A WebDAV client written in JavaScript for NodeJS and the browser.
About
WebDAV is a well-known, stable and highly flexible protocol for interacting with remote filesystems via an API. Being that it is so widespread, many file hosting services such as Box, ownCloud/Nextcloud and Yandex use it as a fallback to their other interfaces.
This library provides a WebDAV client interface that makes interacting with WebDAV enabled services easy. The API returns promises and resolve with the results. It parses and prepares directory-contents requests for easy consumption, as well as providing methods for fetching things like file stats and quotas.
Please read the contribution guide if you plan on making an issue or PR.
Node support
This library is compatibale with NodeJS version 10 and above (For version 6/8 support, use versions in the range of 2.*. For version 4 support, use versions in the range of 1.*). Version 2.x is now in maintenance mode and will receive no further feature additions. It will receive the odd bug fix when necessary. Version 1.x is no longer supported.
Usage in the Browser
As of version 3, WebDAV client is now supported in the browser. The compilation settings specify a minimum supported browser version of Internet Explorer 11, however testing in this browser is not performed regularly.
Although you may choose to transpile this library's default entry point (NodeJS) yourself, it is not advised - use the dedicated web version instead.
You can use the web version via a different entry point:
import { createClient } from "webdav/web";The browser version uses a UMD-style module definition, meaning you can simply load the library within your browser using a <script> tag. When using this method the library is made available on the window object as such: window.WebDAV. For example:
const client = window.WebDAV.createClient(/* ... */);NB: Streams are not available within the browser, so createReadStream and createWriteStream are just stubbed. Calling them will throw an exception.
Installation
Simple install as a dependency using npm:
npm install webdav --save
Usage
Usage entails creating a client adapter instance by calling the factory function createClient:
const { createClient } = require("webdav");
const client = createClient(
"https://webdav.example.com/marie123",
{
username: "marie",
password: "myS3curePa$$w0rd"
}
);
// Get directory contents
const directoryItems = await client.getDirectoryContents("/");
// Outputs a structure like:
// [{
// filename: "/my-file.txt",
// basename: "my-file.txt",
// lastmod: "Mon, 10 Oct 2018 23:24:11 GMT",
// size: 371,
// type: "file"
// }]Make sure to read the API documentation for more information on the available adapter methods.
Authentication & Connection
webdav uses Basic authentication by default, if username and password are provided (if none are provided, no Authorization header is specified). It also supports OAuth tokens and Digest auth.
Basic or no authentication
You can use the client without authentication if the server doesn't require it - simply avoid passing any values to username, password, token or digest in the config.
To use basic authentication, simply pass a username and password in the config.
webdav also allows for overriding the built in HTTP and HTTPS agents by setting the properties httpAgent & httpsAgent accordingly. These should be instances of node's http.Agent and https.Agent respectively.
OAuth tokens
To use a token to authenticate, simply pass the token data to the token field:
createClient(
"https://address.com",
{
token: {
"access_token": "2YotnFZFEjr1zCsicMWpAA",
"token_type": "example",
"expires_in": 3600,
"refresh_token": "tGzv3JOkF0XG5Qx2TlKWIA",
"example_parameter": "example_value"
}
}
);Digest authentication
If a server requires digest-based authentication, you can enable this functionality by setting digest to true:
createClient(
"https://address.com",
{
username: "someUser",
password: "myS3curePa$$w0rd",
digest: true
}
);Methods
copyFile
Copy a file from one remote location to another:
await client.copyFile("/sub/item.txt", "/destination/item.txt");createDirectory
Create a new directory:
await client.createDirectory("/completely/new/path");createReadStream
Create a read stream targeted at a remote file:
client
.createReadStream("/video.mp4")
.pipe(fs.createWriteStream("~/video.mp4"));If you want to stream only part of the file, you can specify the range in the options argument:
client
.createReadStream(
"/video.mp4",
{ range: { start: start, end: end } }
).pipe(fs.createWriteStream("~/video.mp4"));createWriteStream
Create a write stream targeted at a remote file:
fs.createReadStream("~/Music/song.mp3")
.pipe(client.createWriteStream("/music/song.mp3"));deleteFile
Delete a remote file:
await client.deleteFile("/tmp.dat");exists
Check if a file or directory exists:
if (await client.exists("/some/path") === false) {
await client.createDirectory("/some/path");
}getDirectoryContents
Get the contents of a remote directory. Returns an array of item stats.
