A command line tool to generate PDF from URL, HTML or Markdown files with electron.
I have a blog post explain why PDF Generation On The Web
Production ready? See it in action for the Myanmar Election!
Starting with version 4.0.x the master branch will always have the latest electron version.
Semantic Versioning is used, and corresponds to electron versions in the following way:
- electron-pdf 10.0.x (master) => electron=10.1.3, node=12.16.3, chrome=85.0.4183.121
- electron-pdf 7.0.x => electron 7.x (Chromium 78, Node 12.8.1)
- electron-pdf 4.0.x => electron 4.x (Chromium 69, Node 10.11.0)
- electron-pdf 1.3.x => electron 1.6.x (Chromium 56, Node 7.4)
- electron-pdf 1.2.x => electron 1.4.x (Chromium 53, Node 6.5)
Note: The Chromium versions employed by electron have impacts based on the functionality you may be exporting.
Choose the version you need based on Chromium.
npm install electron-pdf
Note: If you're installing electron-pdf as root using the system level npm (vs a user-level install like with NVM) then you may need to run the following command instead:
sudo npm install electron-pdf -g --unsafe-perm
Please see the npm docs for more information.
For gnu/linux installations without a graphical environment:
$ sudo apt-get install xvfb # or equivalent
$ export DISPLAY=':99.0'
$ Xvfb :99 -screen 0 1024x768x24 > /dev/null 2>&1 &
$ electron-pdf ...
There is also an example docker machine here.
Electron PDF can be used inside of an application, or more commonly as the engine for a pdf rendering service. For instance, to handle http requests using Express. The following snippets show you how you can get started.
In package.json
"start": "DEBUG=electronpdf:* electron index.js",
"watch": "DEBUG=electronpdf:* nodemon --exec electron index.js"
var ElectronPDF = require('electron-pdf')
var express = require('express')
var bodyParser = require('body-parser')
var app = express()
app.use(bodyParser.json())
var exporter = new ElectronPDF()
exporter.on('charged', () => {
//Only start the express server once the exporter is ready
app.listen(port, hostname, function() {
console.log(`Export Server running at http://${hostname}:${port}`);
})
})
exporter.start()
app.post('/pdfexport', function(req,res){
// derive job arguments from request here
//
const jobOptions = {
/**
r.results[] will contain the following based on inMemory
false: the fully qualified path to a PDF file on disk
true: The Buffer Object as returned by Electron
Note: the default is false, this can not be set using the CLI
*/
inMemory: false
}
const options = {
pageSize : "A4"
}
exporter.createJob(source, target, options, jobOptions).then( job => {
job.on('job-complete', (r) => {
console.log('pdf files:', r.results)
// Process the PDF file(s) here
})
job.render()
})
})
If you set the inMemory
setting to true, you must also set closeWindow=false
or you will get a segmentation fault anytime the window is closed before the buffer
is sent on the response. You then need to invoke job.destroy
to close the window.
Sample Code:
const jobOptions = { inMemory: true, closeWindow: false }
exporter.createJob(source, target, options, jobOptions).then( job => {
job.on('job-complete', (r) => {
//Send the Buffer here
process.nextTick(() => {job.destroy()})
})
})
The API is designed to emit noteworthy events rather than use callbacks. Full documentation of all events is a work in progress.
ELECTRONPDF_RENDERER_MAX_MEMORY
: The --max-old-space-size option for each Electron renderer process (browser window); default:75% of total system memory up to 8GB
ELECTRONPDF_WINDOW_CLEANUP_INTERVAL
: Interval for which to check for hung windows, in milliseconds; default:30 seconds
ELECTRONPDF_WINDOW_LIFE_THRESHOLD
: How long a window can remain open before it is terminated, in milliseconds; default:5 minutes
ELECTRONPDF_PNG_CAPTURE_DELAY
: Amount of millis to wait before invoking WebContents.capturePage for PNG exports; default:100ms
For Ad-hoc conversions, Electron PDF comes with support for a CLI.
$ electron-pdf index.html ~/Desktop/index.pdf
$ electron-pdf index.md ~/Desktop/index.pdf
$ electron-pdf index.html ~/Desktop/index.pdf -c my-awesome-css.css
$ electron-pdf https://fraserxu.me ~/Desktop/fraserxu.pdf
Electron PDF gives you complete control of how the BrowserWindow should be configured, and when the window contents should be captured.
The BrowserWindow supports many options which you
may define by passing a JSON Object to the --browserConfig
option.
Some common use cases may include:
height
andwidth
- electron-pdf calculates the browser height and width based off of the dimensions of PDF page size multiplied by the HTML standard of 96 pixels/inch. So only set these values if you need to override this behaviorshow
- to display the browser window during generation
$ electron-pdf https://fraserxu.me ~/Desktop/fraserxu.pdf --browserConfig '{"show":true}'
electron-pdf ./index.html ~/Desktop/README.pdf -e
In your application, at the point which the view is ready for rendering
document.body.dispatchEvent(new Event('view-ready'))
Warning: It is possible that your application will be ready and emit the event before the main electron process has had a chance execute the javascript in the renderer process which listens for this event.
