleries</title>`)
			sb.WriteString(`<meta name="description" content="Free Hentai Doujinshi Gallery: [Noxe] Test Gallery: 45 images (ratelimit)" />`)
			sb.WriteString(`</head><body><div class="gtb"><table><tr><td>`)
			sb.WriteString(`<div id="taglist"><table>`)
			sb.WriteString(`<tr><td class="tc">artist:</td><td><div class="gtl"><a href="/tag/artist:noxe">Noxe</a></div></td></tr>`)
			sb.WriteString(`<tr><td class="tc">group:</td><td><div class="gtl"><a href="/tag/group:criticalhai">CriticalHAI</a></div></td></tr>`)
			sb.WriteString(`<tr><td class="tc">other:</td><td><div class="gtl"><a href="/tag:qingxiao">qingxiao</a><a href="/tag:rover">rover</a></div></td></tr>`)
			sb.WriteString(`</table></div><a href="/uploader/velius11">up</a>`)
		}

		for page := from; page <= to; page++ {
			key := fmt.Sprintf("key%02d", page)
			sb.WriteString(fmt.Sprintf(
				`<a href="/s/%s/%s-%d"><div title="Page %d: %03d.png" class="lazy"></div></a>`,
				key, gid, page, page, page))
		}

		if p == 0 {
			sb.WriteString(`<div class="gtb">`)
		}
		sb.WriteString(fmt.Sprintf(`Showing %d - %d of %d images`, from, to, totalImages))
		if p == 0 {
			sb.WriteString(`</div></body></html>`)
		}
		pages[p] = sb.String()
	}
	return pages
}

func TestFetchGalleryPagePagination(t *testing.T) {
	const (
		gid  = "4239250"
		tok  = "58703925db"
		total = 45
	)

	galleryPages := mockGalleryHTML(t, gid, tok, total)

	mux := http.NewServeMux()
	mux.HandleFunc("/g/"+gid+"/"+tok+"/", func(w http.ResponseWriter, r *http.Request) {
		p := 0
		if q := r.URL.Query().Get("p"); q != "" {
			fmt.Sscanf(q, "%d", &p)
		}
		html, ok := galleryPages[p]
		if !ok {
			// 超过范围的屏：EH 会返回空列表页
			w.Write([]byte(`<html>Showing 46 - 45 of 45 images</html>`))
			return
		}
		w.Write([]byte(html))
	})
	mux.HandleFunc("/s/", func(w http.ResponseWriter, r *http.Request) {
		// /s/{key}/{gid}-{page}
		page := strings.TrimPrefix(r.URL.Path, "/s/")
		parts := strings.Split(page, "/")
		if len(parts) != 2 {
			w.WriteHeader(404)
			return
		}
		idx := strings.LastIndex(parts[1], "-")
		pno := parts[1][idx+1:]
		// 每页一个不同的 hath URL（模拟随机子域+端口+keystamp）
		w.Write([]byte(fmt.Sprintf(
			`<html><img src="https://sub%d.gyktyewfonyf.hath.network:%d/h/abc/keystamp=1700000%d;fileindex=%d;xres=800/pic.png"/></html>`,
			len(pno)%10, 39000+len(pno), len(pno), len(pno))))
	})

	srv := httptest.NewServer(mux)
	defer srv.Close()

	// 把 EHBase 换到测试服务器
	oldBase := EHBase
	EHBase = srv.URL
	defer func() { EHBase = oldBase }()

	c := NewClient(5*time.Second, "", "off", "auto")
	g, err := c.FetchGalleryPage(context.Background(), gid, tok, nil)
	if err != nil {
		t.Fatalf("FetchGalleryPage 失败: %v", err)
	}

	// 关键断言：45 张必须全部拿到（分页生效）
	if len(g.Pages) != total {
		t.Errorf("应拿到 %d 页，实际 %d 页 —— 分页可能没生效", total, len(g.Pages))
	}
	if g.TotalPages != total {
		t.Errorf("TotalPages 不对: %d", g.TotalPages)
	}

	// 页码连续且有序
	for i, p := range g.Pages {
		if p.Page != i+1 {
			t.Errorf("第 %d 项页码应为 %d，实际 %d", i, i+1, p.Page)
		}
	}

	// key 必须与页码配对（每页的 key 不同）；SPath 会被补全成绝对地址
	if !strings.HasSuffix(g.Pages[0].SPath, "/s/key01/4239250-1") {
		t.Errorf("第 1 页路径不对: %s", g.Pages[0].SPath)
	}
	if !strings.HasSuffix(g.Pages[1].SPath, "/s/key02/4239250-2") {
		t.Errorf("第 2 页路径不对（key 与 page 必须配对）: %s", g.Pages[1].SPath)
	}

	// 文件名
	if g.Pages[0].Filename != "001.png" {
		t.Errorf("文件名不对: %q", g.Pages[0].Filename)
	}

