mirrors/CyberChef
1
0
Fork 0
mirror of https://github.com/gchq/CyberChef.git synced 2025-04-22 15:56:16 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge branch 'master' into rsa

Browse source
This commit is contained in:
Matt 2020-06-08 15:55:37 +01:00 • committed by GitHub
commit d4ae241758
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
9 changed files with 2056 additions and 2407 deletions
Download patch file Download diff file Expand all files Collapse all files

26
src/core/operations/JavaScriptMinify.mjs
View file

@ -4,10 +4,9 @@
* @license Apache-2.0
*/
import OperationError from "../errors/OperationError.mjs";
import Operation from "../Operation.mjs";
import * as esprima from "esprima";
import escodegen from "escodegen";
import esmangle from "esmangle";
import Terser from "terser";
/**
* JavaScript Minify operation
@ -34,22 +33,11 @@ class JavaScriptMinify extends Operation {
* @returns {string}
*/
run(input, args) {
let result = "";
const AST = esprima.parseScript(input),
optimisedAST = esmangle.optimize(AST, null),
mangledAST = esmangle.mangle(optimisedAST);
result = escodegen.generate(mangledAST, {
format: {
renumber: true,
hexadecimal: true,
escapeless: true,
compact: true,
semicolons: false,
parentheses: false
}
});
return result;
const result = Terser.minify(input);
if (result.error) {
throw new OperationError(`Error minifying JavaScript. (${result.error})`);
}
return result.code;
}
}

2
src/core/operations/NormaliseUnicode.mjs
View file

@ -51,7 +51,7 @@ class NormaliseUnicode extends Operation {
case "NFKD":
return unorm.nfkd(input);
case "NFKC":
return unorm.nfc(input);
return unorm.nfkc(input);
default:
throw new OperationError("Unknown Normalisation Form");
}

26
src/core/operations/OpticalCharacterRecognition.mjs
View file

@ -1,6 +1,7 @@
/**
* @author n1474335 [n1474335@gmail.com]
* @author mshwed [m@ttshwed.com]
* @author Matt C [me@mitt.dev]
* @copyright Crown Copyright 2019
* @license Apache-2.0
*/
@ -12,7 +13,7 @@ import { toBase64 } from "../lib/Base64.mjs";
import { isWorkerEnvironment } from "../Utils.mjs";
import Tesseract from "tesseract.js";
const { TesseractWorker } = Tesseract;
const { createWorker } = Tesseract;
import process from "process";
@ -60,23 +61,30 @@ class OpticalCharacterRecognition extends Operation {
const assetDir = isWorkerEnvironment() ? `${self.docURL}/assets/` : `${process.cwd()}/src/core/vendor/`;
try {
self.sendStatusMessage("Spinning up Tesseract worker...");
const image = `data:${type};base64,${toBase64(input)}`;
const worker = new TesseractWorker({
const worker = createWorker({
workerPath: `${assetDir}tesseract/worker.min.js`,
langPath: `${assetDir}tesseract/lang-data`,
corePath: `${assetDir}tesseract/tesseract-core.wasm.js`,
});
const result = await worker.recognize(image)
.progress(progress => {
logger: progress => {
if (isWorkerEnvironment()) {
self.sendStatusMessage(`Status: ${progress.status} - ${(parseFloat(progress.progress)*100).toFixed(2)}%`);
self.sendStatusMessage(`Status: ${progress.status}${progress.status === "recognizing text" ? ` - ${(parseFloat(progress.progress)*100).toFixed(2)}%`: "" }`);
}
});
}
});
await worker.load();
self.sendStatusMessage("Loading English language...");
await worker.loadLanguage("eng");
self.sendStatusMessage("Intialising Tesseract API...");
await worker.initialize("eng");
self.sendStatusMessage("Finding text...");
const result = await worker.recognize(image);
if (showConfidence) {
return `Confidence: ${result.confidence}%\n\n${result.text}`;
return `Confidence: ${result.data.confidence}%\n\n${result.data.text}`;
} else {
return result.text;
return result.data.text;
}
} catch (err) {
throw new OperationError(`Error performing OCR on image. (${err})`);