🌜
🌞
drzzle-streamsaver

drzzle-streamsaver

v2.0.5

StreamSaver writes stream to the filesystem directly - asynchronous

npm install drzzle-streamsaver

README

StreamSaver.js

npm version

First I want to thank Eli Grey for a fantastic work implementing the FileSaver.js to save files & blobs so easily! But there is one obstacle - The RAM it can hold and the max blob size limitation

StreamSaver.js takes a different approach. Instead of saving data in client-side storage or in memory you could now actually create a writable stream directly to the file system (I'm not talking about chromes sandboxed file system or any other web storage)

StreamSaver.js is the solution to saving streams on the client-side. It is perfect for webapps that need to save really large amounts of data created on the client-side, where the RAM is really limited, like on mobile devices.

Getting started

StreamSaver in it's simplest form

<script src="https://cdn.jsdelivr.net/npm/[email protected]/dist/ponyfill.min.js"></script>
<script src="https://cdn.jsdelivr.net/npm/[email protected]/StreamSaver.min.js"></script>
<script>
    import streamSaver from 'streamsaver'
    const streamSaver = require('streamsaver')
    const streamSaver = window.streamSaver
</script>
<script>
  const fileStream = streamSaver.createWriteStream('filename.txt', {
    size: 22, // (optional) Will show progress
    writableStrategy: undefined, // (optional)
    readableStrategy: undefined  // (optional)
  })

  new Response('StreamSaver is awesome').body
    .pipeTo(fileStream)
    .then(success, error)
</script>

Some browser have ReadableStream but not WritableStream. web-streams-polyfill can fix this gap. It's better to load the ponyfill instead of the polyfill and override the existing implementation because StreamSaver works better when a native ReadableStream is transferable to the service worker. hopefully MattiasBuelens will fix the missing implementations instead of overriding the existing. If you think you can help out here is the issue

Best practice

Use https if you can. That way you don't have to open the man in the middle in a popup to install the service worker from another secure context. Popups are often blocked but if you can't it's best that you initiate the createWriteStream on user interaction. Even if you don't have any data ready - this is so that you can get around the popup blockers. (In secure context this don't matter) Another benefit of using https is that the mitm-iframe can ping the service worker to prevent it from going idle. (worker goes idle after 30 sec in firefox, 5 minutes in blink) but also this won't mater if the browser supports transferable streams throught postMessage since service worker don't have to handle any logic. (the stream that you transfer to the service worker will be the stream we respond with)

Handle unload event when user leaves the page. The download gets broken when you leave the page. Because it looks like a regular native download process some might think that it's okey to leave the page beforehand since it's is downloading in the background directly from some a server, but it isn't.

// abort so it dose not look stuck
window.onunload = () => {
  writableStream.abort()
  // also possible to call abort on the writer you got from `getWriter()`
  writer.abort()
}

window.onbeforeunload = evt => {
  if (!done) {
    evt.returnValue = `Are you sure you want to leave?`;
  }
}

Note that when using insecure context StreamSaver will navigate to the download url instead of using an hidden iframe to initiate the download, this will trigger the onbefureunload event when the download starts, but it will not call the onunload event... In secure context you can add this handler immediately. Otherwise this has to be added sometime later.

Configuration

There a some few settings you can apply to StreamSaver to configure what it should use

// StreamSaver can detect and use the Ponyfill that is loaded from the cdn.
streamSaver.WritableStream = streamSaver.WritableStream
streamSaver.TransformStream = streamSaver.TransformStream
// if you decide to host mitm + sw yourself
streamSaver.mitm = 'https://example.com/custom_mitm.html'

Examples

There are a few examples in the examples directory

In the wild

How does it work?

There is no magical saveAs() function that saves a stream, file or blob. (at least not if/when native-filesystem api becomes avalible) The way we mostly save Blobs/Files today is with the help of Object URLs and a[download] attribute FileSaver.js takes advantage of this and create a convenient saveAs(blob, filename). fantastic! But you can't create a objectUrl from a stream and attach it to a link...

link = document.createElement('a')
link.href = URL.createObjectURL(stream) // DOES NOT WORK
link.download = 'filename'
link.click() // Save

So the one and only other solution is to do what the server does: Send a stream with Content-Disposition header to tell the browser to save the file. But we don't have a server or the content isn't on a server! So the solution is to create a service worker that can intercept request and use respondWith() and act as a server.
But a service workers are only allowed in secure contexts and it requires some effort to put up. Most of the time you are working in the main thread and the service worker are only alive for < 5 minutes before it goes idle.

  1. So StreamSaver creates a own man in the middle that installs the service worker in a secure context hosted on github static pages. either from a iframe (in secure context) or a new popup if your page is insecure.
  2. Transfer the stream (or DataChannel) over to the service worker using postMessage.
  3. And then the worker creates a download link that we then open.

if a "transferable" readable stream was not passed to the service worker then the mitm will also try to keep the service worker alive by pinging it every x second to prevent it from going idle.

To test this locally, spin up a local server
(we don't use any pre compiler or such)

# A simple php or python server is enough
php -S localhost:3001
python -m SimpleHTTPServer 3001
# then open localhost:3001/example.html

Release Notes

2.0.4
By Jimmy Wärting • Published on January 5, 2020

Forces IOS to build blob instead of using streams - dc0c9da, #148

Hopefully this also means chrome, FF and whatever on IOS that can't have a own render engine.

2.0.3
By Jimmy Wärting • Published on January 5, 2020

Don't close the messageChannel prematurely

Solved #103

General

License
MIT
Typescript Types
None found
Tree-shakeable
No

Popularity

GitHub Stargazers
3,147
Community Interest
3,029
Number of Forks
344

Maintenance

Commits
11/2110/2203
Last Commit
Open Issues
50
Closed Issues
165
Open Pull Requests
4
Closed Pull Requests
15

Versions

Versions Released
11/2110/2201
Latest Version Released
Jan 5, 2020
Current Tags
latest2.0.5

Dependencies

Dependencies (0)
Dev Dependencies (0)

Contributors

jimmywarting
jimmywarting
Commits: 179
TexKiller
TexKiller
Commits: 4
machawk1
machawk1
Commits: 3
egoroof
egoroof
Commits: 3
frank-dspeed
frank-dspeed
Commits: 2
vobruba-martin
vobruba-martin
Commits: 2
petethomas
petethomas
Commits: 1
taniadaniela
taniadaniela
Commits: 1
OrekiSH
OrekiSH
Commits: 1
KewlKris
KewlKris
Commits: 1
supermaai
supermaai
Commits: 1
Haraldson
Haraldson
Commits: 1
Bonjur
Bonjur
Commits: 1
Nomeji
Nomeji
Commits: 1
sirbarrence
sirbarrence
Commits: 1