feat(cbz): adapt reader to real archive structure — PageIndex skips macOS junk (__MACOSX/ and ._ AppleDouble were polluting pages with 163-byte blacks), pages endpoint derives chapters[{title,start}] from folder layout, imageless zip now lands state=error instead of an empty reader, CbzReader gains 目录 sidebar + prev/next chapter (TextReader pattern, rd-* classes); pagesidx cache key bumped for one-time invalidation; e2e on 311MB 調教開關：第二季.zip: 15048→7524 real pages, 52 chapters, all JPEGs

2026-09-08 00:22:13 +08:00
parent 7d650107d2
commit b8da9fd92e
9 changed files with 216 additions and 3 deletions
@@ -4,6 +4,7 @@ import (
 	"fmt"
 	"net/http"
 	"os"
+	"path"
 	"path/filepath"
 	"strconv"
 	"strings"
@@ -118,7 +119,7 @@ func (h *H) openBook(c *gin.Context, b store.Book, root string) (*os.File, int64

 func (h *H) pageIndex(c *gin.Context, b store.Book, root string) ([]string, error) {
 	hash := bookfile.Hash(b.FileSize, b.ModTS)
-	key := fmt.Sprintf("pagesidx:%d:%s", b.ID, hash)
+	key := fmt.Sprintf("pagesidx2:%d:%s", b.ID, hash) // 前缀换代 = 索引逻辑变更时一次性作废旧缓存
 	if v, ok := h.rdb.Get(c, key); ok && v != "" {
 		return strings.Split(v, "\n"), nil
 	}
@@ -158,7 +159,48 @@ func (h *H) PagesCount(c *gin.Context) {
 		err(c, http.StatusUnprocessableEntity, "broken", e.Error())
 		return
 	}
-	c.JSON(http.StatusOK, gin.H{"count": len(idx)})
+	c.JSON(http.StatusOK, gin.H{"count": len(idx), "chapters": chaptersOf(idx)})
+}
+
+type cbzChapter struct {
+	Title string `json:"title"`
+	Start int    `json:"start"`
+}
+
+// chaptersOf 页索引已自然序排列,按父目录分组:每个新目录开一章,标题取目录名;扁平包或只有一组返回 nil
+func chaptersOf(idx []string) []cbzChapter {
+	type grp struct {
+		dir   string
+		start int
+	}
+	var grps []grp
+	last := "\x00"
+	for i, n := range idx {
+		d := path.Dir(n)
+		if d == last {
+			continue
+		}
+		last = d
+		if d == "." { // 根目录散页不开章
+			continue
+		}
+		grps = append(grps, grp{d, i})
+	}
+	if len(grps) < 2 {
+		return nil
+	}
+	titles := make(map[string]int)
+	out := make([]cbzChapter, len(grps))
+	for i, g := range grps {
+		out[i] = cbzChapter{Title: path.Base(g.dir), Start: g.start}
+		titles[out[i].Title]++
+	}
+	for i, g := range grps { // 同名目录(不同父级)撞车 → 用全路径消歧
+		if titles[out[i].Title] > 1 {
+			out[i].Title = g.dir
+		}
+	}
+	return out
 }

 func (h *H) Page(c *gin.Context) {
@@ -1,6 +1,8 @@
 package handlers_test

 import (
+	"archive/zip"
+	"bytes"
 	"context"
 	"encoding/json"
 	"net/http"
@@ -103,6 +105,76 @@ func TestPages(t *testing.T) {
 	}
 }

