This project is the core of Notify.gov. It's cloned from the brilliant work of the team at GOV.UK Notify, cheers!
This repo contains:
- A public-facing REST API for Notify.gov, which teams can integrate with using API clients built by UK.
- An internal-only REST API built using Flask to manage services, users, templates, etc., which the Notify.gov Admin UI talks to.
- Asynchronous workers built using Celery to put things on queues and read them off to be processed, sent to providers, updated, etc.
Our other repositories are:
You will need the following items:
- An active cloud.gov account with the correct permissions - speak with your
onboarding buddy for help with
setting up an account (requires a
.mil,.gov, or.fed.usemail address) and getting access to thenotify-local-devandnotify-stagingspaces. - Admin priviliges and SSH access on your machine; you may need to work with your organization's IT support staff if you're not sure or don't currently have this access.
This project currently works with these major versions of the following main components:
- Python 3.12.x
- PostgreSQL 15.x (version 12.x is used in the hosted environments)
These instructions will walk you through how to set your machine up with all of the required tools for this project.
On MacOS, using Homebrew for package management is highly recommended. This helps avoid some known installation issues. Start by following the installation instructions on the Homebrew homepage.
Note: You will also need Xcode or the Xcode Command Line Tools installed. The quickest way to do this is is by installing the command line tools in the shell:
xcode-select –-installIf this is your first time installing Homebrew on your machine, you may need to
add its binaries to your system's $PATH environment variable so that you can
use the brew command. Try running brew help to see if Homebrew is
recognized and runs properly. If that fails, then you'll need to add a
configuration line to wherever your $PATH environment variable is set.
Your system $PATH environment variable is likely set in one of these
locations:
For BASH shells:
~/.bashrc~/.bash_profile~/.profile
For ZSH shells:
~/.zshrc~/.zprofile
There may be different files that you need to modify for other shell environments.
Which file you need to modify depends on whether or not you are running an interactive shell or a login shell (see this Stack Overflow post for an explanation of the differences). If you're still not sure, please ask the team for help!
Once you determine which file you'll need to modify, add these lines before any
lines that add or modify the $PATH environment variable; near or at the top
of the file is appropriate:
# Homebrew setup
eval "$(/opt/homebrew/bin/brew shellenv)"This will make sure Homebrew gets setup correctly. Once you make these changes,
either start a new shell session or source the file
(source ~/.FILE-YOU-MODIFIED) you modified to have your system recognize the
changes.
Verify that Homebrew is now working by trying to run brew help again.
There are several packages you will need to install for your system in order to get the app running (and these are good to have in general for any software development).
Start off with these packages since they're quick and don't require additional configuration after installation to get working out of the box:
- jq - for working with JSON in the command line
- git - for version control management
- tfenv - for managing Terraform installations
- cf-cli@8 - for working with a Cloud Foundry platform (e.g., cloud.gov)
- redis - required as the backend for the API's asynchronous job processing
- vim - for editing files more easily in the command line
- wget - for retrieving files in the command line
You can install them by running the following:
brew install jq git tfenv cloudfoundry/tap/cf-cli@8 redis vim wgetAs a part of the installation above, you just installed tfenv to manage
Terraform installations. This is great, but you still need to install Terraform
itself, which can be done with this command:
tfenv install "latest:^1.7"
tfenv use 1.7.x # x = the patch version installedNow we're going to install a tool to help us manage Python versions and virtual environments on our system. First, we'll install pyenv and one of its plugins, pyenv-virtualenv, with Homebrew:
brew install pyenv pyenv-virtualenvWhen these finish installing, you'll need to make another adjustment in the
file that you adjusted for your $PATH environment variable and Homebrew's
setup. Open the file, and add these lines to it:
# pyenv setup
export PYENV_ROOT="$HOME/.pyenv"
command -v pyenv >/dev/null || export PATH="$PYENV_ROOT/bin:$PATH"
eval "$(pyenv init -)"
eval "$(pyenv virtualenv-init -)"
Once again, start a new shell session or source the file in your current shell session to make the changes take effect.
