Updated setImage to resolve memory leak per #678

src/worker-script/index.js

@@ -294,7 +294,6 @@ const recognize = async ({
     // When the auto-rotate option is True, setImage is called with no angle,
     // then the angle is calculated by Tesseract and then setImage is re-called.
     // Otherwise, setImage is called once using the user-provided rotateRadiansFinal value.
-    let ptr;
     let rotateRadiansFinal;
     if (options.rotateAuto) {
       // The angle is only detected if auto page segmentation is used
@@ -306,7 +305,7 @@ const recognize = async ({
         api.SetVariable('tessedit_pageseg_mode', String(PSM.AUTO));
       }
 
-      ptr = setImage(TessModule, api, image);
+      setImage(TessModule, api, image);
       api.FindLines();
       const rotateRadiansCalc = api.GetAngle();
 
@@ -318,17 +317,17 @@ const recognize = async ({
       // Small angles (<0.005 radians/~0.3 degrees) are ignored to save on runtime
       if (Math.abs(rotateRadiansCalc) >= 0.005) {
         rotateRadiansFinal = rotateRadiansCalc;
-        ptr = setImage(TessModule, api, image, rotateRadiansFinal);
+        setImage(TessModule, api, image, rotateRadiansFinal);
       } else {
         // Image needs to be reset if run with different PSM setting earlier
         if (psmEdit) {
-          ptr = setImage(TessModule, api, image);
+          setImage(TessModule, api, image);
         }
         rotateRadiansFinal = 0;
       }
     } else {
       rotateRadiansFinal = options.rotateRadians || 0;
-      ptr = setImage(TessModule, api, image, rotateRadiansFinal);
+      setImage(TessModule, api, image, rotateRadiansFinal);
     }
 
     const rec = options.rectangle;
@@ -351,7 +350,6 @@ const recognize = async ({
     }
 
     res.resolve(result);
-    TessModule._free(ptr);
   } catch (err) {
     res.reject(err.toString());
   }
@@ -360,11 +358,10 @@ const recognize = async ({
 
 const detect = async ({ payload: { image } }, res) => {
   try {
-    const ptr = setImage(TessModule, api, image);
+    setImage(TessModule, api, image);
     const results = new TessModule.OSResults();
 
     if (!api.DetectOS(results)) {
-      TessModule._free(ptr);
 
       res.resolve({
         tesseract_script_id: null,
@@ -378,8 +375,6 @@ const detect = async ({ payload: { image } }, res) => {
       const oid = best.orientation_id;
       const sid = best.script_id;
 
-      TessModule._free(ptr);
-
       res.resolve({
         tesseract_script_id: sid,
         script: results.unicharset.get_script_from_script_id(sid),

src/worker-script/utils/setImage.js

@@ -9,55 +9,26 @@ const fileType = require('file-type');
  * @access public
  */
 module.exports = (TessModule, api, image, angle = 0) => {
-  const buf = Buffer.from(Array.from({ ...image, length: Object.keys(image).length }));
-  const type = fileType(buf);
-  let bytesPerPixel = 0;
-  let data = null;
-  let pix = null;
-  let w = 0;
-  let h = 0;
 
-  const exif = buf.slice(0, 500).toString().match(/\x01\x12\x00\x03\x00\x00\x00\x01\x00(.)/)?.[1]?.charCodeAt(0) || 1;
+  const type = fileType(image);
 
-  /*
+  const exif = image.slice(0, 500).toString().match(/\x01\x12\x00\x03\x00\x00\x00\x01\x00(.)/)?.[1]?.charCodeAt(0) || 1;
-   * Leptonica supports uncompressed but not compressed bmp files
+
-   * @see https://github.com/DanBloomberg/leptonica/issues/607#issuecomment-1068802516
+  // /*
-   * We therefore use bmp-js to process all bmp files
+  //  * Leptonica supports some but not all bmp files
-   */
+  //  * @see https://github.com/DanBloomberg/leptonica/issues/607#issuecomment-1068802516
+  //  * We therefore use bmp-js to convert all bmp files into a format Leptonica is known to support
+  //  */
   if (type && type.mime === 'image/bmp') {
+    // Not sure what this line actually does, but removing breaks the function
+    const buf = Buffer.from(Array.from({ ...image, length: Object.keys(image).length }));
     const bmpBuf = bmp.decode(buf);
-    data = TessModule._malloc(bmpBuf.data.length * Uint8Array.BYTES_PER_ELEMENT);
+    TessModule.FS.writeFile("/input", bmp.encode(bmpBuf).data);
-    TessModule.HEAPU8.set(bmpBuf.data, data);
-    w = bmpBuf.width;
-    h = bmpBuf.height;
-    bytesPerPixel = 4;
   } else {
-    const ptr = TessModule._malloc(buf.length * Uint8Array.BYTES_PER_ELEMENT);
+    TessModule.FS.writeFile("/input", image);
-    TessModule.HEAPU8.set(buf, ptr);
-    pix = TessModule._pixReadMem(ptr, buf.length);
-    if (TessModule.getValue(pix + (7 * 4), 'i32') === 0) {
-      /*
-       * Set a yres default value to prevent warning from tesseract
-       * See kMinCredibleResolution in tesseract/src/ccstruct/publictypes.h
-       */
-      TessModule.setValue(pix + (7 * 4), 300, 'i32');
-    }
-    [w, h] = Array(2).fill(0)
-      .map((v, idx) => (
-        TessModule.getValue(pix + (idx * 4), 'i32')
-      ));
   }
 
-  /*
+  api.SetImageFile(exif, angle);
-   * As some image format (ex. bmp) is not supported natiely by tesseract,
+  return;
-   * sometimes it will not return pix directly, but data and bytesPerPixel
+
-   * for another SetImage usage.
-   *
-   */
-  if (data === null) {
-    api.SetImage(pix, undefined, undefined, undefined, undefined, exif, angle);
-  } else {
-    api.SetImage(data, w, h, bytesPerPixel, w * bytesPerPixel, exif, angle);
-  }
-  return data === null ? pix : data;
 };