This commit is contained in:
zadam 2023-01-26 09:42:11 +01:00
parent c68a67d148
commit 63c62df787
4 changed files with 104 additions and 80 deletions

View File

@ -90,33 +90,34 @@ class BNoteAttachment extends AbstractBeccaEntity {
}
setContent(content) {
this.contentCheckSum = this.calculateCheckSum(content);
this.save(); // also explicitly save note_attachment to update contentCheckSum
sql.transactional(() => {
this.contentCheckSum = this.calculateCheckSum(content);
this.save(); // also explicitly save note_attachment to update contentCheckSum
const pojo = {
noteAttachmentId: this.noteAttachmentId,
content: content,
utcDateModified: dateUtils.utcNowDateTime()
};
const pojo = {
noteAttachmentId: this.noteAttachmentId,
content: content,
utcDateModified: dateUtils.utcNowDateTime()
};
if (this.isProtected) {
if (protectedSessionService.isProtectedSessionAvailable()) {
pojo.content = protectedSessionService.encrypt(pojo.content);
if (this.isProtected) {
if (protectedSessionService.isProtectedSessionAvailable()) {
pojo.content = protectedSessionService.encrypt(pojo.content);
} else {
throw new Error(`Cannot update content of noteAttachmentId=${this.noteAttachmentId} since we're out of protected session.`);
}
}
else {
throw new Error(`Cannot update content of noteAttachmentId=${this.noteAttachmentId} since we're out of protected session.`);
}
}
sql.upsert("note_attachment_contents", "noteAttachmentId", pojo);
sql.upsert("note_attachment_contents", "noteAttachmentId", pojo);
entityChangesService.addEntityChange({
entityName: 'note_attachment_contents',
entityId: this.noteAttachmentId,
hash: this.contentCheckSum,
isErased: false,
utcDateChanged: pojo.utcDateModified,
isSynced: true
entityChangesService.addEntityChange({
entityName: 'note_attachment_contents',
entityId: this.noteAttachmentId,
hash: this.contentCheckSum,
isErased: false,
utcDateChanged: pojo.utcDateModified,
isSynced: true
});
});
}

View File

@ -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();
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`);
runOcr(note, buffer);
});
}
@ -136,7 +144,9 @@ function saveImage(parentNoteId, uploadBuffer, originalName, shrinkImageSwitch,
note.save();
note.setContent(buffer);
})
});
runOcr(note, buffer);
});
return {

View File

@ -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() {

View File

@ -40,63 +40,75 @@ 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`)) {
if (!inputNoteSet.hasNoteId(noteId) || !(noteId in becca.notes)) {
continue;
}
this.findInText(row, inputNoteSet, resultNoteSet);
}
if (isProtected) {
if (!protectedSessionService.isProtectedSessionAvailable()) {
continue;
}
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`)) {
try {
content = protectedSessionService.decryptString(content);
} catch (e) {
log.info(`Cannot decrypt content of note ${noteId}`);
continue;
}
}
content = this.preprocessContent(content, type, mime);
if (this.tokens.length === 1) {
const [token] = this.tokens;
if ((this.operator === '=' && token === content)
|| (this.operator === '!=' && token !== content)
|| (this.operator === '*=' && content.endsWith(token))
|| (this.operator === '=*' && content.startsWith(token))
|| (this.operator === '*=*' && content.includes(token))
|| (this.operator === '%=' && getRegex(token).test(content))) {
resultNoteSet.add(becca.notes[noteId]);
}
}
else {
const nonMatchingToken = this.tokens.find(token =>
!content.includes(token) &&
(
// in case of default fulltext search we should consider both title, attrs and content
// so e.g. "hello world" should match when "hello" is in title and "world" in content
!this.flatText
|| !becca.notes[noteId].getFlatText().includes(token)
)
);
if (!nonMatchingToken) {
resultNoteSet.add(becca.notes[noteId]);
}
}
this.findInText(row, inputNoteSet, resultNoteSet);
}
return resultNoteSet;
}
findInText({noteId, isProtected, content, type, mime}, inputNoteSet, resultNoteSet) {
if (!inputNoteSet.hasNoteId(noteId) || !(noteId in becca.notes)) {
return;
}
if (isProtected) {
if (!protectedSessionService.isProtectedSessionAvailable()) {
return;
}
try {
content = protectedSessionService.decryptString(content);
} catch (e) {
log.info(`Cannot decrypt content of note ${noteId}`);
return;
}
}
content = this.preprocessContent(content, type, mime);
if (this.tokens.length === 1) {
const [token] = this.tokens;
if ((this.operator === '=' && token === content)
|| (this.operator === '!=' && token !== content)
|| (this.operator === '*=' && content.endsWith(token))
|| (this.operator === '=*' && content.startsWith(token))
|| (this.operator === '*=*' && content.includes(token))
|| (this.operator === '%=' && getRegex(token).test(content))) {
resultNoteSet.add(becca.notes[noteId]);
}
} else {
const nonMatchingToken = this.tokens.find(token =>
!content.includes(token) &&
(
// in case of default fulltext search we should consider both title, attrs and content
// so e.g. "hello world" should match when "hello" is in title and "world" in content
!this.flatText
|| !becca.notes[noteId].getFlatText().includes(token)
)
);
if (!nonMatchingToken) {
resultNoteSet.add(becca.notes[noteId]);
}
}
return content;
}
preprocessContent(content, type, mime) {
content = utils.normalize(content.toString());