ocr

src/becca/entities/bnote_attachment.js

@@ -90,6 +90,7 @@ class BNoteAttachment extends AbstractBeccaEntity {
     }
 
     setContent(content) {
+        sql.transactional(() => {
             this.contentCheckSum = this.calculateCheckSum(content);
             this.save(); // also explicitly save note_attachment to update contentCheckSum
 
@@ -102,8 +103,7 @@ class BNoteAttachment extends AbstractBeccaEntity {
             if (this.isProtected) {
                 if (protectedSessionService.isProtectedSessionAvailable()) {
                     pojo.content = protectedSessionService.encrypt(pojo.content);
-            }
+                } else {
-            else {
                     throw new Error(`Cannot update content of noteAttachmentId=${this.noteAttachmentId} since we're out of protected session.`);
                 }
             }
@@ -118,6 +118,7 @@ class BNoteAttachment extends AbstractBeccaEntity {
                 utcDateChanged: pojo.utcDateModified,
                 isSynced: true
             });
+        });
     }
 
     calculateCheckSum(content) {

src/services/image.js

@@ -65,6 +65,24 @@ function getImageMimeFromExtension(ext) {
     return `image/${ext === 'svg' ? 'svg+xml' : ext}`;
 }
 
+function runOcr(note, buffer) {
+    if (!optionService.getOptionBool('ocrImages')) {
+        return;
+    }
+
+    const start = Date.now();
+    const img = new Canvas.Image();
+    img.src = 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`);
+
+    note.saveNoteAttachment('plainText', 'text/plain', plainText);
+}
+
 function updateImage(noteId, uploadBuffer, originalName) {
     log.info(`Updating image ${noteId}: ${originalName}`);
 
@@ -85,17 +103,7 @@ function updateImage(noteId, uploadBuffer, originalName) {
             note.setContent(buffer);
         });
 
-        const start = Date.now();
+        runOcr(note, buffer);
-        const img = new Canvas.Image();
-        img.src = 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 text = OCRAD(canvas);
-
-        console.log(text);
-
-        log.info(`OCR of ${buffer.byteLength} bytes took ${Date.now() - start}ms`);
     });
 }
 
@@ -136,7 +144,9 @@ function saveImage(parentNoteId, uploadBuffer, originalName, shrinkImageSwitch,
             note.save();
 
             note.setContent(buffer);
-        })
+        });
+
+        runOcr(note, buffer);
     });
 
     return {

src/services/options_init.js

@@ -90,6 +90,7 @@ const defaultOptions = [
     { name: 'checkForUpdates', value: 'true', isSynced: true },
     { name: 'disableTray', value: 'false', isSynced: false },
     { name: 'userGuideSha256Hash', value: '', isSynced: true },
+    { name: 'ocrImages', value: 'true', isSynced: true },
 ];
 
 function initStartupOptions() {

src/services/search/expressions/note_content_fulltext.js

@@ -40,25 +40,40 @@ class NoteContentFulltextExp extends Expression {
         const resultNoteSet = new NoteSet();
         const sql = require('../../sql');
 
-        for (let {noteId, type, mime, content, isProtected} of sql.iterateRows(`
+        for (const row of sql.iterateRows(`
                 SELECT noteId, type, mime, content, isProtected
                 FROM notes JOIN note_contents USING (noteId) 
                 WHERE type IN ('text', 'code', 'mermaid') AND isDeleted = 0`)) {
 
+            this.findInText(row, inputNoteSet, resultNoteSet);
+        }
+
+        for (const row of sql.iterateRows(`
+                SELECT noteId, 'plainText' as type, mime, content, isProtected
+                FROM note_attachments JOIN note_attachment_contents USING (noteAttachmentId) 
+                WHERE name IN ('plainText') AND isDeleted = 0`)) {
+
+            this.findInText(row, inputNoteSet, resultNoteSet);
+        }
+
+        return resultNoteSet;
+    }
+
+    findInText({noteId, isProtected, content, type, mime}, inputNoteSet, resultNoteSet) {
         if (!inputNoteSet.hasNoteId(noteId) || !(noteId in becca.notes)) {
-                continue;
+            return;
         }
 
         if (isProtected) {
             if (!protectedSessionService.isProtectedSessionAvailable()) {
-                    continue;
+                return;
             }
 
             try {
                 content = protectedSessionService.decryptString(content);
             } catch (e) {
                 log.info(`Cannot decrypt content of note ${noteId}`);
-                    continue;
+                return;
             }
         }
 
@@ -76,8 +91,7 @@ class NoteContentFulltextExp extends Expression {
 
                 resultNoteSet.add(becca.notes[noteId]);
             }
-            }
+        } else {
-            else {
             const nonMatchingToken = this.tokens.find(token =>
                 !content.includes(token) &&
                 (
@@ -92,9 +106,7 @@ class NoteContentFulltextExp extends Expression {
                 resultNoteSet.add(becca.notes[noteId]);
             }
         }
-        }
+        return content;
-
-        return resultNoteSet;
     }
 
     preprocessContent(content, type, mime) {