Heroku Slugify

A command line tool to deploy your application to Heroku without using git push. It creates a source tarball, uploads it, builds a slug and deploys it on Heroku. The Heroku Dev Center article Building and Releasing Using the Platform API explains the details.

Installation

$ npm install heroku-slugify --global

Configuration

heroku-slugify uses node-heroku-client to talk to the Heroku API. You must set the HEROKU_API_KEY environment variable, otherwise authentication will fail.

Usage

$ heroku-slugify --help
usage: heroku-slugify [-h] -a APP -v VERSION [-s SOURCE] [-t TIMEOUT]
                      [-i INTERVAL]


Deploy a new Heroku slug

Optional arguments:
  -h, --help            Show this help message and exit.
  -a APP, --app APP     Heroku app id
  -v VERSION, --version VERSION
                        Version to deploy
  -d DIRECTORY, --directory DIRECTORY
                        Name of the directory containing the files for the
                        tarball (default: heroku)
  -s SOURCE, --source SOURCE
                        File name of source tarball (default: source.tar.gz)
  -t TIMEOUT, --timeout TIMEOUT
                        Timeout for build in seconds (default: 120 s)
  -i INTERVAL, --interval INTERVAL
                        Interval to check build status in seconds (default:
                        10 s)

Create a source tarball, upload it, build a slug and deploy it on Heroku.

Example: Deploying Jupyter Notebook on Heroku

A simple example is to deploy a Jupyter Notebook on a Heroku Dyno. Instead of creating a Git repository and pushing it to the heroku remote we just create three configuration files and deploy them using heroku-slugify.

The following steps show you how to do this. You just need a Heroku account to start!

By default heroku-slugify will collect all files in the directory heroku to create the source tarball that contains the Heroku slug configuration. So your first step is to create this directory:

$ mkdir heroku

Now you specify the Python packages you want to install by writing them into the heroku/requirements.txt file:

$ echo "jupyter" > heroku/requirements.txt

The next step sets the Python version your Dyno will use:

$ echo "python-3.5.1" > heroku/runtime.txt

You also need a Procfile to configure the Jupyter Notebook webserver:

$ echo "web: jupyter notebook --no-browser --ip=* --port=$PORT /tmp" > heroku/Procfile

Now you should have a heroku directory with three files in it:

$ tree heroku/
heroku/
├── Procfile
├── requirements.txt
└── runtime.txt

0 directories, 3 files

That's all you need to deploy your Heroku Dyno!

If you don't have a Heroku app you can use for testing, create one now:

$ heroku apps:create --no-remote

As explained in the configuration section you must set the HEROKU_API_KEY environment variable, otherwise authentication will fail:

$ export HEROKU_API_KEY=<YOUR_HEROKU_API_KEY_HERE>

Now you are ready to deploy the Jupyter Notebook using heroku-slugify:

$ heroku-slugify --app <YOUR_HEROKU_APP_ID_HERE> --version 1.0
[2016-03-21 16:22:12.637] [INFO]  app id: rocky-oasis-34863
[2016-03-21 16:22:12.642] [INFO]  version: 1.0
[2016-03-21 16:22:12.643] [INFO]  source tarball: source.tar.gz
[2016-03-21 16:22:12.643] [INFO]  timeout: 120 s
[2016-03-21 16:22:12.643] [INFO]  interval: 10 s
[2016-03-21 16:22:12.643] [LOG]   rm -f source.tar.gz
[2016-03-21 16:22:12.653] [LOG]   tar czvf source.tar.gz -C heroku/ $(ls heroku/)
[2016-03-21 16:22:12.663] [LOG]   a Procfile
[2016-03-21 16:22:12.663] [LOG]   a requirements.txt
[2016-03-21 16:22:12.663] [LOG]   a runtime.txt
[2016-03-21 16:22:12.663] [LOG]
[2016-03-21 16:22:13.270] [LOG]   curl "https://s3-external-1.amazonaws.com/heroku-sources-production/heroku.com/ef0dc77f-6047-4be8-ab25-d3154215af31?AWSAccessKeyId=AKIAJURUZ6XB34ESX54A&Signature=mcxRSHmAS6hor9ekpIVb%2B5RDBBg%3D&Expires=1458577333" -X PUT -H "Content-Type:" --data-binary @source.tar.gz
[2016-03-21 16:22:14.230] [LOG]     % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
[2016-03-21 16:22:14.230] [LOG]                                    Dload  Upload   Total   Spent    Left  Speed
100   272    0     0  100   272      0    286 --:--:-- --:--:-- --:--:--   286
[2016-03-21 16:22:14.230] [LOG]
[2016-03-21 16:22:15.830] [DIR]   { app: { id: '8ea98634-b053-44cc-a26c-d17c966dc2a7' },
  buildpacks: null,
  created_at: '2016-03-21T15:22:15+00:00',
  id: '704b7a99-bc7c-441f-b00c-4d54c5b36f16',
  output_stream_url: 'https://build-output.heroku.com/streams/8e/8ea98634-b053-44cc-a26c-d17c966dc2a7/logs/70/704b7a99-bc7c-441f-b00c-4d54c5b36f16.log?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=AKIAJQUUZPBDLMDG7K7Q%2F20160321%2Fus-east-1%2Fs3%2Faws4_request&X-Amz-Date=20160321T152215Z&X-Amz-Expires=86400&X-Amz-SignedHeaders=host&X-Amz-Signature=8b89368f7e5c9302950592463cad4de85e7530f07b1caf0149e529278e8b3ca4',
  slug: null,
  source_blob:
   { checksum: null,
     url: 'https://s3-external-1.amazonaws.com/heroku-sources-production/heroku.com/ef0dc77f-6047-4be8-ab25-d3154215af31?AWSAccessKeyId=AKIAJURUZ6XB34ESX54A&Signature=QTPd7vWCQNXl0SbFrkJSpxeHAOk%3D&Expires=1458577333',
     version: '1.0',
     version_description: null },
  status: 'pending',
  updated_at: '2016-03-21T15:22:15+00:00',
  user:
   { email: '[email protected]',
     id: '573b6879-718b-4beb-a075-1324534e5dd8' } }
[2016-03-21 16:22:16.413] [LOG]   build status: pending
[2016-03-21 16:22:27.039] [LOG]   build status: pending
[2016-03-21 16:22:37.660] [LOG]   build status: pending
[2016-03-21 16:22:48.288] [LOG]   build status: pending
[2016-03-21 16:22:58.891] [LOG]   build status: pending
[2016-03-21 16:23:09.522] [LOG]   build status: pending
[2016-03-21 16:23:20.174] [LOG]   build status: pending
[2016-03-21 16:23:30.806] [LOG]   build status: pending
[2016-03-21 16:23:41.403] [LOG]   build status: pending
[2016-03-21 16:23:52.014] [LOG]   build status: pending
[2016-03-21 16:24:02.787] [LOG]   build status: succeeded

Now open the Jupyter Notebook in your browser:

$ heroku open --app <YOUR_HEROKU_APP_ID_HERE>

That's it! heroku-slugify makes it as simply as possible to deploy your packaged application to a Heroku Dyno. You can not only deploy Python packages, but any package format a Heroku buildpack exists for.

In case you want to destroy your Heroku Dyno, run the following command:

$ heroku apps:destroy --app <YOUR_HEROKU_APP_ID_HERE>

SECURITY ADVICE

Please note that anyone who has the URL of your Heroku Dyno has full access to the host! This is just an example and not something you should run in production!