+func zipTo(t *testing.T, path string, kv map[string]string) {
+	t.Helper()
+	os.MkdirAll(filepath.Dir(path), 0o755)
+	buf := &bytes.Buffer{}
+	zw := zip.NewWriter(buf)
+	for n, v := range kv {
+		w, err := zw.Create(n)
+		if err != nil {
+			t.Fatal(err)
+		}
+		w.Write([]byte(v))
+	}
+	zw.Close()
+	os.WriteFile(path, buf.Bytes(), 0o644)
+}
+
+func TestPagesChaptersAndNoImageZip(t *testing.T) {
+	st, sc, h, booksDir := setupAPI(t)
+	tok := adminToken(t, h)
+	lib, root := newLibrary(t, st, h, tok, booksDir, "ch")
+	zipTo(t, filepath.Join(root, "show.cbz"), map[string]string{
+		"第2季/第2話/0001.jpg":     "IMG2",
+		"第2季/第1話/0001.jpg":     "IMG1a",
+		"第2季/第1話/._0001.jpg":   "junk",
+		"__MACOSX/第2季/._0001.jpg": "junk",
+	})
+	zipTo(t, filepath.Join(root, "videos.zip"), map[string]string{"ep/01.mkv": "x"})
+	scanNow(t, sc, lib)
+
+	bookID := func(q string) string {
+		w := do(h, "GET", "/api/books?q="+q, tok, nil)
+		var bs []map[string]any
+		json.Unmarshal(w.Body.Bytes(), &bs)
+		if len(bs) != 1 {
+			t.Fatalf("q=%s books: %s", q, w.Body)
+		}
+		return itoa(bs[0]["id"])
+	}
+	cbzID := bookID("show")
+	w := do(h, "GET", "/api/books/"+cbzID+"/pages", tok, nil)
+	var pg struct {
+		Count    int `json:"count"`
+		Chapters []struct {
+			Title string `json:"title"`
+			Start int    `json:"start"`
+		} `json:"chapters"`
+	}
+	json.Unmarshal(w.Body.Bytes(), &pg)
+	if pg.Count != 2 || len(pg.Chapters) != 2 {
+		t.Fatalf("pages want 2/2ch got %s", w.Body)
+	}
+	if pg.Chapters[0].Title != "第1話" || pg.Chapters[0].Start != 0 || pg.Chapters[1].Start != 1 {
+		t.Fatalf("chapters: %s", w.Body)
+	}
+	ww := do(h, "GET", "/api/books/"+cbzID+"/pages/0", tok, nil)
+	if ww.Code != 200 || ww.Body.String() != "IMG1a" {
+		t.Fatalf("page0 must be real image (junk filtered): %d %s", ww.Code, ww.Body)
+	}
+	if ww := do(h, "GET", "/api/books/"+cbzID+"/pages/2", tok, nil); ww.Code != 404 {
+		t.Fatalf("junk-indexed page must be gone: %d", ww.Code)
+	}
+	// 无图 zip → state=error,而不是空白 reader
+	w = do(h, "GET", "/api/books/"+bookID("videos"), tok, nil)
+	var bk map[string]any
+	json.Unmarshal(w.Body.Bytes(), &bk)
+	if bk["state"] != "error" || !strings.Contains(bk["error"].(string), "no images") {
+		t.Fatalf("empty-image zip want error, got %s", w.Body)
+	}
+}
+
 func TestBrokenCBZ(t *testing.T) {
 	st, sc, h, booksDir := setupAPI(t)
 	atok := adminToken(t, h)
@@ -26,6 +26,14 @@ func isImage(name string) bool {
 	return false
 }

+// junkEntry: macOS 打包混入的资源叉垃圾(__MACOSX/ 目录与 ._* AppleDouble),不是页
+func junkEntry(name string) bool {
+	if strings.HasPrefix(strings.ToLower(name), "__macosx/") {
+		return true
+	}
+	return strings.HasPrefix(path.Base(name), "._")
+}
+
 func PageIndex(f io.ReaderAt, size int64) ([]string, error) {
 	zr, err := zip.NewReader(f, size)
 	if err != nil {
@@ -36,6 +44,9 @@ func PageIndex(f io.ReaderAt, size int64) ([]string, error) {
 		if unsafeEntry(zf.Name) {
 			return nil, fmt.Errorf("%w: %s", ErrUnsafeZip, zf.Name)
 		}
+		if junkEntry(zf.Name) {
+			continue
+		}
 		if isImage(zf.Name) {
 			names = append(names, zf.Name)
 		}
@@ -38,6 +38,23 @@ func TestPageIndexSortAndFilter(t *testing.T) {
 	}
 }

+func TestPageIndexSkipsAppleDouble(t *testing.T) {
+	r := zipOf(t, "第2季/第2話/0001.jpg", "第2季/第1話/._0001.jpg", "__MACOSX/第2季/._0001.jpg", "第2季/第1話/0001.jpg", "._top.jpg")
+	idx, err := PageIndex(r, int64(r.Len()))
+	if err != nil {
+		t.Fatal(err)
+	}
+	want := []string{"第2季/第1話/0001.jpg", "第2季/第2話/0001.jpg"}
+	if len(idx) != len(want) {
+		t.Fatalf("got %v want %v", idx, want)
+	}
+	for i := range want {
+		if idx[i] != want[i] {
+			t.Fatalf("got %v want %v", idx, want)
+		}
+	}
+}
+
 func TestPageIndexRejectsSlip(t *testing.T) {
 	for _, bad := range []string{"../evil.jpg", "/etc/passwd.jpg", "a\\..\\b.jpg", "pag\ne.jpg", "pag\re.jpg"} {
 		r := zipOf(t, bad)
@@ -2,6 +2,7 @@ package scanner

