feat(backend): cover extraction for cbz and epub (opf cover-image/meta)

backend/internal/bookfile/cover.go

@@ -0,0 +1,125 @@
+package bookfile
+
+import (
+	"archive/zip"
+	"encoding/xml"
+	"errors"
+	"io"
+	"net/url"
+	"path"
+	"strings"
+)
+
+func CBZCover(f io.ReaderAt, size int64) ([]byte, string, error) {
+	idx, err := PageIndex(f, size)
+	if err != nil {
+		return nil, "", err
+	}
+	if len(idx) == 0 {
+		return nil, "", errors.New("no page images")
+	}
+	img, err := ReadEntry(f, size, idx[0])
+	return img, strings.ToLower(path.Ext(idx[0])), err
+}
+
+func readZipEntry(zr *zip.Reader, name string) ([]byte, error) {
+	for _, zf := range zr.File {
+		if zf.Name == name {
+			rc, err := zf.Open()
+			if err != nil {
+				return nil, err
+			}
+			defer rc.Close()
+			return io.ReadAll(io.LimitReader(rc, 8<<20))
+		}
+	}
+	return nil, errors.New("entry not found: " + name)
+}
+
+type opfDoc struct {
+	XMLName  xml.Name `xml:"package"`
+	Metadata struct {
+		Meta []struct {
+			Name    string `xml:"name,attr"`
+			Content string `xml:"content,attr"`
+		} `xml:"meta"`
+	} `xml:"metadata"`
+	Manifest struct {
+		Items []struct {
+			ID         string `xml:"id,attr"`
+			Href       string `xml:"href,attr"`
+			Properties string `xml:"properties,attr"`
+			MediaType  string `xml:"media-type,attr"`
+		} `xml:"item"`
+	} `xml:"manifest"`
+}
+
+func EPUBCover(f io.ReaderAt, size int64) ([]byte, string, error) {
+	zr, err := zip.NewReader(f, size)
+	if err != nil {
+		return nil, "", ErrNotZip
+	}
+	cont, err := readZipEntry(zr, "META-INF/container.xml")
+	if err != nil {
+		return nil, "", err
+	}
+	var root struct {
+		XMLName   xml.Name `xml:"container"`
+		RootFiles []struct {
+			FullPath string `xml:"full-path,attr"`
+		} `xml:"rootfiles>rootfile"`
+	}
+	if err := xml.Unmarshal(cont, &root); err != nil || len(root.RootFiles) == 0 {
+		return nil, "", errors.New("bad container.xml")
+	}
+	opfPath := root.RootFiles[0].FullPath
+	opfData, err := readZipEntry(zr, opfPath)
+	if err != nil {
+		return nil, "", err
+	}
+	var doc opfDoc
+	if err := xml.Unmarshal(opfData, &doc); err != nil {
+		return nil, "", errors.New("bad opf")
+	}
+	metaCoverID := ""
+	for _, m := range doc.Metadata.Meta {
+		if m.Name == "cover" {
+			metaCoverID = m.Content
+		}
+	}
+	href := ""
+	for _, it := range doc.Manifest.Items {
+		if strings.Contains(it.Properties, "cover-image") {
+			href = it.Href
+			break
+		}
+		if metaCoverID != "" && it.ID == metaCoverID {
+			href = it.Href
+			break
+		}
+	}
+	if href == "" {
+		for _, it := range doc.Manifest.Items {
+			if strings.HasPrefix(it.MediaType, "image/") && strings.Contains(strings.ToLower(it.ID), "cover") {
+				href = it.Href
+				break
+			}
+		}
+	}
+	if href == "" {
+		return nil, "", errors.New("no cover found")
+	}
+	unescaped, e := url.PathUnescape(href)
+	if e == nil {
+		href = unescaped
+	}
+	name := path.Join(path.Dir(opfPath), href)
+	if name == "." {
+		name = href
+	}
+	img, err := readZipEntry(zr, name)
+	if err != nil {
+		return nil, "", err
+	}
+	return img, strings.ToLower(path.Ext(name)), nil
+}

backend/internal/bookfile/cover_test.go

@@ -0,0 +1,75 @@
+package bookfile
+
+import (
+	"archive/zip"
+	"bytes"
+	"testing"
+)
+
+func TestCBZCoverIsFirstPage(t *testing.T) {
+	buf := &bytes.Buffer{}
+	zw := zip.NewWriter(buf)
+	w, _ := zw.Create("02.jpg")
+	w.Write([]byte("second"))
+	w, _ = zw.Create("01.jpg")
+	w.Write([]byte("first"))
+	zw.Close()
+	r := bytes.NewReader(buf.Bytes())
+	img, ext, err := CBZCover(r, int64(r.Len()))
+	if err != nil || string(img) != "first" || ext != ".jpg" {
+		t.Fatalf("%q %q %v", img, ext, err)
+	}
+}
+
+func zipEntries(t *testing.T, kv map[string]string) *bytes.Reader {
+	t.Helper()
+	buf := &bytes.Buffer{}
+	zw := zip.NewWriter(buf)
+	for name, data := range kv {
+		w, _ := zw.Create(name)
+		w.Write([]byte(data))
+	}
+	zw.Close()
+	return bytes.NewReader(buf.Bytes())
+}
+
+const containerXML = `<?xml version="1.0"?><container><rootfiles><rootfile full-path="OEBPS/content.opf"/></rootfiles></container>`
+
+func opf(manifest, spine string) string {
+	return `<?xml version="1.0"?><package xmlns="http://www.idpf.org/2007/opf"><metadata>` +
+		spine + `</metadata><manifest>` + manifest + `</manifest></package>`
+}
+
+func TestEPUBCoverProperties(t *testing.T) {
+	r := zipEntries(t, map[string]string{
+		"META-INF/container.xml":  containerXML,
+		"OEBPS/content.opf":       opf(`<item id="c" href="img/cover%20one.png" properties="cover-image" media-type="image/png"/><item id="t" href="t.xhtml" media-type="application/xhtml+xml"/>`, ""),
+		"OEBPS/img/cover one.png": "PNGDATA",
+	})
+	img, ext, err := EPUBCover(r, int64(r.Len()))
+	if err != nil || string(img) != "PNGDATA" || ext != ".png" {
+		t.Fatalf("%q %q %v", img, ext, err)
+	}
+}
+
+func TestEPUBCoverMetaID(t *testing.T) {
+	r := zipEntries(t, map[string]string{
+		"META-INF/container.xml": containerXML,
+		"OEBPS/content.opf":      opf(`<item id="cv" href="art.jpg" media-type="image/jpeg"/><item id="t" href="t.xhtml" media-type="application/xhtml+xml"/>`, `<meta name="cover" content="cv"/>`),
+		"OEBPS/art.jpg":          "JPGDATA",
+	})
+	img, _, err := EPUBCover(r, int64(r.Len()))
+	if err != nil || string(img) != "JPGDATA" {
+		t.Fatalf("%q %v", img, err)
+	}
+}
+
+func TestEPUBNoCover(t *testing.T) {
+	r := zipEntries(t, map[string]string{
+		"META-INF/container.xml": containerXML,
+		"OEBPS/content.opf":      opf(`<item id="t" href="t.xhtml" media-type="application/xhtml+xml"/>`, ""),
+	})
+	if _, _, err := EPUBCover(r, int64(r.Len())); err == nil {
+		t.Fatal("want no-cover error")
+	}
+}