Fix #789

pdfcpu · Jan 31, 2024 · 96659b7 · 96659b7
1 parent dae09eb
commit 96659b7
Show file tree

Hide file tree

Showing 2 changed files with 55 additions and 24 deletions.
diff --git a/pkg/pdfcpu/extract.go b/pkg/pdfcpu/extract.go
@@ -314,14 +314,7 @@ func prepareExtractImage(sd *types.StreamDict) (string, string, types.Dict, bool
 
 	return filters, lastFilter, d, imgMask
 }
-
-func img(
-	ctx *model.Context,
-	sd *types.StreamDict,
-	thumb, imgMask bool,
-	resourceID, filters, lastFilter string,
-	objNr int) (*model.Image, error) {
-
+func decodeImage(ctx *model.Context, sd *types.StreamDict, filters, lastFilter string, objNr int) error {
 	// CCITTDecoded images / (bit) masks don't have a ColorSpace attribute, but we render image files.
 	if lastFilter == filter.CCITTFax {
 		if _, err := ctx.DereferenceDictEntry(sd.Dict, "ColorSpace"); err != nil {
@@ -332,7 +325,7 @@ func img(
 	if lastFilter == filter.DCT {
 		comp, err := ColorSpaceComponents(ctx.XRefTable, sd)
 		if err != nil {
-			return nil, err
+			return err
 		}
 		sd.CSComponents = comp
 	}
@@ -341,7 +334,7 @@ func img(
 
 	case filter.DCT, filter.JPX, filter.Flate, filter.CCITTFax, filter.RunLength:
 		if err := sd.Decode(); err != nil {
-			return nil, err
+			return err
 		}
 
 	default:
@@ -352,7 +345,25 @@ func img(
 		if log.CLIEnabled() {
 			log.CLI.Println(msg)
 		}
-		return nil, nil
+		return nil
+	}
+
+	return nil
+}
+
+func img(
+	ctx *model.Context,
+	sd *types.StreamDict,
+	thumb, imgMask bool,
+	resourceID, filters, lastFilter string,
+	objNr int) (*model.Image, error) {
+
+	if sd.FilterPipeline == nil {
+		sd.Content = sd.Raw
+	} else {
+		if err := decodeImage(ctx, sd, filters, lastFilter, objNr); err != nil {
+			return nil, err
+		}
 	}
 
 	r, t, err := RenderImage(ctx.XRefTable, sd, thumb, resourceID, objNr)
@@ -383,10 +394,6 @@ func ExtractImage(ctx *model.Context, sd *types.StreamDict, thumb bool, resource
 		return imageStub(ctx, sd, resourceID, filters, lastFilter, decodeParms, thumb, imgMask, objNr)
 	}
 
-	if sd.FilterPipeline == nil {
-		return nil, nil
-	}
-
 	return img(ctx, sd, thumb, imgMask, resourceID, filters, lastFilter, objNr)
 }
 

diff --git a/pkg/pdfcpu/writeImage.go b/pkg/pdfcpu/writeImage.go
@@ -259,7 +259,7 @@ func imageForCMYKWithoutSoftMask(im *PDFImage) image.Image {
 	for y := 0; y < im.h; y++ {
 		for x := 0; x < im.w; x++ {
 			img.Set(x, y, color.CMYK{C: b[i], M: b[i+1], Y: b[i+2], K: b[i+3]})
-			i += 4
+			i += im.comp
 		}
 	}
 
@@ -773,25 +773,45 @@ func renderIndexed(xRefTable *model.XRefTable, im *PDFImage, resourceName string
 }
 
 func renderDeviceN(xRefTable *model.XRefTable, im *PDFImage, resourceName string, cs types.Array) (io.Reader, string, error) {
+	if im.comp <= 4 {
+		switch im.comp {
+		case 1:
+			// Gray
+			return renderDeviceGrayToPNG(im, resourceName)
 
-	switch im.comp {
-	case 1:
+		case 3:
+			// RGB
+			return renderDeviceRGBToPNG(im, resourceName)
+
+		case 4:
+			// CMYK
+			return renderDeviceCMYKToTIFF(im, resourceName)
+		}
+	}
+
+	alternateCS, ok := cs[2].(types.Name)
+	if !ok {
+		return nil, "", nil
+	}
+
+	switch alternateCS {
+	case model.DeviceGrayCS:
 		// Gray
 		return renderDeviceGrayToPNG(im, resourceName)
 
-	case 3:
+	case model.DeviceRGBCS:
 		// RGB
 		return renderDeviceRGBToPNG(im, resourceName)
 
-	case 4:
+	case model.DeviceCMYKCS:
 		// CMYK
 		return renderDeviceCMYKToTIFF(im, resourceName)
 	}
 
 	return nil, "", nil
 }
 
-func renderFlateEncodedImage(xRefTable *model.XRefTable, sd *types.StreamDict, thumb bool, resourceName string, objNr int) (io.Reader, string, error) {
+func renderImage(xRefTable *model.XRefTable, sd *types.StreamDict, thumb bool, resourceName string, objNr int) (io.Reader, string, error) {
 	// If color space is CMYK then write .tif else write .png
 
 	pdfImage, err := pdfImage(xRefTable, sd, thumb, objNr)
@@ -820,7 +840,7 @@ func renderFlateEncodedImage(xRefTable *model.XRefTable, sd *types.StreamDict, t
 
 		default:
 			if log.InfoEnabled() {
-				log.Info.Printf("renderFlateEncodedImage: objNr=%d, unsupported name colorspace %s\n", objNr, cs.String())
+				log.Info.Printf("renderImage: objNr=%d, unsupported name colorspace %s\n", objNr, cs.String())
 			}
 		}
 
@@ -846,7 +866,7 @@ func renderFlateEncodedImage(xRefTable *model.XRefTable, sd *types.StreamDict, t
 
 		default:
 			if log.InfoEnabled() {
-				log.Info.Printf("renderFlateEncodedImage: objNr=%d, unsupported array colorspace %s\n", objNr, csn)
+				log.Info.Printf("renderImage: objNr=%d, unsupported array colorspace %s\n", objNr, csn)
 			}
 		}
 
@@ -908,12 +928,16 @@ func renderDCTToPNG(xRefTable *model.XRefTable, sd *types.StreamDict, thumb bool
 func RenderImage(xRefTable *model.XRefTable, sd *types.StreamDict, thumb bool, resourceName string, objNr int) (io.Reader, string, error) {
 	// Image compression is the last filter in the pipeline.
 
+	if len(sd.FilterPipeline) == 0 {
+		return renderImage(xRefTable, sd, thumb, resourceName, objNr)
+	}
+
 	f := sd.FilterPipeline[len(sd.FilterPipeline)-1].Name
 
 	switch f {
 
 	case filter.Flate, filter.CCITTFax, filter.RunLength:
-		return renderFlateEncodedImage(xRefTable, sd, thumb, resourceName, objNr)
+		return renderImage(xRefTable, sd, thumb, resourceName, objNr)
 
 	case filter.DCT:
 		if sd.CSComponents == 4 {