 import (
 	"context"
+	"errors"
 	"fmt"
 	"io"
 	"io/fs"
@@ -146,6 +147,9 @@ func (s *Scanner) add(ctx context.Context, libID int64, root, rel string, ds dis
 		idx, err := s.zipIndex(root, rel)
 		pageCount = len(idx)
 		idxErr = err
+		if idxErr == nil && pageCount == 0 { // 视频/文档 zip 不是漫画,空白 reader 没有意义
+			idxErr = errors.New("no images in archive")
+		}
 	}
 	id, err := s.st.InsertBook(ctx, libID, rel, titleOf(rel), format, ds.size, ds.modTS, pageCount)
 	if err != nil {
@@ -167,6 +171,9 @@ func (s *Scanner) update(ctx context.Context, libID, bookID int64, root, rel str
 		idx, err := s.zipIndex(root, rel)
 		pageCount = len(idx)
 		idxErr = err
+		if idxErr == nil && pageCount == 0 { // 视频/文档 zip 不是漫画,空白 reader 没有意义
+			idxErr = errors.New("no images in archive")
+		}
 	}
 	if err := s.st.UpdateBookFile(ctx, bookID, ds.size, ds.modTS, pageCount); err != nil {
 		log.Printf("scan: update %s: %v", rel, err)
@@ -10,11 +10,17 @@ The format loosely follows Keep a Changelog and can be adapted to the team's hab

 ### Added / 新增

+- API: CBZ page indexing now skips macOS packaging junk (`__MACOSX/…` and `._*` AppleDouble files), which used to land in the page list as ~163-byte black "pages"; `GET /api/books/:id/pages` additionally returns `chapters:[{title,start}]` derived from the archive's folder structure (e.g. 第1話…), so per-folder comics expose their real organization.
+- API：CBZ 页索引现会跳过 macOS 打包垃圾（`__MACOSX/…` 与 `._*` 资源叉文件），此前它们以 ~163 字节黑页混入页列表；`GET /api/books/:id/pages` 新增 `chapters:[{title,start}]`，按压缩包内目录结构（如 第1話…）给出真实章节。
+
 - API: resumable chunked upload protocol for large files — `POST /api/libraries/:id/upload/init` (fingerprint-derived deterministic `uploadId`, rejects totals over `UPLOAD_MAX_MB` with `413 too_large`), `PUT /api/uploads/:uid/parts/:index` (parts ≤ 32MB), `GET /api/uploads/:uid` (received parts, for resume), `POST /api/uploads/:uid/complete` (assemble + atomic land, same path contract as single-POST upload). Sessions persist under `BOOKS_DIR/.uploads/` with 24h opportunistic sweep. `UPLOAD_MAX_MB` is now wired through both compose stacks/`.env`; `.env.example` sets 2048 and drops `NGINX_CLIENT_MAX_BODY_SIZE` to 32m (nginx only ever sees one chunk).
 - API: 新增大文件可续传分片上传协议——`POST /api/libraries/:id/upload/init`（按指纹派生确定性 `uploadId`，总量超 `UPLOAD_MAX_MB` 返回 `413 too_large`）、`PUT /api/uploads/:uid/parts/:index`（单片 ≤32MB）、`GET /api/uploads/:uid`（查询已传分片以续传）、`POST /api/uploads/:uid/complete`（拼接后原子落盘，返回与单发上传一致的 `path`）。会话存于 `BOOKS_DIR/.uploads/`，超 24h 顺手清理。`UPLOAD_MAX_MB` 已接入两份 compose/`.env`；`.env.example` 调至 2048 并将 `NGINX_CLIENT_MAX_BODY_SIZE` 降为 32m（nginx 只见单个分片）。

 ### Changed / 变更

+- Scanner: an image-list-less archive (`.zip`/`.cbz` with no page images — video packs, document dumps) is now recorded as `state=error` ("no images in archive") instead of registering as an empty CBZ with a blank reader.
+- 扫描器：不含任何图片条目的 `.zip`/`.cbz`（视频包、文档包）现记录为 `state=error`（"no images in archive"），不再注册成空 CBZ 留下一个白板阅读器。
+
 - API: upload over `UPLOAD_MAX_MB` now returns `413 too_large` with the limit in the message; previously the size abort was misreported as `400 bad_request "multipart field 'file' required"`.
 - API：超过 `UPLOAD_MAX_MB` 的上传现在返回 `413 too_large` 并在消息中带上限额；此前体积超限被误报为 `400 bad_request "multipart field 'file' required"`。
 - API: `POST /api/libraries` now takes only `{name}`; `root_path` is generated server-side as `BOOKS_DIR/<sanitized name>` (no client-supplied paths, validated at creation).
@@ -10,6 +10,9 @@ The format loosely follows Keep a Changelog and can be adapted to the team's hab