If you are finding that the event is not effective and your page waits until the full timeout has occurred, then you should use setInterval
to emit the event until it is acknowledged like so:
var eventEmitInterval = setInterval(function () {
document.body.dispatchEvent(new Event('view-ready'))
}, 25)
document.body.addEventListener('view-ready-acknowledged', function(){
clearInterval(eventEmitInterval)
})
When the main process first receives your ready event it will emit a single acknowlegement on document.body
with whatever event name you are using suffixed with -acknowledged
. So the default would be view-ready-acknowledged
If the page you are rending is under your control, and you wish to modify the behavior of the rendering process you can use a CustomEvent and an observer that will be triggered after the view is ready but before it is captured.
document.body.dispatchEvent(new CustomEvent('view-ready', { detail: {layout: landscape} }))
You are required to provide a function that accepts the detail
object from
the CustomEvent and returns a Promise. You may optionally fulfill the promise with
and object that will amend/override any of the contextual attributes assigned to resource (url)
currently being exported.
As an example, suppose you wanted to change the orientation of the PDF, and capture the output as PNG instead of a PDF.
job.observeReadyEvent( (detail) => {
return new Promise( (resolve,reject) => {
const context = {}
if( detail && detail.landscape ){
job.changeArgValue('landscape', true)
context.type = 'png'
}
resolve(context)
})
})
Note: Future versions of the library will only allow you to provide context overrides, and not allow you to change job level attributes.
Electron PDF exposes the printToPDF settings (i.e. pageSize, orientation, margins, etc.) available from the Electron API. See the following options for usage.
A command line tool to generate PDF from URL, HTML or Markdown files
Options
--help Show this help
--version Current version of package
-i | --input String - The path to the HTML file or url
-o | --output String - The path of the output PDF
-b | --printBackground Boolean - Whether to print CSS backgrounds.
--acceptLanguage String - A valid value for the 'Accept-Language' http request header
--browserConfig String - A valid JSON String that will be parsed into the options passed to electron.BrowserWindow
-c | --css String - The path to custom CSS (can be specified more than once)
-d | --disableCache Boolean - Disable HTTP caching
false - default
-e | --waitForJSEvent String - The name of the event to wait before PDF creation
'view-ready' - default
-l | --landscape Boolean - true for landscape, false for portrait (don't pass a string on the CLI, just the `-l` flag)
false - default
-m | --marginsType Integer - Specify the type of margins to use
0 - default margins
1 - no margins (electron-pdf default setting)
2 - minimum margins
--noprint Boolean - Do not run printToPDF, useful if the page downloads a file that needs captured instead of a PDF.
The Electron `win.webContents.session.on('will-download')` event will be implemented
and the file saved to the location provided in `--output`.
Currently only supports a single import url.
The page is responsible for initiating the download itself.
-p | --pageSize String - Can be A3, A4, A5, Legal, Letter, Tabloid or an Object containing height and width in microns
"A4" - default
-r | --requestHeaders String - A valid JSON String that will be parsed into an Object where each key/value pair is: <headerName>: <headerValue>
Example: '{"Authorization": "Bearer token", "X-Custom-Header": "Hello World"}'
-s | --printSelectionOnly Boolean - Whether to print selection only
false - default
-t | --trustRemoteContent Boolean - Whether to trust remote content loaded in the Electron webview. False by default.
--type String - The type of export, will dictate the output file type. 'png': PNG image, anything else: PDF File
-w | --outputWait Integer β Time to wait (in MS) between page load and PDF creation.
If used in conjunction with -e this will override the default timeout of 10 seconds
--ignoreCertificateErrors Boolean - If true, all certificate errors thrown by Electron will be ignored. This can be used to accept self-signed and untrusted certificates. You should be aware of the security implications of setting this flag.
false - default
Find more information on Electron Security here.
If you have a Sentry account and setup a new app to get a new DSN, you can set a SENTRY_DSN
environment variable which will activate sentry logs.
See lib/sentry.js
for implementation details.
This will allow you to easily see/monitor errors that are occuring inside of the Chromium renderer (browser window). It also automatically integrates with Electron's Crash Reporter
You can see some additional logging (if you're getting errors or unexpected output) by setting DEBUG=electron*
For example: DEBUG=electron* electron-pdf <input> <output> -l
Usage
$ electron-pdf <input> <output>
$ electron-pdf <input> <output> -l
Examples
$ electron-pdf http://fraserxu.me ~/Desktop/fraserxu.pdf
$ electron-pdf ./index.html ~/Desktop/index.pdf
$ electron-pdf ./README.md ~/Desktop/README.pdf -l
$ electron-pdf ./README.md ~/Desktop/README.pdf -l -c my-awesome-css.css
Inspired by electron-mocha
Want to use the same options, but export to PNG or snapshot the rendered HTML? Just set the output filename to end in .png or .html instead!
Examples
$ electron-pdf http://fraserxu.me ~/Desktop/fraserxu.pdf
$ electron-pdf http://fraserxu.me ~/Desktop/fraserxu.html
$ electron-pdf http://fraserxu.me ~/Desktop/fraserxu.png
If you need powerpoint support, pdf-powerpoint picks up where Electron PDF leaves off by converting each page in the PDF to a PNG and placing them on individual slides.
MIT