degit makes copies of git repositories. When you run degit some-user/some-repo
, it will find the latest commit on https://github.com/some-user/some-repo and download the associated tar file to ~/.degit/some-user/some-repo/commithash.tar.gz
if it doesn't already exist locally. (This is much quicker than using git clone
, because you're not downloading the entire git history.)
Requires Node 8 or above, because async
and await
are the cat's pyjamas
npm install -g degit
The simplest use of degit is to download the master branch of a repo from GitHub to the current working directory:
degit user/repo
# these commands are equivalent
degit github:user/repo
degit [email protected]:user/repo
degit https://github.com/user/repo
Or you can download from GitLab and BitBucket:
# download from GitLab
degit gitlab:user/repo
degit [email protected]:user/repo
degit https://gitlab.com/user/repo
# download from BitBucket
degit bitbucket:user/repo
degit [email protected]:user/repo
degit https://bitbucket.org/user/repo
# download from Sourcehut
degit git.sr.ht/user/repo
degit [email protected]:user/repo
degit https://git.sr.ht/user/repo
The default branch is master
.
degit user/repo#dev # branch
degit user/repo#v1.2.3 # release tag
degit user/repo#1234abcd # commit hash
If the second argument is omitted, the repo will be cloned to the current directory.
degit user/repo my-new-project
To clone a specific subdirectory instead of the entire repo, just add it to the argument:
degit user/repo/subdirectory
If you have an https_proxy
environment variable, Degit will use it.
Private repos can be cloned by specifying --mode=git
(the default is tar
). In this mode, Degit will use git
under the hood. It's much slower than fetching a tarball, which is why it's not the default.
Note: this clones over SSH, not HTTPS.
degit --help
- Private repositories
Pull requests are very welcome!
A few salient differences:
- If you
git clone
, you get a.git
folder that pertains to the project template, rather than your project. You can easily forget to re-init the repository, and end up confusing yourself - Caching and offline support (if you already have a
.tar.gz
file for a specific commit, you don't need to fetch it again). - Less to type (
degit user/repo
instead ofgit clone --depth 1 [email protected]:user/repo
) - Composability via actions
- Future capabilities — interactive mode, friendly onboarding and postinstall scripts
You can also use degit inside a Node script:
const degit = require('degit');
const emitter = degit('user/repo', {
cache: true,
force: true,
verbose: true,
});
emitter.on('info', info => {
console.log(info.message);
});
emitter.clone('path/to/dest').then(() => {
console.log('done');
});
You can manipulate repositories after they have been cloned with actions, specified in a degit.json
file that lives at the top level of the working directory. Currently, there are two actions — clone
and remove
. Additional actions may be added in future.
// degit.json
[
{
"action": "clone",
"src": "user/another-repo"
}
]
This will clone user/another-repo
, preserving the contents of the existing working directory. This allows you to, say, add a new README.md or starter file to a repo that you do not control. The cloned repo can contain its own degit.json
actions.
// degit.json
[
{
"action": "remove",
"files": ["LICENSE"]
}
]
Remove a file at the specified path.
- zel by Vu Tran
- gittar by Luke Edwards
MIT.