feat(converter): add --keep-filenames flag to preserve original page filenames

Add a new `--keep-filenames` boolean flag (no shorthand, default false) on
both `optimize` and `watch` commands. When enabled, page entries in the
output CBZ preserve their original base filenames instead of being
renumbered to `%04d` sequential indices. Extension is swapped on format
conversion (e.g. `001.png` → `001.webp`), split pages get `-NN` suffixes
on the original stem, and duplicate base names fall back to
`<stem>_<index>.<ext>`. Watch mode supports the flag via the
`keep-filenames` viper config key.

Design: data-driven `OriginalName` field on `manga.PageFile`, set at
extraction in `cbz_loader.ExtractChapter`. The `Converter` interface is
unchanged. Flag off → behavior identical to before.

Closes #216
This commit is contained in:
Antoine Aflalo
2026-07-21 19:38:17 -04:00
parent 650623b340
commit f3ba585c13
12 changed files with 245 additions and 37 deletions
+7
View File
@@ -58,6 +58,12 @@ cbzconverter optimize [folder] --format=webp
cbzconverter optimize [folder] --format WEBP
```
Preserve original page filenames in the output CBZ instead of sequential renumbering:
```sh
cbzconverter optimize [folder] --keep-filenames --quality 85 --format webp
```
With timeout to avoid hanging on problematic chapters:
```sh
@@ -103,6 +109,7 @@ docker run -v /path/to/comics:/comics ghcr.io/belphemur/cbzoptimizer:latest watc
- `--format`, `-f`: Format to convert the images to (currently supports: webp). Default is webp.
- Can be specified as: `--format webp`, `-f webp`, or `--format=webp`
- Case-insensitive: `webp`, `WEBP`, and `WebP` are all valid
- `--keep-filenames`: Preserve each page's original base filename in the output CBZ (with the extension swapped for format conversion) instead of renumbering pages to `0000`, `0001`, ... Default is false. When two source pages share the same base filename, the second one falls back to the indexed form so the output stays a valid zip.
- `--timeout`, `-t`: Maximum time allowed for converting a single chapter (e.g., 30s, 5m, 1h). 0 means no timeout. Default is 0.
- `--backfill`: *(`watch` only)* Optimize CBZ/CBR files that already exist in the watched folder at startup, before watching for new changes. Default is false.
- `--log`, `-l`: Set log level; can be 'panic', 'fatal', 'error', 'warn', 'info', 'debug', or 'trace'. Default is info.
+14
View File
@@ -70,6 +70,19 @@ func setupSplitFlag(cmd *cobra.Command, defaultValue bool, bindViper bool) {
}
}
// setupKeepFilenamesFlag sets up the keep-filenames flag for a command.
//
// Parameters:
// - cmd: The Cobra command to add the keep-filenames flag to
// - defaultValue: The default keep-filenames value
// - bindViper: If true, binds the flag to viper for configuration file support
func setupKeepFilenamesFlag(cmd *cobra.Command, defaultValue bool, bindViper bool) {
cmd.Flags().Bool("keep-filenames", defaultValue, "Preserve original page filenames instead of renumbering to sequential indices")
if bindViper {
_ = viper.BindPFlag("keep-filenames", cmd.Flags().Lookup("keep-filenames"))
}
}
// setupTimeoutFlag sets up the timeout flag for a command.