	// 元信息只在第一页解析
	if g.Title != "[Noxe] Test Gallery" {
		t.Errorf("标题不对: %q", g.Title)
	}
	if g.Meta.Artist != "Noxe" {
		t.Errorf("作者不对: %q", g.Meta.Artist)
	}
	if g.Meta.Category != "doujinshi" {
		t.Errorf("分类不对: %q", g.Meta.Category)
	}
	if len(g.Meta.Tags) != 2 {
		t.Errorf("标签不对: %v", g.Meta.Tags)
	}
}

// 解析 hath 直链 + xres 改写
func TestResolveImageURLEndToEnd(t *testing.T) {
	mux := http.NewServeMux()
	mux.HandleFunc("/s/", func(w http.ResponseWriter, r *http.Request) {
		w.Write([]byte(`<html><body>
			<img src="https://oclzqqk.gyktyewfonyf.hath.network:39656/h/4e9af9f-44168-707-754-wbp/keystamp=1791448200-8b38ea8f6d;fileindex=270107377;xres=800/NH.webp">
		</body></html>`))
	})
	srv := httptest.NewServer(mux)
	defer srv.Close()

	oldBase := EHBase
	EHBase = srv.URL
	defer func() { EHBase = oldBase }()

	c := NewClient(5*time.Second, "", "off", "auto")
	got, err := c.ResolveImageURL(context.Background(), srv.URL+"/s/abc/123-1")
	if err != nil {
		t.Fatalf("解析失败: %v", err)
	}
	if !strings.HasPrefix(got, "https://oclzqqk.gyktyewfonyf.hath.network:39656/") {
		t.Errorf("直链不对: %s", got)
	}
	if !strings.Contains(got, "xres=800") {
		t.Errorf("应保留站点给的 xres: %s", got)
	}
	if strings.Contains(got, `"`) || strings.Contains(got, ">") {
		t.Errorf("URL 里混进了 HTML 符号: %s", got)
	}

	// 切到 org 应该改写
	c.SetXres("org")
	got2, _ := c.ResolveImageURL(context.Background(), srv.URL+"/s/abc/123-1")
	if !strings.Contains(got2, "xres=org") {
		t.Errorf("xres=org 改写失败: %s", got2)
	}
}

// 断点续传：文件已存在且非空就该跳过
func TestDownloadImageResume(t *testing.T) {
	payload := []byte("fake webp bytes for testing resume")
	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
		w.Write(payload)
	}))
	defer srv.Close()

	dir := t.TempDir()
	c := NewClient(5*time.Second, "", "off", "auto")

	dest := filepath.Join(dir, "001.webp")
	n, err := c.DownloadImage(context.Background(), srv.URL+"/pic.webp", dest, nil)
	if err != nil {
		t.Fatalf("下载失败: %v", err)
	}
	if n != int64(len(payload)) {
		t.Errorf("下载字节数不对: %d", n)
	}

	got, err := os.ReadFile(dest)
	if err != nil {
		t.Fatal(err)
	}
	if string(got) != string(payload) {
		t.Errorf("内容不对")
	}

	// .part 文件不该残留
	if _, err := os.Stat(dest + ".part"); err == nil {
		t.Error("临时 .part 文件没清理")
	}

	// 进度回调
	var received int64
	c2 := NewClient(5*time.Second, "", "off", "auto")
	dest2 := filepath.Join(dir, "002.webp")
	c2.DownloadImage(context.Background(), srv.URL+"/pic.webp", dest2, func(recv, total int64) {
		received = recv
	})
	if received != int64(len(payload)) {
		t.Errorf("进度回调不对: %d", received)
	}
}

// 取消要能真的打断
func TestDownloadImageCancel(t *testing.T) {
	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
		// 慢慢吐
		w.Write([]byte("start"))
		if f, ok := w.(http.Flusher); ok {
			f.Flush()
		}
		time.Sleep(3 * time.Second)
	}))
	defer srv.Close()

	ctx, cancel := context.WithCancel(context.Background())
	go func() {
		time.Sleep(200 * time.Millisecond)
		cancel()
	}()

	c := NewClient(5*time.Second, "", "off", "auto")
	dest := filepath.Join(t.TempDir(), "x.webp")
	_, err := c.DownloadImage(ctx, srv.URL+"/slow.webp", dest, nil)
	if err == nil {
		t.Error("取消后应该返回错误")
	}
	// 临时文件要被清理
	if _, err := os.Stat(dest + ".part"); err == nil {
		t.Error("取消后 .part 没清理")
	}
}

// 受限画廊要给出明确错误
func TestFetchGalleryRestricted(t *testing.T) {
	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
		w.Write([]byte(`<html><body>Login Required - restricted gallery</body></html>`))
	}))
	defer srv.Close()

	oldBase := EHBase
	EHBase = srv.URL
	defer func() { EHBase = oldBase }()

	c := NewClient(5*time.Second, "", "off", "auto")
	_, err := c.FetchGalleryPage(context.Background(), "12345678", "abc", nil)
	if err == nil {
		t.Fatal("受限画廊应该报错")
	}
	if !strings.Contains(err.Error(), "受限") {
		t.Errorf("错误信息不明确: %v", err)
	}
}