This is a port of MuPDF to javascript and webassembly, giving you the following:
- π₯ Blazing fast rendering of PDFs to PNG, SVG and even HTML
- πΌ Run in the web browser or your server. Basically any platform that supports Webassembly!
- β Supports Typescript
- πΊοΈ A super simple API that's also completely flexible, see below...
yarn add mupdf-js
# or
npm i mupdf-js
Before you do any processing, you'll need to initialise the MuPdf library:
import { createMuPdf } from "mupdf-js";
async function handleSomePdf(file: File) {
const mupdf = await createMuPdf();
//...
}
In the browser, you'll most likely retrieve a File or Blob object from an html <input type="file">
tag, supplied by a user.
You'll need to convert the file firstly to an ArrayBuffer
, then to a Uint8Array
:
import { createMuPdf } from "mupdf-js";
async function handleSomePdf(file) {
const mupdf = await createMuPdf();
const buf = await file.arrayBuffer();
const arrayBuf = new Uint8Array(buf);
//...
}
Once you have this, you can load the file into the MuPdf environment, creating a MuPdf document:
import { createMuPdf } from "mupdf-js";
async function handleSomePdf(file) {
const mupdf = await createMuPdf();
const buf = await file.arrayBuffer();
const arrayBuf = new Uint8Array(buf);
const doc = mupdf.load(arrayBuf);
}
You now have three different options to render the PDF document:
import { createMuPdf } from "mupdf-js";
async function handleSomePdf(file) {
const mupdf = await createMuPdf();
const buf = await file.arrayBuffer();
const arrayBuf = new Uint8Array(buf);
const doc = mupdf.load(arrayBuf);
// Each of these returns a string:
const png = mupdf.drawPageAsPNG(doc, 1, 300);
const svg = mupdf.drawPageAsSVG(doc, 1);
const html = mupdf.drawPageAsHTML(doc, 1);
// This method returns Uint8Array
const pngRaw = mupdf.drawPageAsPNGRaw(doc, 1, 300);
}
// Returns PNG as data uri string
mupdf.drawPageAsPNG(document, page, resolution);
// Returns PNG data as Uint8Array
mupdf.drawPageAsPNGRaw(document, page, resolution);
Arguments:
- document: a MuPdf document object
- page: the page number to be rendered, starting from 1
- resolution: the DPI to use for rendering the file
Returns: an uncompressed PNG image, encoded as a base64 data URI.
mupdf.drawPageAsSVG(document, page);
Arguments:
- document: a MuPdf document object
- page: the page number to be rendered, starting from 1
Returns: an SVG file with the PDF document rendered as image tiles.
mupdf.drawPageAsHTML(document, page);
Arguments:
- document: a MuPdf document object
- page: the page number to be rendered, starting from 1
Returns: an HTML file that uses absolute positioned elements for layout.
mupdf.getPageText(document, page);
Arguments:
- document: a MuPdf document object
- page: the page number to be rendered, starting from 1
Returns: string containing all text collected from page
mupdf.searchPageText(document, page, searchString, maxHits);
Arguments:
- document: a MuPdf document object
- page: the page number to be rendered, starting from 1
- searchString: string to search
- maxHits: the maximum possible number of matches (it stops search when reaches this limit)
Returns: array of found rectangles of text matches ({x: number, y: number, w: number, h: number}[])
You should set maxHits
to an appropriate level that a user would expect (for example 100), or allow users to set their own limit. Alternatively, if you want to allow effectively unlimited search hits (and risk running out of memory), you can set it to C's maximum unsigned 32-bit integer size, which is 4294967295.
By default, mupdf-js creates a MuPDF context upon initialization and uses it for all calls. However, since the context includes a cache, over time this can lead to an increase in the application's memory consumption. To manage the context independently, mupdf-js supports the following:
import { createMuPdfWithoutContext } from "mupdf-js";
async function handleSomePdf(file) {
const mupdf = await createMuPdfWithoutContext();
const ctx = mupdf.createContext();
const buf = await file.arrayBuffer();
const arrayBuf = new Uint8Array(buf);
const doc = mupdf.load(ctx, arrayBuf);
// Each of these returns a string:
const png = mupdf.drawPageAsPNG(ctx, doc, 1, 300);
const svg = mupdf.drawPageAsSVG(ctx, doc, 1);
const html = mupdf.drawPageAsHTML(ctx, doc, 1);
// This method returns Uint8Array
const pngRaw = mupdf.drawPageAsPNGRaw(ctx, doc, 1, 300);
mupdf.freeDocument(doc);
mupdf.freeContext(ctx);
}
By default, console.log and console.warn are used for printing errors and other messages. If you prefer to use your custom logger (e.g., pino), you can do the following:
import {createMuPdf} from "mupdf-js";
import pino from 'pino';
const logger = pino();
async function handleSomePdf(file) {
const mupdf = await createMuPdf();
mupdf.setLogger({
log: (...args: any[]) => {
logger.debug(...args);
},
errorLog: (...args: any[]) => {
logger.error(...args);
},
});
//...
}
See CONTRIBUTING.md
AGPL, subject to the MuPDF license.