//
// Parameters:
@@ -96,5 +109,6 @@ func setupCommonFlags(cmd *cobra.Command, converterType *constant.ConversionForm
setupQualityFlag(cmd, qualityDefault, bindViper)
setupOverrideFlag(cmd, overrideDefault, bindViper)
setupSplitFlag(cmd, splitDefault, bindViper)
setupKeepFilenamesFlag(cmd, false, bindViper)
setupTimeoutFlag(cmd, bindViper)
}
+16 -8
View File
@@ -73,6 +73,13 @@ func ConvertCbzCommand(cmd *cobra.Command, args []string) error {
}
log.Debug().Bool("split", split).Msg("Split parameter parsed")
keepFilenames, err := cmd.Flags().GetBool("keep-filenames")
if err != nil {
log.Error().Err(err).Msg("Failed to parse keep-filenames flag")
return fmt.Errorf("invalid keep-filenames value")
}
log.Debug().Bool("keep-filenames", keepFilenames).Msg("Keep-filenames parameter parsed")
timeout, err := cmd.Flags().GetDuration("timeout")
if err != nil {
log.Error().Err(err).Msg("Failed to parse timeout flag")
@@ -121,14 +128,15 @@ func ConvertCbzCommand(cmd *cobra.Command, args []string) error {
log.Debug().Int("worker_id", workerID).Msg("Worker started")
for path := range fileChan {
log.Debug().Int("worker_id", workerID).Str("file_path", path).Msg("Worker processing file")
err := utils2.Optimize(&utils2.OptimizeOptions{
ChapterConverter: chapterConverter,
Path: path,
Quality: quality,
Override: override,
Split: split,
Timeout: timeout,
})
err := utils2.Optimize(&utils2.OptimizeOptions{
ChapterConverter: chapterConverter,
Path: path,
Quality: quality,
Override: override,
Split: split,
KeepFilenames: keepFilenames,
Timeout: timeout,
})
if err != nil {
log.Error().Int("worker_id", workerID).Str("file_path", path).Err(err).Msg("Worker encountered error")
errMutex.Lock()
@@ -86,6 +86,7 @@ func TestConvertCbzCommand(t *testing.T) {
cmd.Flags().IntP("parallelism", "n", 2, "Number of chapters to convert in parallel")
cmd.Flags().BoolP("override", "o", false, "Override the original CBZ/CBR files")
cmd.Flags().BoolP("split", "s", false, "Split long pages into smaller chunks")
cmd.Flags().Bool("keep-filenames", false, "Preserve original page filenames instead of renumbering to sequential indices")
cmd.Flags().DurationP("timeout", "t", 0, "Maximum time allowed for converting a single chapter (e.g., 30s, 5m, 1h). 0 means no timeout")
// Execute the command
@@ -197,6 +198,7 @@ func setupTestCommand(t *testing.T) (*cobra.Command, func()) {
cmd.Flags().IntP("parallelism", "n", 1, "Number of chapters to convert in parallel")
cmd.Flags().BoolP("override", "o", false, "Override the original CBZ/CBR files")
cmd.Flags().BoolP("split", "s", false, "Split long pages into smaller chunks")
cmd.Flags().Bool("keep-filenames", false, "Preserve original page filenames instead of renumbering to sequential indices")
cmd.Flags().DurationP("timeout", "t", 0, "Maximum time allowed for converting a single chapter")
// Reset converterType to default before test for consistency
@@ -433,6 +435,7 @@ func TestConvertCbzCommand_ManyFiles_NoDeadlock(t *testing.T) {
cmd.Flags().IntP("parallelism", "n", 2, "Number of chapters to convert in parallel")
cmd.Flags().BoolP("override", "o", false, "Override the original CBZ/CBR files")
cmd.Flags().BoolP("split", "s", false, "Split long pages into smaller chunks")
cmd.Flags().Bool("keep-filenames", false, "Preserve original page filenames instead of renumbering to sequential indices")
cmd.Flags().DurationP("timeout", "t", 0, "Maximum time allowed for converting a single chapter")
converterType = constant.DefaultConversion
@@ -535,6 +538,7 @@ func TestConvertCbzCommand_HighParallelism_NoDeadlock(t *testing.T) {
cmd.Flags().IntP("parallelism", "n", 8, "Number of chapters to convert in parallel")
cmd.Flags().BoolP("override", "o", false, "Override the original CBZ/CBR files")
cmd.Flags().BoolP("split", "s", false, "Split long pages into smaller chunks")
cmd.Flags().Bool("keep-filenames", false, "Preserve original page filenames instead of renumbering to sequential indices")
cmd.Flags().DurationP("timeout", "t", 0, "Maximum time allowed for converting a single chapter")
converterType = constant.DefaultConversion
+4 -1
View File