Now we're ready to install the Python version we need with pyenv, like so:
pyenv install 3.12This will install the latest version of Python 3.12.
NOTE: This project currently runs on Python 3.12.x.
Lastly, we need to install the tool we use to manage Python dependencies within the project, which is poetry.
Visit the official installer instructions page and follow the steps to install Poetry directly with the script.
This will ensure poetry doesn't conflict with any project virtual environments
and can update itself properly.
We now need to install a database - this project uses PostgreSQL, and Homebrew requires a version number to be included with it when installing it:
brew install postgresql@15NOTE: This project currently works with PostgreSQL version 15.x; version 12.x is currently used in our hosted environments.
NOTE: If you have a pre-existing instance of PSQL installed because of another product like PGAdmin, your database configuration may differ from the instructions above, which uses Homebrew to install and configure PostgreSQL. If this is the case for you, you may have to either account for slightly different user permissions with the database, or uninstall PGAdmin and/or PostgreSQL itself, and reinstall it with Homebrew to follow the steps above.
You'll now need to modify (or create, if it doesn't already exist) the $PATH
environment variable to include the PostgreSQL binaries. Open the file you have
worked with before to adjust your shell environment with the previous steps and
do one of the following:
If you already have a line that modifies the $PATH environment variable, just
add this path into the existing string:
/opt/homebrew/opt/postgresql@15/bin
If you don't have a line for your $PATH environment variable, add it in like
this, which will include the PostgreSQL binaries:
export PATH="/opt/homebrew/opt/postgresql@15/bin:$PATH
NOTE: You don't want to overwrite your existing $PATH environment variable! Hence the reason why it is included on the end like this; paths are separated by a colon.
With both PostgreSQL and Redis installed, you now need to start the services. Run this command so that they're available at all times going forward on your machine:
brew services start postgresql@15
brew services start redisIf they're already running, you can run this command instead to make sure the latest updates are applied to both services:
brew services restart postgresql@15
brew services restart redisOnce all of pre-requisites for the project are installed and you have a cloud.gov account, you can now set up the API project and get things running locally!
First, clone the repository in the directory of your choosing on your machine:
git clone git@github.com:GSA/notifications-api.gitNow go into the project directory (notifications-api by default), create a
virtual environment, and set the local Python version to point to the virtual
environment (assumes version Python 3.12.2 is what is installed on your
machine):
cd notifications-api
pyenv virtualenv 3.12.2 notify-api
pyenv local notify-apiIf you're not sure which version of Python was installed with pyenv, you can check by running pyenv versions and it'll list everything available currently.
Now log into cloud.gov in the command line by using this command:
cf login -a api.fr.cloud.gov --ssoIf you are offered a choice of orgs, select gsa-tts-benefits-studio.
For the space, choose notify-local-dev to start with (assuming you are
setting up local development).
REMINDER: Ensure you have access to the notify-local-dev and notify-staging spaces in cloud.gov
Now run the development Terraform setup by navigating to the development folder and running the script in it:
cd terraform/development
./run.shIf this runs correctly, Terraform will ask you if you want to create some
resources. Answer yes.
The script will also create a local .env file for you in the project's
root directory, which will include a handful of project-specific environment
variables.
Lastly, if you didn't already start PostgreSQL and Redis above, be sure to do so now:
brew services start postgresql@15
brew services start redisIf you're upgrading an existing project to a newer version of Python, you can follow these steps to get yourself up-to-date.
First, use pyenv to install the newer version of Python you'd like to use;
we'll use 3.12 in our example here since we recently upgraded to this version:
pyenv install 3.12Next, delete the virtual environment you previously had set up. If you followed
the instructions above with the first-time set up, you can do this with pyenv:
pyenv virtualenv-delete notify-apiNow, make sure you are in your project directory and recreate the same virtual environment with the newer version of Python you just installed:
cd notifications-api
pyenv virtualenv 3.12.2 notify-api
pyenv local notify-apiAt this point, proceed with the rest of the instructions here in the README and you'll be set with an upgraded version of Python.
