From 7eabaf0de619bfcc1c3d90ae94b96d746321826b Mon Sep 17 00:00:00 2001
From: n1474335 <n1474335@gmail.com>
Date: Fri, 13 Sep 2019 14:34:08 +0100
Subject: [PATCH] Cleaned up and improved OCR operation

---
 src/core/config/Categories.json               |  4 +-
 ...CR.mjs => OpticalCharacterRecognition.mjs} | 55 ++++++++++---------
 tests/operations/tests/Image.mjs              | 17 +++++-
 3 files changed, 48 insertions(+), 28 deletions(-)
 rename src/core/operations/{OCR.mjs => OpticalCharacterRecognition.mjs} (50%)

diff --git a/src/core/config/Categories.json b/src/core/config/Categories.json
index b990393c..94f7fd30 100755
--- a/src/core/config/Categories.json
+++ b/src/core/config/Categories.json
@@ -383,6 +383,7 @@
         "ops": [
             "Render Image",
             "Play Media",
+            "Optical Character Recognition",
             "Remove EXIF",
             "Extract EXIF",
             "Split Colour Channels",
@@ -405,8 +406,7 @@
             "Hex Density chart",
             "Scatter chart",
             "Series chart",
-            "Heatmap chart",
-            "OCR"
+            "Heatmap chart"
         ]
     },
     {
diff --git a/src/core/operations/OCR.mjs b/src/core/operations/OpticalCharacterRecognition.mjs
similarity index 50%
rename from src/core/operations/OCR.mjs
rename to src/core/operations/OpticalCharacterRecognition.mjs
index 5b4e6d79..ccfbcd1f 100644
--- a/src/core/operations/OCR.mjs
+++ b/src/core/operations/OpticalCharacterRecognition.mjs
@@ -1,4 +1,5 @@
 /**
+ * @author n1474335 [n1474335@gmail.com]
  * @author mshwed [m@ttshwed.com]
  * @copyright Crown Copyright 2019
  * @license Apache-2.0
@@ -7,30 +8,36 @@
 import Operation from "../Operation.mjs";
 import OperationError from "../errors/OperationError.mjs";
 import { isImage } from "../lib/FileType.mjs";
+import { toBase64 } from "../lib/Base64.mjs";
 import { isWorkerEnvironment } from "../Utils.mjs";
 
-import jimp from "jimp";
 import Tesseract from "tesseract.js";
 const { TesseractWorker } = Tesseract;
 
 /**
- * OCR operation
+ * Optical Character Recognition operation
  */
-class OCR extends Operation {
+class OpticalCharacterRecognition extends Operation {
 
     /**
-     * OCR constructor
+     * OpticalCharacterRecognition constructor
      */
     constructor() {
         super();
 
-        this.name = "OCR";
-        this.module = "Default";
-        this.description = "Optical character recognition or optical character reader (OCR) is the mechanical or electronic conversion of images of typed, handwritten or printed text into machine-encoded text.";
-        this.infoURL = "https://en.wikipedia.org/wiki/Optical_character_recognition";
+        this.name = "Optical Character Recognition";
+        this.module = "Image";
+        this.description = "Optical character recognition or optical character reader (OCR) is the mechanical or electronic conversion of images of typed, handwritten or printed text into machine-encoded text.<br><br>Supported image formats: png, jpg, bmp, pbm.";
+        this.infoURL = "https://wikipedia.org/wiki/Optical_character_recognition";
         this.inputType = "ArrayBuffer";
         this.outputType = "string";
-        this.args = [];
+        this.args = [
+            {
+                name: "Show confidence",
+                type: "boolean",
+                value: true
+            }
+        ];
     }
 
     /**
@@ -39,34 +46,32 @@ class OCR extends Operation {
      * @returns {string}
      */
     async run(input, args) {
-        if (!isImage(input)) {
+        const [showConfidence] = args;
+
+        const type = isImage(input);
+        if (!type) {
             throw new OperationError("Invalid File Type");
         }
 
         try {
-            if (isWorkerEnvironment())
-                self.sendStatusMessage("Performing OCR on image...");
-
-            let image;
-            try {
-                image = await jimp.read(input);
-                image = await image.getBase64Async(jimp.AUTO);
-            } catch (err) {
-                throw new OperationError(`Error loading image. (${err})`);
-            }
-
+            const image = `data:${type};base64,${toBase64(input)}`;
             const worker = new TesseractWorker();
-
             const result = await worker.recognize(image)
                 .progress(progress => {
-                    if (isWorkerEnvironment()) self.sendStatusMessage(`${progress.status} - ${(parseFloat(progress.progress)*100).toFixed(2)}%`);
+                    if (isWorkerEnvironment()) {
+                        self.sendStatusMessage(`Status: ${progress.status} - ${(parseFloat(progress.progress)*100).toFixed(2)}%`);
+                    }
                 });
 
-            return result.text;
+            if (showConfidence) {
+                return `Confidence: ${result.confidence}%\n\n${result.text}`;
+            } else {
+                return result.text;
+            }
         } catch (err) {
             throw new OperationError(`Error performing OCR on image. (${err})`);
         }
     }
 }
 
-export default OCR;
+export default OpticalCharacterRecognition;
diff --git a/tests/operations/tests/Image.mjs b/tests/operations/tests/Image.mjs
index ea5c64a4..307420f2 100644
--- a/tests/operations/tests/Image.mjs
+++ b/tests/operations/tests/Image.mjs
@@ -247,5 +247,20 @@ TestRegister.addTests([
                 args: ["None"]
             }
         ]
-    }
+    },
+    /*{ Commented out as it takes a while to run and drops a file to disk (eng.traineddata)
+        name: "Optical Character Recognition",
+        input: "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",
+        expectedOutput: "Tesseract.js\n",
+        recipeConfig: [
+            {
+                "op": "From Base64",
+                "args": ["A-Za-z0-9+/=", true]
+            },
+            {
+                "op": "Optical Character Recognition",
+                "args": [false]
+            }
+        ]
+    }*/
 ]);