MLTSHP

Status

Development Environment

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:

http://mltshp.localhost:8000/

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.

Logs and Data

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)

AWS S3 Storage

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,

Database Migrations

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.

Tests

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.

Connecting to the MLTSHP shell

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.

Cleanup

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.

Relationship with MLTSHP-Patterns

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 / mltshp