Merge pull request #18204 from calixteman/issue16782

Fix decoding of JPX images having an alpha channel

src/core/base_stream.js

@@ -49,8 +49,8 @@ class BaseStream {
    *       to be fully loaded, since otherwise intermittent errors may occur;
    *       note the `ObjectLoader` class.
    */
-  async getImageData(length, ignoreColorSpace) {
+  async getImageData(length, decoderOptions) {
-    return this.getBytes(length, ignoreColorSpace);
+    return this.getBytes(length, decoderOptions);
   }
 
   async asyncGetBytes() {

src/core/colorspace.js

@@ -370,6 +370,8 @@ class ColorSpace {
         case "RGB":
         case "DeviceRGB":
           return this.singletons.rgb;
+        case "DeviceRGBA":
+          return this.singletons.rgba;
         case "CMYK":
         case "DeviceCMYK":
           return this.singletons.cmyk;
@@ -511,6 +513,9 @@ class ColorSpace {
       get rgb() {
         return shadow(this, "rgb", new DeviceRgbCS());
       },
+      get rgba() {
+        return shadow(this, "rgba", new DeviceRgbaCS());
+      },
       get cmyk() {
         return shadow(this, "cmyk", new DeviceCmykCS());
       },
@@ -778,6 +783,23 @@ class DeviceRgbCS extends ColorSpace {
   }
 }
 
+/**
+ * The default color is `new Float32Array([0, 0, 0, 1])`.
+ */
+class DeviceRgbaCS extends ColorSpace {
+  constructor() {
+    super("DeviceRGBA", 4);
+  }
+
+  getOutputLength(inputLength, _alpha01) {
+    return inputLength * 4;
+  }
+
+  isPassthrough(bits) {
+    return bits === 8;
+  }
+}
+
 /**
  * The default color is `new Float32Array([0, 0, 0, 1])`.
  */

src/core/decode_stream.js

@@ -73,7 +73,7 @@ class DecodeStream extends BaseStream {
     return this.buffer[this.pos++];
   }
 
-  getBytes(length, ignoreColorSpace = false) {
+  getBytes(length, decoderOptions = null) {
     const pos = this.pos;
     let end;
 
@@ -82,7 +82,7 @@ class DecodeStream extends BaseStream {
       end = pos + length;
 
       while (!this.eof && this.bufferLength < end) {
-        this.readBlock(ignoreColorSpace);
+        this.readBlock(decoderOptions);
       }
       const bufEnd = this.bufferLength;
       if (end > bufEnd) {
@@ -90,7 +90,7 @@ class DecodeStream extends BaseStream {
       }
     } else {
       while (!this.eof) {
-        this.readBlock(ignoreColorSpace);
+        this.readBlock(decoderOptions);
       }
       end = this.bufferLength;
     }
@@ -99,12 +99,12 @@ class DecodeStream extends BaseStream {
     return this.buffer.subarray(pos, end);
   }
 
-  async getImageData(length, ignoreColorSpace = false) {
+  async getImageData(length, decoderOptions = null) {
     if (!this.canAsyncDecodeImageFromBuffer) {
-      return this.getBytes(length, ignoreColorSpace);
+      return this.getBytes(length, decoderOptions);
     }
     const data = await this.stream.asyncGetBytes();
-    return this.decodeImage(data, ignoreColorSpace);
+    return this.decodeImage(data, decoderOptions);
   }
 
   reset() {

src/core/flate_stream.js

@@ -149,7 +149,7 @@ class FlateStream extends DecodeStream {
     this.codeBuf = 0;
   }
 
-  async getImageData(length, _ignoreColorSpace) {
+  async getImageData(length, _decoderOptions) {
     const data = await this.asyncGetBytes();
     return data?.subarray(0, length) || this.getBytes(length);
   }

src/core/image.js

@@ -18,7 +18,6 @@ import {
   FeatureTest,
   FormatError,
   ImageKind,
-  info,
   warn,
 } from "../shared/util.js";
 import {
@@ -104,7 +103,6 @@ class PDFImage {
     localColorSpaceCache,
   }) {
     this.image = image;
-    let jpxDecode = false;
     const dict = image.dict;
 
     const filter = dict.get("F", "Filter");
@@ -126,7 +124,11 @@ class PDFImage {
           bitsPerComponent: image.bitsPerComponent,
         } = JpxImage.parseImageProperties(image.stream));
         image.stream.reset();
-        jpxDecode = true;
+        this.jpxDecoderOptions = {
+          numComponents: 0,
+          isIndexedColormap: false,
+          smaskInData: dict.has("SMaskInData"),
+        };
         break;
       case "JBIG2Decode":
         image.bitsPerComponent = 1;
@@ -180,23 +182,31 @@ class PDFImage {
 
     if (!this.imageMask) {
       let colorSpace = dict.getRaw("CS") || dict.getRaw("ColorSpace");
-      if (!colorSpace) {
+      const hasColorSpace = !!colorSpace;
-        info("JPX images (which do not require color spaces)");
+      if (!hasColorSpace) {
-        switch (image.numComps) {
+        if (this.jpxDecoderOptions) {
-          case 1:
+          colorSpace = Name.get("DeviceRGBA");
-            colorSpace = Name.get("DeviceGray");
+        } else {
-            break;
+          switch (image.numComps) {
-          case 3:
+            case 1:
-            colorSpace = Name.get("DeviceRGB");
+              colorSpace = Name.get("DeviceGray");
-            break;
+              break;
-          case 4:
+            case 3:
-            colorSpace = Name.get("DeviceCMYK");
+              colorSpace = Name.get("DeviceRGB");
-            break;
+              break;
-          default:
+            case 4:
-            throw new Error(
+              colorSpace = Name.get("DeviceCMYK");
-              `JPX images with ${image.numComps} color components not supported.`
+              break;
-            );
+            default:
+              throw new Error(
+                `Images with ${image.numComps} color components not supported.`
+              );
+          }
         }
+      } else if (this.jpxDecoderOptions?.smaskInData) {
+        // If the jpx image has a color space then it mustn't be used in order
+        // to be able to use the color space that comes from the pdf.
+        colorSpace = Name.get("DeviceRGBA");
       }
 