 ### Added / 新增

+- Folder-structured comics gain a 目录 sidebar in the CBZ reader: each archive folder (第N話…) becomes a chapter that jumps to its first page, with 上一章/下一章 buttons, a current-chapter counter, and the active row highlighted and scrolled into view — flat archives show no extra UI.
+- 按目录组织的漫画在 CBZ 阅读器新增「目录」侧栏：压缩包内每个文件夹（第N話…）即一章，点击直达该话首页；工具条配上一章/下一章与当前章计数，目录内高亮当前章并自动居中；平铺无目录的包不显示多余控件。
+
 - Library uploads over 16MB now auto-switch to the resumable chunked protocol (8MB parts): failed parts retry in place up to 3 times and already-sent parts are skipped, so a flaky upload continues instead of starting over; small files keep the original single request.
 - 书库上传超过 16MB 自动切换为可续传分片协议（每片 8MB）：失败的分片原地重试至多 3 次，已传片自动跳过，中断后继续而不是从头再来；小文件仍走原单次上传。

@@ -172,7 +172,8 @@ export const api = {
  },
  getBook: (id: number) => apiFetch<Book>(`/books/${id}`),
  deleteBook: (id: number) => apiFetch<void>(`/books/${id}`, { method: "DELETE" }),
-  pageCount: (pagesUrl: string) => apiFetch<{ count: number }>(pagesUrl),
+  pageCount: (pagesUrl: string) =>
+    apiFetch<{ count: number; chapters?: { title: string; start: number }[] | null }>(pagesUrl),

  putProgress: (id: number, locator: Record<string, unknown>, percent: number, keepalive = false) =>
    apiFetch<void>(`/books/${id}/progress`, { method: "PUT", body: { locator, percent }, keepalive }),
@@ -84,6 +84,9 @@ export default function CbzReader({ book, initialLocator, chrome }: ReaderProps)
    retry: 0,
  });
  const count = countQ.data?.count ?? 0;
+  const chapters = countQ.data?.chapters ?? null; // 包内目录结构(第N話/上中下卷)
+  const [toc, setToc] = useState(false);
+  const activeRow = useRef<HTMLButtonElement>(null);
  const saver = useProgressSaver(book.id);
  const boxRef = useRef<HTMLDivElement>(null);
  const [width, setWidth] = useState(480);
@@ -168,6 +171,10 @@ export default function CbzReader({ book, initialLocator, chrome }: ReaderProps)
  }

  const cur = ph.pageAt(top + vh / 2);
+  const ci = chapters ? chapters.reduce((acc, c, i) => (c.start <= cur ? i : acc), 0) : 0;
+  useEffect(() => {
+    if (toc) activeRow.current?.scrollIntoView({ block: "center" });
+  }, [toc]);

  function jump(i: number, smooth = true) {
    const el = boxRef.current;
@@ -284,6 +291,26 @@ export default function CbzReader({ book, initialLocator, chrome }: ReaderProps)
            >
              {MODE_LABEL[mode]}
            </button>
+            {chapters && (
+              <>
+                <button className="rd-btn" aria-expanded={toc} aria-controls="cbz-toc" onClick={() => setToc((v) => !v)}>
+                  目录
+                </button>
+                <span className="text-xs tabular-nums opacity-60">
+                  {ci + 1}/{chapters.length}
+                </span>
+                <button className="rd-btn" disabled={ci === 0} onClick={() => jump(chapters[ci - 1].start)}>
+                  ← 上一章
+                </button>
+                <button
+                  className="rd-btn"
+                  disabled={ci === chapters.length - 1}
+                  onClick={() => jump(chapters[ci + 1].start)}
+                >
+                  下一章 →
+                </button>
+              </>
+            )}
          </div>
        </div>
      ) : (
@@ -294,6 +321,33 @@ export default function CbzReader({ book, initialLocator, chrome }: ReaderProps)
          <span className="opacity-60">{Math.round(((cur + 1) / Math.max(1, count)) * 100)}%</span>
        </div>
      )}
+      {toc && chapters && (
+        <>
+          <div className="fixed inset-0 z-10" aria-hidden="true" onClick={() => setToc(false)} />
+          <nav
+            id="cbz-toc"
+            aria-label="章节目录"
+            className="rd-divider absolute inset-y-0 left-0 z-20 w-64 overflow-y-auto border-r bg-[var(--rd-bg)] p-2 shadow-2xl"
+          >
+            <ul className="space-y-0.5">
+              {chapters.map((c, i) => (
+                <li key={i}>
+                  <button
+                    ref={i === ci ? activeRow : undefined}
+                    className={"rd-row" + (i === ci ? " rd-btn-on" : "")}
+                    onClick={() => {
+                      jump(c.start);
+                      setToc(false);
+                    }}
+                  >
+                    {i + 1}. {c.title}
+                  </button>
+                </li>
+              ))}
+            </ul>
+          </nav>
+        </>
+      )}
    </div>
  );
 }