@@ -65,6 +65,8 @@ func WatchCommand(_ *cobra.Command, args []string) error {
split := viper.GetBool("split")
keepFilenames := viper.GetBool("keep-filenames")
timeout := viper.GetDuration("timeout")
backfill := viper.GetBool("backfill")
@@ -79,7 +81,7 @@ func WatchCommand(_ *cobra.Command, args []string) error {
if err != nil {
return fmt.Errorf("failed to prepare converter: %w", err)
}
log.Info().Str("path", path).Bool("override", override).Uint8("quality", quality).Str("format", converterType.String()).Bool("split", split).Bool("backfill", backfill).Msg("Watching directory")
log.Info().Str("path", path).Bool("override", override).Uint8("quality", quality).Str("format", converterType.String()).Bool("split", split).Bool("keep_filenames", keepFilenames).Bool("backfill", backfill).Msg("Watching directory")
watcher, err := fsnotify.NewWatcher()
if err != nil {
@@ -98,6 +100,7 @@ func WatchCommand(_ *cobra.Command, args []string) error {
Quality: quality,
Override: override,
Split: split,
KeepFilenames: keepFilenames,
Timeout: timeout,
})
defer queue.Stop()
+52 -7
View File
@@ -5,6 +5,8 @@ import (
"fmt"
"io"
"os"
"path/filepath"
"strings"
"time"
"github.com/belphemur/CBZOptimizer/v2/internal/manga"
@@ -12,6 +14,49 @@ import (
"github.com/rs/zerolog/log"
)
// resolvePageName picks the final archive entry name for a page.
//
// Order of resolution:
// 1. When page.OriginalName is set (keep-filenames mode), use its stem with
// the current page.Extension. Split pages append the -NN suffix so all
// parts of the same source still land in the archive.
// 2. If that final name has already been used in this archive, fall back to
// the indexed form (stem + "_%04d" + extension) so the zip stays valid.
// 3. Otherwise (OriginalName is empty), use the historical %04d / %04d-NN
// sequential naming.
//
// usedNames tracks every name already chosen for this archive and is mutated
// in place so the caller can keep a single map across the whole chapter.
func resolvePageName(page *manga.PageFile, usedNames map[string]struct{}) string {
if page.OriginalName != "" {
stem := strings.TrimSuffix(page.OriginalName, filepath.Ext(page.OriginalName))
var candidate string
if page.IsSplitted {
candidate = fmt.Sprintf("%s-%02d%s", stem, page.SplitPartIndex, page.Extension)
} else {
candidate = stem + page.Extension
}
if _, taken := usedNames[candidate]; !taken {
usedNames[candidate] = struct{}{}
return candidate
}
// Collision: fall back to the indexed form so the entry still gets
// written, but make it visibly distinct from the preserved one.
fallback := fmt.Sprintf("%s_%04d%s", stem, page.Index, page.Extension)
usedNames[fallback] = struct{}{}
return fallback
}
if page.IsSplitted {
name := fmt.Sprintf("%04d-%02d%s", page.Index, page.SplitPartIndex, page.Extension)
usedNames[name] = struct{}{}
return name
}
name := fmt.Sprintf("%04d%s", page.Index, page.Extension)
usedNames[name] = struct{}{}
return name
}
// WriteChapterToCBZ creates a CBZ file from a Chapter by streaming page files
// from disk directly into the zip archive. No image data is held in memory.
func WriteChapterToCBZ(chapter *manga.Chapter, outputFilePath string) (err error) {
@@ -34,14 +79,14 @@ func WriteChapterToCBZ(chapter *manga.Chapter, outputFilePath string) (err error
zipWriter := zip.NewWriter(zipFile)
defer errs.Capture(&err, zipWriter.Close, "failed to close .cbz writer")
// Write each page to the archive by streaming from disk
// Write each page to the archive by streaming from disk.
// Final name resolution: when a page carries an OriginalName (recorded by
// ExtractChapter when --keep-filenames is on), preserve its stem and only
// swap the extension to the current page.Extension. Duplicates in the
// archive fall back to the indexed naming so the output stays a valid zip.
usedNames := make(map[string]struct{}, len(chapter.Pages))
for _, page := range chapter.Pages {
var fileName string
if page.IsSplitted {
fileName = fmt.Sprintf("%04d-%02d%s", page.Index, page.SplitPartIndex, page.Extension)
} else {
fileName = fmt.Sprintf("%04d%s", page.Index, page.Extension)
}
fileName := resolvePageName(page, usedNames)
log.Debug().
