Skip to content

rickecon/PolicyBrain

 
 

Repository files navigation

Build Status

Taxbrain Web Application

Deploying To Heroku

Production

Heroku relies upon GIT for deployment. The following commands are for deployment. Before deploying to Heroku make certain to run the following command* locally and commit the changes:

./manage.py collectstatic --noinput

* Currently this command fails to create the staticfiles dir on Heroku, so this has to be run locally and the changes committed.

Deploy to Heroku:

git push heroku master

If any of the models have been updated in a deployment then those changes have to migrated so the database schema is updated. The migration should be created locally, committed, and pushed to Heroku as opposed to running the makemigration command on Heroku. Locally run the following:

./manage.py makemigrations or python manage.py makemigrations (both commands have the same effect.)

Commit the files that are created.

In order to run those migrations, run the following commands:

heroku run ./manage.py migrate or heroku run python manage.py migrate (both commands have the same effect.)

Your schema will now be udpated.

DEBUG: This should always be False on the production site. Make certain the environment variable DEV_DEBUG is set to anything but the string 'True'. To check what variables are set to in Heroku's environment run the following command:

heroku config

This will list all config variables (alternatively go to the Settings section on the Heroku dashboard and click the 'Reveal Config Vars' button.)

Development

A live development server works exactly the same way as a production server on Heroku does. All commands and configurations are the same. The only difference is that the requirements_dev.txt can be used on this server to have access to the various debugging tools.

Post deployment to a dev server runt he following command in order to install the dev tools (if desired): heroku run pip install -r requirement_dev.txt

The migration commands presented in the Production section also hold true for development. Post deployment make sure to run the migration commands.

DEBUG: In order to turn DEBUG on, set the DEV_DEBUG environment variable to the string 'True'. DEBUG will be off id set to anything but 'True'. In order to change the DEV_DEBUG variable use the following command: heroku config:set DEV_DEBUG='True'

The above command will turn DEBUG on. Setting to any other string will turn DEBUG off.

Release Process

To review the steps for the release process, see RELEASE_PROCESS.md

Installing Locally

To develop the application locally, you need a script containing the environment variables necessary for communicating with the computation nodes. You should ask your project manager or administrator for this script.

Installing PostgreSQL

Postgres is used for this project. If PostgreSQL is not installed on your machine follow the instructions in the following links.

OS X: http://www.postgresql.org/download/macosx/

Ubunutu: https://help.ubuntu.com/community/PostgreSQL

Windows (should support 7 & 8): http://www.postgresql.org/download/windows/

Create database taxcalc

After PostgreSQL has been successfully installed make sure to create the taxcalc database.

For Ubuntu the command is:

createdb taxcalc

If you need to add a user and grant permissions make sure to have psql installed as well.

Create a directory

Make sure to create a directory wherever you keep your projects.

mkdir project_name

Clone the repo in the new directory

git clone git@github.com:OpenSourcePolicyCenter/webapp-public.git

Create a conda environment for your local development

Follow these instructions in README_local_development.md for installing all the Python packages you'll need and installing the right versions of the open-source-economics packages.

Then make sure you have done:

source activate aei_dropq

before continuing the remaining instructions.

Using foreman

Since Django can be a more involved process using foreman might be a better tool for some situations. If no work will be done on the back end, then foreman might be the tool of choice (although once Django is setup all you will have to do is run the server locally to make use of it.)

Some important points:

  • You must start the virtual environment.
  • The Heroku toolbar that is installed with the requirements.txt file is expecting PostgreSQL to exist on your system.

After the environment is activated the following command will start foreman:

foreman start

Once the server has started foreman will use port 5000.

Building Static Files

To setup your environment for building static assets:

npm install
npm install -g bower
bower install

To compile LESS assets:

gulp less

Collect static files and make sure changes get committed back:

python manage.py collectstatic

Using Django

Once all the dependecies are installed a couple of commands are necessary to get the Django project up and running and onto ./manage.py runserver. Make sure the virtual environment is activated.

First update the settings.py file located in the webapp dir (webapp/settings.py)

In the settings.py file there is a database configuration that looks like:

DATABASES = {
'default': {
    'ENGINE': 'django.db.backends.postgresql_psycopg2',
    'NAME': 'taxcalc',
    'USER': 'postgres',
    'PASSWORD': '',
    'HOST': '',
    'PORT': '5432',
    }
}
  1. Change the USER to your user name (the one you used to setup Postgres).
  2. Change the PASSWORD to the password you used to setup Postgres.
  3. Change HOST to 127.0.0.1

Using SQLite as an alternative to Postgres

Your environment settings file should include a variable named DATABASE_URL. If you would rather use SQLite than Postgres for your development, simply comment out or delete this line. The application will default to using SQLite as your database.

PLEASE DO NOT COMMIT YOUR LOCAL CHANGES TO THE DATABASE CONFIG IN THE SETTINGS FILE. GIT STASH THEM!

ALTERNATIVELY, STOP TRACKING LOCAL CHANGES TO THIS FILE WITH:

git update-index --assume-unchanged webapp/settings.py

Next change the DEBUG & TEMPLATE_DEBUG settings to True.

Second migrations have to been run. Migrations manage the schema for all database objects. Go to the root of the project simply run:

./manage.py migrate

Django will then run the migrations and all the tables will be created in the db. NOTE: it is critical that migrations are run when updating models, otherwise the changes to the models will not be recognized and errors will be thrown by Django (these errors tend to be increasingly informative and you will usually be prompted by Django that there are unmigrated changes when you run the local server.)

Third, it's time run the server. Simply run:

./manage.py runserver

Now you have a live project being run locally at least for the front end Django app. You'll be able to see the pages and forms of ospc.org on http://localhost:8000 but you won't be able to submit and run jobs without error unless you follow the instructions below for starting flask, celery and redis.

Fourth, run celery:

In another separate terminal tab or window:

taxbrain-celery-worker

Fifth, run flask:

In another separate terminal tab or window:

taxbrain-flask-worker

Sixth, run redis

In another separate terminal tab or window:

redis-server

Seventh, if you edit your code...

If you edit your code or reinstall any part of the code, including even javascript assets, then you must to Ctrl+C to kill and restart each of the following processes:

  • The runserver process
  • The taxbrain-celery-worker process
  • The taxbrain-flask-worker process

About

Source code for PolicyBrain, ospc.org, and related assets.

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • CSS 73.6%
  • Python 18.1%
  • HTML 4.2%
  • JavaScript 3.8%
  • Other 0.3%