mirrors/CyberChef
1
0
Fork 0
mirror of https://github.com/gchq/CyberChef.git synced 2025-04-21 23:36:16 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions

Fixed Optical Character Recognition and added tests

Browse source
This commit is contained in:
n1474335 2024-06-21 14:12:28 +01:00
parent c23a8de5a0
commit ab37c1e562
No known key found for this signature in database
GPG key ID: D15457B7B4AF3F37
4 changed files with 34 additions and 24 deletions
Download patch file Download diff file Expand all files Collapse all files

21
src/core/operations/OpticalCharacterRecognition.mjs
View file

@ -12,9 +12,10 @@ import { isImage } from "../lib/FileType.mjs";
import { toBase64 } from "../lib/Base64.mjs";
import { isWorkerEnvironment } from "../Utils.mjs";
import process from "process";
import { createWorker } from "tesseract.js";
const OEM_MODES = ["Tesseract only", "LSTM only", "Tesseract/LSTM Combined"];
/**
* Optical Character Recognition operation
*/
@ -37,6 +38,12 @@ class OpticalCharacterRecognition extends Operation {
name: "Show confidence",
type: "boolean",
value: true
},
{
name: "OCR Engine Mode",
type: "option",
value: OEM_MODES,
defaultIndex: 1
}
];
}
@ -47,7 +54,7 @@ class OpticalCharacterRecognition extends Operation {
* @returns {string}
*/
async run(input, args) {
const [showConfidence] = args;
const [showConfidence, oemChoice] = args;
if (!isWorkerEnvironment()) throw new OperationError("This operation only works in a browser");
@ -56,12 +63,13 @@ class OpticalCharacterRecognition extends Operation {
throw new OperationError("Unsupported file type (supported: jpg,png,pbm,bmp) or no file provided");
}
const assetDir = isWorkerEnvironment() ? `${self.docURL}/assets/` : `${process.cwd()}/src/core/vendor/`;
const assetDir = `${self.docURL}/assets/`;
const oem = OEM_MODES.indexOf(oemChoice);
try {
self.sendStatusMessage("Spinning up Tesseract worker...");
const image = `data:${type};base64,${toBase64(input)}`;
const worker = createWorker({
const worker = await createWorker("eng", oem, {
workerPath: `${assetDir}tesseract/worker.min.js`,
langPath: `${assetDir}tesseract/lang-data`,
corePath: `${assetDir}tesseract/tesseract-core.wasm.js`,
@ -71,11 +79,6 @@ class OpticalCharacterRecognition extends Operation {
}
}
});
await worker.load();
self.sendStatusMessage(`Loading English language pack...`);
await worker.loadLanguage("eng");
self.sendStatusMessage("Intialising Tesseract API...");
await worker.initialize("eng");
self.sendStatusMessage("Finding text...");
const result = await worker.recognize(image);