MLTSHP is a Dockerized application. This greatly simplifies running the application for local development. We also deploy the application using Docker, so it makes for a more consistent environment to build and test against.
With Docker and a git client installed on your computer, clone the MLTSHP
code from Github. If you intend to develop features for MLTSHP, you should
clone from your own fork of the code. Once you have a copy checked out
locally, use this command to create a settings.py and celeryconfig.py
file suitable for local development (edit these as needed, but the defaults
should be okay):
$ make init-dev
You should be able to start the app itself using:
$ make run
This will do a lot of things. But ultimately, you should end up with a
copy of MLTSHP running locally. It expects to be accessed via a hostname
of mltshp.localhost and s.mltshp.localhost. Add these entries to your /etc/hosts
file:
127.0.0.1 mltshp.localhost s.mltshp.localhost
The web app itself runs on port 8000. You should be able to reach it via:
Subsequent invocations of make run should be faster, once you have
the dependency images downloaded.
You can login as the admin user using the password password. You
can also register new user accounts.
While running the service, you can open an editor to the git checkout and make updates to the code. Changes should be reflected as you save your files (no need to restart the service).
The MySQL instance that is launched will be accessible on 127.0.0.1:3306
if you want to look at the database directly (since this is using the
default MySQL port, you will probably need to shutdown any existing MySQL
server you may have running). The login for the database is root with
no password. If you want to mark any of your user accounts as paid users,
find them in the user table and set their is_paid value to 1 and
their stripe_plan_id column value to mltshp-double.
When you run the application, it launches it into a background process. But if you want to watch the realtime logs emitted by each service, just use this command:
$ docker-compose logs -f
In addition to that, the web app produces some log files that are captured under the "mounts/logs" folder of your git repository. The directory structure looks like this:
mounts/
uploaded/
(transient uploaded file storage)
logs/
access.log - nginx access log file
error.log - nginx error log file
main-8000.log - python app log file
celeryd-01.log - celery worker log file
fakes3/
(local S3 storage)
mysql/
(mysql data files)
MLTSHP utilizes AWS S3 for storing uploaded images. The development
environment provides a dummy S3 server for local operation. But it requires
a license key in order to use it. Visit this page
to obtain a license key. For individual developers and small organizations,
there is no cost. Add the following to a local .env file in the root
of the project:
FAKES3_LICENSE_KEY=your-license-key-here
You will find any uploaded files under the `mounts/fakes3' directory.
Note: As of this time, the Docker image for the fake S3 server is incompatible with Apple Silicon CPU architectures. If you are using a computer with an Apple ARM CPU, you will need to use a real S3 bucket (see below).
If you would rather use a real S3 bucket, you can do that too. Create one and then assign these in your local settings.py file:
"aws_bucket": "your-mltshp-bucket-name",
"aws_key": "your-aws-key",
"aws_secret": "your-aws-secret",
## Comment these entries out:
##"aws_host": "fakes3",
##"aws_port": 8000,
Occassionally, a database migration will need to be performed to
bring an older database schema up to date. This will be necessary
if you have a MySQL database you use locally for development and
testing and keep it versus using the destroy and init-dev
commands to make a new one. To update your database, just do this:
$ make shell
docker-shell$ cd /srv/mltshp.com/mltshp; python migrate.py
That should do it.
With your copy of MLTSHP running, you may want to run unit tests.
Some of the unit tests actually test against Twitter itself, so you'll
want to generate a custom Twitter app with your own set of keys.
Configure the test_settings in your settings.py file appropriately:
"twitter_consumer_key" : "twitter_consumer_key_here",
"twitter_consumer_secret" : "twitter_consumer_secret_key_here",
"twitter_access_key" : "twitter_access_key_here",
"twitter_access_secret" : "twitter_access_secret_here",
Then, just run:
$ make test
Which will invoke a Docker process to run the unit test suite.
If you ever need to access the Docker image running the application, you can use this command to create a shell (specifically, a shell to the Python web application container):
$ make shell
This should place you in the /srv/mltshp.com/mltshp directory as the
root user. You can use apt-get commands to install utilities you
may need.
If you ever want to wipe your local data and rebuild your Docker containers, just use this command:
$ make destroy
If you just wish to rebuild the Docker container, use the Docker compose command:
$ docker-compose down
Then, run another make run.
The CSS in this repo is just the compiled version of the styles from the MLTSHP pattern library, which can be found in the mltshp-patterns repo. Please do not edit the CSS in this repo, since any changes will be lost the next time we update from the pattern library.
MLTSHP is open-source software, ©2017 the MLTSHP team and released to the public under the terms of the Mozilla Public License. A copy of the MPL can be found in the LICENSE file.
MLTSHP is proudly powered by Fastly.