Str("output_path", outputFilePath).
+64
View File
@@ -101,6 +101,70 @@ func TestWriteChapterToCBZ(t *testing.T) {
},
expectedFiles: []string{"0000-01.jpg"},
},
{
// --keep-filenames: the page stem is reused as-is with the
// current Extension, regardless of the source extension.
name: "Preserve original filename, extension swap",
chapter: func(t *testing.T, dir string) *manga.Chapter {
return &manga.Chapter{
Pages: []*manga.PageFile{
{
Index: 0,
Extension: ".webp",
FilePath: createTempPage(t, dir, "image data", ".webp"),
OriginalName: "page01.png",
},
},
}
},
expectedFiles: []string{"page01.webp"},
},
{
// --keep-filenames for a split page: stem is preserved, the
// split part index still gets appended to keep parts distinct.
name: "Split page with OriginalName",
chapter: func(t *testing.T, dir string) *manga.Chapter {
return &manga.Chapter{
Pages: []*manga.PageFile{
{
Index: 0,
Extension: ".webp",
FilePath: createTempPage(t, dir, "split data", ".webp"),
IsSplitted: true,
SplitPartIndex: 2,
OriginalName: "tall.png",
},
},
}
},
expectedFiles: []string{"tall-02.webp"},
},
{
// Two pages share the same OriginalName (rare but possible
// with subdirectories): the first keeps the stem, the second
// falls back to the indexed form to avoid duplicate zip
// entries.
name: "Duplicate OriginalName falls back to indexed form",
chapter: func(t *testing.T, dir string) *manga.Chapter {
return &manga.Chapter{
Pages: []*manga.PageFile{
{
Index: 0,
Extension: ".webp",
FilePath: createTempPage(t, dir, "first", ".webp"),
OriginalName: "cover.png",
},
{
Index: 1,
Extension: ".webp",
FilePath: createTempPage(t, dir, "second", ".webp"),
OriginalName: "cover.png",
},
},
}
},
expectedFiles: []string{"cover.webp", "cover_0001.webp"},
},
}
for _, tc := range testCases {
+14 -4
View File
@@ -142,7 +142,13 @@ func IsAlreadyConverted(ctx context.Context, filePath string) (converted bool, e
// ExtractChapter extracts an archive (CBZ/CBR) to a temp directory on disk.
// Pages are streamed directly to files — no image data is held in memory.
// Returns a Chapter with PageFile entries pointing to extracted files.
func ExtractChapter(ctx context.Context, filePath string) (*manga.Chapter, error) {
//
// When keepFilenames is true, each PageFile has its OriginalName set to the
// base filename of the entry inside the archive. Downstream code uses that
// name to preserve the original page identity in the output CBZ (with the
// extension swapped for format conversion). When false, OriginalName stays
// empty and the sequential %04d naming convention is used instead.
func ExtractChapter(ctx context.Context, filePath string, keepFilenames bool) (*manga.Chapter, error) {
log.Debug().Str("file_path", filePath).Msg("Extracting chapter to disk")
// Create temp directory for extraction
@@ -279,6 +285,9 @@ func ExtractChapter(ctx context.Context, filePath string) (*manga.Chapter, error
Extension: ext,
FilePath: outputPath,
}
if keepFilenames {
page.OriginalName = filepath.Base(path)
}
chapter.Pages = append(chapter.Pages, page)
log.Debug().
@@ -320,8 +329,9 @@ func isJunkFile(path string) bool {
}
// LoadChapter extracts the chapter from a CBZ/CBR file to disk.
// It delegates to ExtractChapter and always extracts all pages.
// Use IsAlreadyConverted for a fast conversion status check without extraction.
