ocr wip

package.json

@@ -33,7 +33,7 @@
     "archiver": "5.3.1",
     "async-mutex": "0.4.0",
     "axios": "1.2.5",
-    "better-sqlite3": "7.4.5",
+    "better-sqlite3": "8.0.1",
     "canvas": "2.11.0",
     "chokidar": "3.5.3",
     "cls-hooked": "4.2.2",
@@ -95,7 +95,7 @@
   },
   "devDependencies": {
     "cross-env": "7.0.3",
-    "electron": "16.2.8",
+    "electron": "23.0.0-beta.6",
     "electron-builder": "23.6.0",
     "electron-packager": "17.1.1",
     "electron-rebuild": "3.2.9",

src/services/cls.js

@@ -48,6 +48,14 @@ function isEntityEventsDisabled() {
     return !!namespace.get('disableEntityEvents');
 }
 
+function isOcrDisabled() {
+    return !!namespace.get('disableOcr');
+}
+
+function disableOcr() {
+    namespace.set('disableOcr', true);
+}
+
 function getAndClearEntityChangeIds() {
     const entityChangeIds = namespace.get('entityChangeIds') || [];
 
@@ -92,5 +100,7 @@ module.exports = {
     reset,
     getAndClearEntityChangeIds,
     addEntityChange,
-    ignoreEntityChangeIds
+    ignoreEntityChangeIds,
+    isOcrDisabled,
+    disableOcr
 };

src/services/image.js

@@ -12,8 +12,7 @@ const sanitizeFilename = require('sanitize-filename');
 const isSvg = require('is-svg');
 const isAnimated = require('is-animated');
 const htmlSanitizer = require("./html_sanitizer");
-const OCRAD = require('ocrad.js');
+const textExtractingService = require("./text_extracting");
-const Canvas = require('canvas');
 
 async function processImage(uploadBuffer, originalName, shrinkImageSwitch) {
     const compressImages = optionService.getOptionBool("compressImages");
@@ -128,7 +127,7 @@ function saveImage(parentNoteId, uploadBuffer, originalName, shrinkImageSwitch,
             note.setContent(buffer);
         });
 
-        runOcr(note, buffer);
+        textExtractingService.runOcr(note, buffer);
     });
 
     return {

src/services/notes.js

@@ -729,27 +729,12 @@ function scanForLinks(note, content) {
     }
 }
 
-function runOcr(note, buffer) {
-    if (!note.isImage() || !optionService.getOptionBool('ocrImages')) {
-        return;
-    }
-
-    try {
-        const plainText = textExtractingService.ocrTextFromBuffer(buffer);
-
-        note.saveNoteAncillary('plainText', 'text/plain', plainText);
-    }
-    catch (e) {
-        log.error(`OCR on note '${note.noteId}' failed with error '${e.message}', stack ${e.stack}`);
-    }
-}
-
 /**
  * Things which have to be executed after updating content, but asynchronously (separate transaction)
  */
 async function asyncPostProcessContent(note, content) {
     scanForLinks(note, content);
-    runOcr(note, content);
+    await textExtractingService.runOcr(note, content);
     await textExtractingService.extractTextFromPdf(note, content);
 }
 

src/services/text_extracting.js

@@ -1,7 +1,8 @@
 const Canvas = require("canvas");
 const OCRAD = require("ocrad.js");
-const log = require("./log.js");
+const log = require("./log");
-const optionService = require("./options.js");
+const optionService = require("./options");
+const cls = require("./cls");
 
 function ocrFromByteArray(img) {
     // byte array contains raw uncompressed pixel data
@@ -85,7 +86,7 @@ async function extractTextFromPdf(note, buffer) {
             content.items.forEach(({str}) => strings.push(str));
 
             try {
-                if (optionService.getOptionBool('ocrImages')) {
+                if (optionService.getOptionBool('ocrImages') && !cls.isOcrDisabled()) {
                     await ocrTextFromPdfImages(pdfjsLib, page, strings);
                 }
             }
@@ -117,13 +118,37 @@ async function ocrTextFromBuffer(buffer) {
     const canvas = new Canvas.createCanvas(img.width, img.height);
     const ctx = canvas.getContext('2d');
     ctx.drawImage(img, 0, 0, img.width, img.height);
+
     const plainText = OCRAD(canvas);
 
     log.info(`OCR of ${buffer.byteLength} image bytes into ${plainText.length} chars of text took ${Date.now() - start}ms`);
     return plainText;
 }
 
+async function runOcr(note, buffer) {
+    console.log("buffer length", buffer.length);
+
+    if (!note.isImage()
+        || !optionService.getOptionBool('ocrImages')
+        || cls.isOcrDisabled()
+        || buffer.length === 0
+    ) {
+        return;
+    }
+
+    try {
+        const plainText = await ocrTextFromBuffer(buffer);
+
+        console.log("OCR", plainText);
+
+        note.saveNoteAncillary('plainText', 'text/plain', plainText);
+    }
+    catch (e) {
+        log.error(`OCR on note '${note.noteId}' failed with error '${e.message}', stack ${e.stack}`);
+    }
+}
+
 module.exports = {
-    ocrTextFromBuffer,
+    runOcr,
     extractTextFromPdf
 };

src/services/user_guide_import.js

@@ -13,6 +13,7 @@ const yauzl = require("yauzl");
 const htmlSanitizer = require('./html_sanitizer');
 const sql = require('./sql');
 const options = require('./options');
+const cls = require('./cls');
 const {USER_GUIDE_ZIP_DIR} = require('./resource_dir');
 
 async function importUserGuideIfNeeded() {
@@ -33,6 +34,8 @@ async function importUserGuideIfNeeded() {
     const hiddenRoot = becca.getNote("_hidden");
     const data = await fs.readFile(USER_GUIDE_ZIP_DIR + "/user-guide.zip", "binary");
 
+    cls.disableOcr(); // no OCR needed for user guide images
+
     await importZip(Buffer.from(data, 'binary'), hiddenRoot);
 
     options.setOption('userGuideSha256Hash', userGuideSha256HashInFile);