pdf-ocr-ts
v1.0.17
Published
Javascript-only library to perform OCR on scanned PDFs to turn them into searchable PDFs
Downloads
23
Readme
pdf-ocr-ts
pdf-ocr-ts creates searchable PDF files out of PDF files that only contain images of scanned documents. It is javascript-only and hence works without the need to install further tools. Under the hood it uses pdf.js to render the pages within a pdf to png files, Jimp to create compressed jpeg images, tesseract.js to perform ocr and pdf-lib to merge the single page pdfs tesseract.js is creating into a final searchable output PDF file.
Usage
To create a searchable PDF with filename outputFilename
from inputFilename
use:
const { default: PdfOcr } = require('pdf-ocr-ts');
const inputFilename = './input/scan_test.pdf';
const outputFilename = './output/scan_test-searchable.pdf';
PdfOcr.createSearchablePdf(inputFilename, outputFilename);
In certain contexts it might be more handy to read the input file in some other function and also output the searchable PDF in another component. In these cases pdf-ocr-ts offers the function getSearchablePdfBufferBased(Uint8Array)
that takes a Uint8Array
(e.g. created by fs.readFile()), performs ocr and returns the searchable PDF file as Uint8Array
. Which can than be used again in fs.writeFile().
const { default: PdfOcr } = require('pdf-ocr-ts');
const fs = require('fs');
const path = require('path');
const inputFilename = './input/scan_test.pdf';
const outputFilename = './output/scan_test-searchable.pdf';
(async () => {
const pdf = new Uint8Array(fs.readFileSync(path.resolve(__dirname, inputFilename)));
const { pdfBuffer, text } = await PdfOcr.getSearchablePdfBufferBased(pdf);
fs.writeFile(path.resolve(__dirname, outputFilename), pdfBuffer, (error) => {
if (error) {
console.error(`Error: ${error}`);
} else {
console.log(`Finished merging PDFs into ${outputFilename}.`);
}
});
})();
To generate log output, pdf-ocr-ts supports logging frameworks. It ships with the most simple logger simpleLog
and supports any logger with the call signature (level: string, message: string) => void;
(see ./utils/Logger.ts
).
const { default: PdfOcr } = require('pdf-ocr-ts');
const { simpleLog } = require("pdf-ocr-ts/build/utils/Logger");
const inputFilename = './input/scan_test.pdf';
const outputFilename = './output/scan_test-searchable.pdf';
PdfOcr.createSearchablePdf(inputFilename, outputFilename, simpleLog);
Here's an example for the log library winston.js
via a simple wrapper like logHelper(level: string, message: string)
. Internally pdf-ocr-ts uses the log levels: info
, error
and debug
.
const { default: PdfOcr } = require('pdf-ocr-ts');
const { createLogger, transports, format } = require("winston");
// create the winston logger
const logger = createLogger({
transports: [new transports.Console()],
format: format.combine(
format.colorize(),
format.timestamp(),
format.printf(({ timestamp, level, message }) => {
return `[${timestamp}] ${level}: ${message}`;
})
),
});
// wrap winston logger in logHelper to comply with the call signature
// (level: string, message: string) => void;
function logHelper(level: string, message: string) {
logger.log(level, message);
}
// pass the logHelper function
PdfOcr.createSearchablePdf(inputFilename, outputFilename, logHelper);
To build the module from source run npm run build
.