If you're not sure about the details of your current virtual environment, you can run poetry env info to get more information. If you've been using pyenv for everything, you can also see all available virtual environments with pyenv virtualenvs.
There's one final thing to adjust in the newly created .env file. This
project has support for end-to-end (E2E) tests and has some additional checks
for the presence of an E2E test user so that it can be authenticated properly.
In the .env file, you should see this section:
#############################################################
# E2E Testing
NOTIFY_E2E_TEST_EMAIL=example@fake.gov
NOTIFY_E2E_TEST_PASSWORD="don't write secrets to the sample file"
You can leave the email address alone or change it to something else to your liking.
You should absolutely change the NOTIFY_E2E_TEST_PASSWORD environment
variable to something else, preferably a lengthy passphrase.
With those two environment variable set, the database migrations will run properly and an E2E test user will be ready to go for use in the admin project.
Note: Whatever you set these two environment variables to, you'll need to match their values on the admin side. Please see the admin README and documentation for more details.
The first time you run the project you'll need to run the project setup from the root project directory:
make bootstrapThis command is handled by the Makefile file in the root project directory, as
are a few others.
NOTE: You'll want to occasionally run make bootstrap to keep your project up-to-date, especially when there are dependency updates.
Now you can run the web server and background workers for asynchronous jobs:
make run-procfileIf it runs correctly, you will be able to visit http://127.0.0.1:6011/ and see JSON from the API in your web browser.
This will run all of the services within the same shell session. If you need to run them separately to help with debugging or tracing logs, you can do so by opening three sepearate shell sessions and running one of these commands in each one separately:
make run-celery- Handles the asynchronous jobsmake run-celery-beat- Handles the scheduling of asynchronous jobsmake run-flask- Runs the web server
We're using Poetry for managing our Python
dependencies and local virtual environments. When it comes to managing the
Python dependencies, there are a couple of things to bear in mind.
For situations where you manually manipulate the pyproject.toml file, you
should use the make py-lock command to sync the poetry.lock file. This will
ensure that you don't inadvertently bring in other transitive dependency updates
that have not been fully tested with the project yet.
If you're just trying to update a dependency to a newer (or the latest) version, you should let Poetry take care of that for you by running the following:
poetry update <dependency> [<dependency>...]You can specify more than one dependency together. With this command, Poetry will do the following for you:
- Find the latest compatible version(s) of the specified dependency/dependencies
- Install the new versions
- Update and sync the
poetry.lockfile
In either situation, once you are finished and have verified the dependency
changes are working, please be sure to commit both the pyproject.toml and
poetry.lock files.
On M1 Macs, if you get a fatal error: 'Python.h' file not found message, try a
different method of installing Python. The recommended approach is to use
pyenv, as noted above in the installation
instructions.
If you're using PyCharm for Python development, we've noticed some quirkiness with the IDE and the interaction between Poetry and virtual environment management that could cause a variety of problems to come up during project setup and dependency management. Other tools, such as Visual Studio Code, have proven to be a smoother experience for folks.
A direct installation of PostgreSQL will not put the createdb command on your
$PATH. It can be added there in your shell startup script, or a
Homebrew-managed installation of PostgreSQL will take care of it. See the
instructions above for more details.
- Infrastructure overview
- Using the logs
- Testing
- Deploying
- Database management
- One-off tasks
- Test Loading Commands
- How messages are queued and sent
- Writing public APIs
- API Usage
- Queues and tasks
- Notify.gov
- Pull Requests
- Code Reviews
- Run Book
- Data Storage Policies & Procedures
Work through
commit e604385
is licensed by the UK government under the MIT license. Work after that commit
is in the worldwide public domain. See LICENSE.md for more
information.
As stated in CONTRIBUTING.md, all contributions to this project will be released under the CC0 dedication. By submitting a pull request, you are agreeing to comply with this waiver of copyright interest.