// Get current directory contents:
const contents = await client.getDirectoryContents("/");
// Get all contents:
const contents = await client.getDirectoryContents("/", { deep: true });Files can be globbed using the glob option (processed using minimatch). When using a glob pattern it is recommended to fetch deep contents:
const images = await client.getDirectoryContents("/", { deep: true, glob: "/**/*.{png,jpg,gif}" });getFileContents
Fetch the contents of a remote file. Binary contents are returned by default (Buffer):
const buff = await client.getFileContents("/package.zip");It is recommended to use streams if the files being transferred are large.
Text files can also be fetched:
const str = await client.getFileContents("/config.json", { format: "text" });Specify the maxContentLength option to alter the maximum number of bytes the client can receive in the request. NodeJS only.
getFileDownloadLink
Return a public link where a file can be downloaded. This exposes authentication details in the URL.
const downloadLink = client.getFileDownloadLink("/image.png");Not all servers may support this feature. Only Basic authentication and unauthenticated connections support this method.
getFileUploadLink
Return a URL for a file upload:
const uploadLink = client.getFileUploadLink("/image.png");See getFileDownloadLink for support details.
getQuota
Get the quota information for the current account:
const quota = await client.getQuota();
// {
// "used": 1938743,
// "available": "unlimited"
// }moveFile
Move a remote file to another remote location:
await client.moveFile("/file1.png", "/file2.png");putFileContents
Write data to a remote file:
// Write a buffer:
await client.putFileContents("/my/file.jpg", imageBuffer, { overwrite: false });
// Write a text file:
await client.putFileContents("/my/file.txt", str);Specify the maxBodyLength option to alter the maximum number of bytes the client can send in the request. NodeJS only.
Handling Upload Progress (browsers only):
This uses the axios onUploadProgress callback which uses the native XMLHttpRequest progress event.
// Upload a file and log the progress to the console:
await client.putFileContents("/my/file.jpg", imageFile, { onUploadProgress: progress => {
console.log(`Uploaded ${progress.loaded} bytes of ${progress.total}`);
} });stat
Get a file or directory stat object:
const stat = await client.stat("/some/file.tar.gz");Returns an item stat.
Custom requests
Custom requests can be made to the attached host:
const contents = await client.customRequest("/alrighty.jpg", {
method: "PROPFIND",
headers: {
Accept: "text/plain",
Depth: 0
},
responseType: "text"
});Returned data structures
Directory contents items
Each item returned by getDirectoryContents is basically an item stat. If the details: true option is set, each item stat (as mentioned in the stat documentation) will also include the props property containing extra properties returned by the server. No particular property in props, not its format or value, is guaranteed.
You can request all files in the file-tree (infinite depth) by calling getDirectoryContents with the option deep: true. All items will be returned in a flat array, where the filename will hold the absolute path.
Detailed responses
Requests that return results, such as getDirectoryContents, getFileContents, getQuota and stat, can be configured to return more detailed information, such as response headers. Pass { details: true } to their options argument to receive an object like the following:
| Property | Type | Description |
|---|---|---|
| data | * | The data returned by the procedure. Will be whatever type is returned by calling without { details: true }
|
| headers | Object | The response headers. |
Item stat
Item stats are objects with properties that descibe a file or directory. They resemble the following:
{
"filename": "/test",
"basename": "test",
"lastmod": "Tue, 05 Apr 2016 14:39:18 GMT",
"size": 0,
"type": "directory",
"etag": null
}or:
{
"filename": "/image.jpg",
"basename": "image.jpg",
"lastmod": "Sun, 13 Mar 2016 04:23:32 GMT",
"size": 42497,
"type": "file",
"mime": "image/jpeg",
"etag": "33a728c7f288ede1fecc90ac6a10e062"
}Properties:
| Property name | Type | Present | Description |
|---|---|---|---|
| filename | String | Always | File path of the remote item |
| basename | String | Always | Base filename of the remote item, no path |
| lastmod | String | Always | Last modification date of the item |
| size | Number | Always | File size - 0 for directories |
| type | String | Always | Item type - "file" or "directory" |
| mime | String | Files only | Mime type - for file items only |
| etag | String / null | When supported | ETag of the file |
| props | Object | details: true |
Props object containing all item properties returned by the server |
Compatibility
This library has been tested to work with the following WebDAV servers or applications:
¹ These services will work if CORS is correctly configured to return the proper headers. This may not work by default.
CORS
CORS is a security enforcement technique employed by browsers to ensure requests are executed to and from expected contexts. It can conflict with this library if the target server doesn't return CORS headers when making requests from a browser. It is your responsibility to handle this.
It is a known issue that ownCloud and Nextcloud servers by default don't return friendly CORS headers, making working with this library within a browser context impossible. You can of course force the addition of CORS headers (Apache or Nginx configs) yourself, but do this at your own risk.