       this.colorSpace = ColorSpace.parse({
@@ -208,9 +218,13 @@ class PDFImage {
       });
       this.numComps = this.colorSpace.numComps;
 
-      // If the jpx image has a color space then it musn't be used in order to
+      if (this.jpxDecoderOptions) {
-      // be able to use the color space that comes from the pdf.
+        this.jpxDecoderOptions.numComponents = hasColorSpace ? this.numComp : 0;
-      this.ignoreColorSpace = jpxDecode && this.colorSpace.name === "Indexed";
+        // If the jpx image has a color space then it musn't be used in order to
+        // be able to use the color space that comes from the pdf.
+        this.jpxDecoderOptions.isIndexedColormap =
+          this.colorSpace.name === "Indexed";
+      }
     }
 
     this.decode = dict.getArray("D", "Decode");
@@ -691,6 +705,28 @@ class PDFImage {
       isOffscreenCanvasSupported &&
       ImageResizer.needsToBeResized(drawWidth, drawHeight);
 
+    if (this.colorSpace.name === "DeviceRGBA") {
+      imgData.kind = ImageKind.RGBA_32BPP;
+      const imgArray = (imgData.data = await this.getImageBytes(
+        originalHeight * originalWidth * 4,
+        {}
+      ));
+
+      if (isOffscreenCanvasSupported) {
+        if (!mustBeResized) {
+          return this.createBitmap(
+            ImageKind.RGBA_32BPP,
+            drawWidth,
+            drawHeight,
+            imgArray
+          );
+        }
+        return ImageResizer.createImage(imgData, false);
+      }
+
+      return imgData;
+    }
+
     if (!forceRGBA) {
       // If it is a 1-bit-per-pixel grayscale (i.e. black-and-white) image
       // without any complications, we pass a same-sized copy to the main
@@ -994,7 +1030,7 @@ class PDFImage {
     this.image.forceRGB = !!forceRGB;
     const imageBytes = await this.image.getImageData(
       length,
-      this.ignoreColorSpace
+      this.jpxDecoderOptions
     );
 
     // If imageBytes came from a DecodeStream, we're safe to transfer it

src/core/jpx.js

@@ -26,9 +26,10 @@ class JpxError extends BaseException {
 class JpxImage {
   static #module = null;
 
-  static decode(data, ignoreColorSpace = false) {
+  static decode(data, decoderOptions) {
+    decoderOptions ||= {};
     this.#module ||= OpenJPEG({ warn });
-    const imageData = this.#module.decode(data, ignoreColorSpace);
+    const imageData = this.#module.decode(data, decoderOptions);
     if (typeof imageData === "string") {
       throw new JpxError(imageData);
     }

src/core/jpx_stream.js

@@ -41,16 +41,16 @@ class JpxStream extends DecodeStream {
     // directly insert all of its data into `this.buffer`.
   }
 
-  readBlock(ignoreColorSpace) {
+  readBlock(decoderOptions) {
-    this.decodeImage(null, ignoreColorSpace);
+    this.decodeImage(null, decoderOptions);
   }
 
-  decodeImage(bytes, ignoreColorSpace) {
+  decodeImage(bytes, decoderOptions) {
     if (this.eof) {
       return this.buffer;
     }
     bytes ||= this.bytes;
-    this.buffer = JpxImage.decode(bytes, ignoreColorSpace);
+    this.buffer = JpxImage.decode(bytes, decoderOptions);
     this.bufferLength = this.buffer.length;
     this.eof = true;
 

test/pdfs/isssue18194.pdf.link

@@ -0,0 +1 @@
+https://github.com/user-attachments/files/15511847/out2.pdf

test/pdfs/issue11306.pdf.link

@@ -0,0 +1 @@
+https://github.com/mozilla/pdf.js/files/3806127/2018.Wrapped.pdf

test/pdfs/issue16782.pdf.link

@@ -0,0 +1 @@
+https://github.com/mozilla/pdf.js/files/12244643/tt1.pdf

test/test_manifest.json

@@ -10031,5 +10031,30 @@
     "rounds": 1,
     "link": true,
     "type": "eq"
+  },
+  {
+    "id": "issue16782",
+    "file": "pdfs/issue16782.pdf",
+    "md5": "b203313ae62c0dac2585fae1c5fa6f8b",
+    "rounds": 1,
+    "link": true,
+    "type": "eq"
+  },
+  {
+    "id": "issue11306",
+    "file": "pdfs/issue11306.pdf",
+    "md5": "dfc626ee307f9488d74341d21641db9e",
+    "rounds": 1,
+    "link": true,
+    "lastPage": 1,
+    "type": "eq"
+  },
+  {
+    "id": "isssue18194",
+    "file": "pdfs/isssue18194.pdf",
+    "md5": "dbc12624353401deac99d94a2962df4d",
+    "rounds": 1,
+    "link": true,
+    "type": "eq"
   }
 ]