// It delegates to ExtractChapter with keepFilenames=false and always
// extracts all pages. Use IsAlreadyConverted for a fast conversion status
// check without extraction.
func LoadChapter(filePath string) (*manga.Chapter, error) {
return ExtractChapter(context.Background(), filePath)
return ExtractChapter(context.Background(), filePath, false)
}
+31 -12
View File
@@ -157,12 +157,12 @@ func TestIsAlreadyConverted_CBZWithNonDateComment(t *testing.T) {
}
func TestExtractChapter_NonexistentFile(t *testing.T) {
_, err := ExtractChapter(context.Background(), "/nonexistent/file.cbz")
_, err := ExtractChapter(context.Background(), "/nonexistent/file.cbz", false)
require.Error(t, err)
}
func TestExtractChapter_PageExtensions(t *testing.T) {
chapter, err := ExtractChapter(context.Background(), "../../testdata/Chapter 128.cbz")
chapter, err := ExtractChapter(context.Background(), "../../testdata/Chapter 128.cbz", false)
require.NoError(t, err)
defer func() { _ = chapter.Cleanup() }()
@@ -174,17 +174,36 @@ func TestExtractChapter_PageExtensions(t *testing.T) {
}
func TestExtractChapter_PagesHaveSequentialIndices(t *testing.T) {
chapter, err := ExtractChapter(context.Background(), "../../testdata/Chapter 128.cbz")
require.NoError(t, err)
defer func() { _ = chapter.Cleanup() }()
t.Run("default sequential naming", func(t *testing.T) {
chapter, err := ExtractChapter(context.Background(), "../../testdata/Chapter 128.cbz", false)
require.NoError(t, err)
defer func() { _ = chapter.Cleanup() }()
for i, page := range chapter.Pages {
assert.Equal(t, uint16(i), page.Index)
}
for i, page := range chapter.Pages {
assert.Equal(t, uint16(i), page.Index)
assert.Empty(t, page.OriginalName, "keep-filenames disabled: OriginalName must stay empty")
}
})
t.Run("keep-filenames records OriginalName", func(t *testing.T) {
chapter, err := ExtractChapter(context.Background(), "../../testdata/Chapter 128.cbz", true)
require.NoError(t, err)
defer func() { _ = chapter.Cleanup() }()
require.NotEmpty(t, chapter.Pages)
for i, page := range chapter.Pages {
assert.Equal(t, uint16(i), page.Index)
assert.NotEmpty(t, page.OriginalName, "keep-filenames enabled: OriginalName must be set for every page")
// OriginalName must be a bare base name (no directory prefix), even
// when the source archive nests pages in subdirectories.
assert.Equal(t, filepath.Base(page.OriginalName), page.OriginalName,
"OriginalName should be a base filename with no path components")
}
})
}
func TestExtractChapter_Cleanup(t *testing.T) {
chapter, err := ExtractChapter(context.Background(), "../../testdata/Chapter 128.cbz")
chapter, err := ExtractChapter(context.Background(), "../../testdata/Chapter 128.cbz", false)
require.NoError(t, err)
tempDir := chapter.TempDir
@@ -199,7 +218,7 @@ func TestExtractChapter_Cleanup(t *testing.T) {
}
func TestExtractChapter_CBR(t *testing.T) {
chapter, err := ExtractChapter(context.Background(), "../../testdata/Chapter 1.cbr")
chapter, err := ExtractChapter(context.Background(), "../../testdata/Chapter 1.cbr", false)
require.NoError(t, err)
defer func() { _ = chapter.Cleanup() }()
@@ -212,7 +231,7 @@ func TestExtractChapter_CBR(t *testing.T) {
}
func TestExtractChapter_ConvertedStatus(t *testing.T) {
chapter, err := ExtractChapter(context.Background(), "../../testdata/Chapter 10_converted.cbz")
chapter, err := ExtractChapter(context.Background(), "../../testdata/Chapter 10_converted.cbz", false)
require.NoError(t, err)
defer func() { _ = chapter.Cleanup() }()
@@ -356,7 +375,7 @@ func TestExtractChapter_WithConvertedTxt(t *testing.T) {
_ = w.Close()
_ = f.Close()
chapter, err := ExtractChapter(context.Background(), cbzPath)
chapter, err := ExtractChapter(context.Background(), cbzPath, false)
require.NoError(t, err)
defer func() { _ = chapter.Cleanup() }()
+6
View File
@@ -13,4 +13,10 @@ type PageFile struct {
IsSplitted bool
// SplitPartIndex is the part index when the page was split.
SplitPartIndex uint16
// OriginalName is the base filename (e.g. "page01.png") of the page as
// it appeared in the source archive, recorded when the --keep-filenames
// flag is enabled. Empty when the flag is off or the source name is
// unknown. When set, downstream code uses this stem (with the final
// Extension swapped in) instead of the default %04d sequential name.
OriginalName string
}
+8 -2
View File
@@ -21,7 +21,12 @@ type OptimizeOptions struct {
Quality uint8
Override bool
Split bool
Timeout time.Duration
// KeepFilenames preserves the original base filename of each page inside
// the output CBZ (with the extension swapped for format conversion)
// instead of the historical %04d sequential naming. Off by default so
// existing behavior is unchanged.
KeepFilenames bool
Timeout time.Duration
}
// Optimize optimizes a CBZ/CBR file using the specified converter.
@@ -38,6 +43,7 @@ func Optimize(options *OptimizeOptions) error {
Uint8("quality", options.Quality).
Bool("override", options.Override).
Bool("split", options.Split).
Bool("keep_filenames", options.KeepFilenames).
Msg("Optimization parameters")
// Step 1: Fast conversion check before extracting (new requirement)
@@ -64,7 +70,7 @@ func Optimize(options *OptimizeOptions) error {
extractCtx = context.Background()
}
chapter, err := cbz.ExtractChapter(extractCtx, options.Path)
chapter, err := cbz.ExtractChapter(extractCtx, options.Path, options.KeepFilenames)
if err != nil {
log.Error().Str("file", options.Path).Err(err).Msg("Failed to extract chapter")
return fmt.Errorf("failed to extract chapter: %w", err)
+25 -3
View File
@@ -25,6 +25,26 @@ import (
const webpMaxHeight = 16383
// intermediatePageName returns the on-disk filename used for a page's WebP
// intermediate during conversion. When the page carries an OriginalName
// (recorded by --keep-filenames), its stem is reused so the temp file line
// up with the name the archive writer will pick. Otherwise the historical
// %04d indexed naming is kept. The splitSuffix argument is appended verbatim
// after the stem (e.g. "-00", "-01") for split parts and is empty for the
// happy-path single output. A leading dash is only added when both the
// split suffix and the original-name stem are present, so the indexed form
// stays as %04d-%02d.
func intermediatePageName(page *manga.PageFile, splitSuffix string) string {
if page.OriginalName != "" {
stem := strings.TrimSuffix(page.OriginalName, filepath.Ext(page.OriginalName))
return stem + splitSuffix + ".webp"
}
if splitSuffix == "" {
return fmt.Sprintf("%04d.webp", page.Index)
}
return fmt.Sprintf("%04d%s.webp", page.Index, splitSuffix)
}
type Converter struct {
maxHeight int
cropHeight int
@@ -218,14 +238,16 @@ func (converter *Converter) convertPageFile(ctx context.Context, page *manga.Pag
Str("input", page.FilePath).
Msg("Converting page file")
// If the page is already WebP, just return it as-is
// If the page is already WebP, just return it as-is. The returned page
// keeps any OriginalName set during extraction so the archive writer can
// honor --keep-filenames for the final entry name.
if strings.ToLower(page.Extension) == ".webp" {
log.Debug().Uint16("page_index", page.Index).Msg("Page already WebP, skipping")
return []*manga.PageFile{page}, nil
}
// Try direct file-to-file conversion first (happy path — no memory allocation)
outputPath := filepath.Join(outputDir, fmt.Sprintf("%04d.webp", page.Index))
outputPath := filepath.Join(outputDir, intermediatePageName(page, ""))
err := EncodeFile(page.FilePath, outputPath, uint(quality))
if err == nil {
@@ -316,7 +338,7 @@ func (converter *Converter) splitAndConvert(ctx context.Context, page *manga.Pag
partHeight = height - yOffset
}
outputPath := filepath.Join(outputDir, fmt.Sprintf("%04d-%02d.webp", page.Index, i))
outputPath := filepath.Join(outputDir, intermediatePageName(page, fmt.Sprintf("-%02d", i)))
err := EncodeFileWithCrop(page.FilePath, outputPath, uint(quality), 0, yOffset, width, partHeight)
if err != nil {