Simple Server

This is the backend for the Simple app to help track hypertensive patients across a population.

Development

We have a bin/setup script that does most of the work of getting things setup, but you need a few things in place first. If you are on a Mac, install homebrew and then install rbenv, redis, postgres@14, and yarn:

brew install rbenv ruby-build redis yarn postgresql@14

After this is done, it is highly recommended to tune your local PostgreSQL installation, otherwise your server will get bogged down when doing things like refreshing materialized views. You can use PGTune to do this, it takes about 2 minutes. You can find your local postgresql.conf file at /opt/homebrew/var/postgresql@14/postgresql.conf on M1 Macs, and /usr/local/var/postgresql@14/postgresql.conf on Intel Macs.

bin/setup

To set up the Simple server for local development, clone the git repository and run the setup script included:

$ git clone [email protected]:simpledotorg/simple-server.git
$ cd simple-server
$ bin/setup

Note: If you already have a previous dev environment you're trying to refresh, it's easiest to drop your database run setup again.

$ rails db:drop
$ rails parallel:drop
$ bin/setup

If you encounter issues with this script, please open a new issue with details. Please include the entire log from bin/setup, as well as your computer / OS details.

Setup Troubleshooting

Setup fails when bundler tries to resolve nokogiri

The error message will look like:

Extracting libxml2-2.9.13.tar.xz into <...>
========================================================================
tar (child): xz: Cannot exec: No such file or directory
tar (child): Error is not recoverable: exiting now
/bin/tar: Child returned status 2
/bin/tar: Error is not recoverable: exiting now
========================================================================

Ensure that you have xz installed and linked with brew link xz.

Apple Silicon M1 Macs

With recent gem updates, all of our gems and dependencies now build ARM native on m1 macs. This means you do not need to use Rosetta to set up simple-server, and in fact using Rosetta will make things more complicated and confusing in day to day dev experience, and also hurts performance.

There is one possible caveat to this -- if you see any problems with google-protobuf, run the following:

gem uninstall google-protobuf
gem install google-protobuf -v 3.21.5 --platform=ruby

Then rerun bundler and everything will work. This is being tracked over in protocolbuffers/protobuf#8682, hopefully there will be a better fix soon.

Beyond that, the setup instructions are now the same for Intel or M1 macs, as you can install homebrew normally and go from there.

Docker Compose

Dev environment setup using docker and docker-compose

Prerequisite

• Docker
• Docker compose

Install docker and docker-compose on Mac

brew install docker
brew install docker-compose

Setup

bin/docker-up

After a successful docker-compose initialisation, an admin dashboard account is automatically created.

username: [email protected]
password: Resolve2SaveLives

Open http://localhost:3000 in your browser to view the simple dashboard

Use below Ngrok guide for Android development setup

Teardown (delete docker containers and volumes)

bin/docker-down

Manual Setup

If the included bin/setup script fails for some reason, you can also manually set up the application step by step. You can do so as follows.

First, you need to install ruby. It is recommended to use rbenv to manage ruby versions. Note that we currently use Bundler version 2.3.22, so that is also hardcoded below.

Next, install NodeJS v18.11.0 using nvm.

gem install bundler -v 2.3.22
bundle _2.3.22_ install
brew install nvm
nvm install 18.11.0
rake yarn:install
rails db:setup

We cleanup old migration files every once in a while and so running db:migrate would not work for the initial setup. When setting up a new database, db:setup will take care of everything (it runs db:structure:load under the hood).

Developing with the Android app

To run simple-android app with the server running locally, you can use ngrok.

brew install --cask ngrok
rails server
ngrok http 3000

The output of the ngrok command is HTTP and HTTPS URLs that can be used to access your local server. The HTTP URL cannot be used since HTTP traffic will not be supported by the emulator. Configure the following places with the HTTPS URL.

In the gradle.properties file in the simple-android repository, set:

manifestEndpoint=<HTTPS URL>/api/
fallbackApiEndpoint=<HTTPS URL>/api/

In the .env.development.local (you can create this file if it doesn't exist),

SIMPLE_SERVER_HOST=<URL>  # i.e. without https://
SIMPLE_SERVER_HOST_PROTOCOL=https

Alternatively, you can make the change on the server side. In the server repo, open app/views/api/manifests/show.json.jbuilder. Change:

json.endpoint "#{ENV["SIMPLE_SERVER_HOST_PROTOCOL"]}://#{ENV["SIMPLE_SERVER_HOST"]}/api/"

to:

json.endpoint "<HTTPS URL>/api/"

Workers

We use sidekiq to run async tasks. To run them locally you need to start redis:

redis-server -v

Testing Email

We use Mailcatcher for testing email in development. Please use the following to set it up on your machine.

Note: Please don't add Mailcatcher to the Gemfile, as it causes conflicts.

gem install mailcatcher
mailcatcher

Now you should be able to see test emails at http://localhost:1080

Testing Web Views

When testing web views like the progress tab or help screens, you will need to authenticate yourself with specific request headers. You can run the following command to get a set of request headers for a user that you can attach to your requests.

$ bundle exec rails get_user_credentials

The command will output a set of request headers that you can attach to your requests using tools like Postman or ModHeader.

Attach the following request headers to your requests:
Authorization: Bearer 9b54814d4b422ee37dad46e7ebee673c59eed088c264e479880cbe7fb5ac1ce7
X-User-ID: 452b96c2-e0cf-49e7-ab73-c328acd3f1e5
X-Facility-ID: dcda7d9d-48f9-47d2-b1cc-93d90c94386e

Here are two Simple App pages you can test on your browser:

• "Progress Tab": http://localhost:3000/api/v3/analytics/user_analytics.html
• "Help Page": http://localhost:3000/api/v3/help.html

Review Apps

Every pull request opened on the simple-server repo creates a Heroku review app with the branch's code deployed to it. The review app is hosted at the URL https://simple-review-pr-.herokuapp.com. This temporary environment can be used to test your changes in a production-like environment easily.

If you need to test your changes with a mobile app build as well, you can generate a mobile app build that points to your review app. To do so:

• Navigate to the GitHub Actions page on the simple-server repository
• Select the "Mobile Review App Build" action
• Trigger a "workflow dispatch" at the top of the screen. You can keep the branch as master (it doesn't matter) and enter your PR number in the required input
• Once the Action is complete, its page will contain the APK as an artifact.

Testing messages

Messages sent through Twilio are currently fixed to specific countries. To override this setting, go to the heroku console and add/update the DEFAULT_COUNTRY config variable on your review app to your desired country. The supported country codes are listed here.

# for US/Canada
DEFAULT_COUNTRY = US

# for UK
DEFAULT_COUNTRY = UK

Updating this config will automatically restart the review app and should allow one to receive messages in their appropriate ISD codes.

Configuration

The app uses a base development configuration using .env.development. To add or override any configurations during local development, create a .env.development.local file and add your necessary configurations there. If a configuration change is applicable to all dev environments, ensure that it is added to .env.development and checked into the codebase.

Running the application locally

Foreman can be used to run the application locally. First, install foreman.

$ gem install foreman

Then, run the following command to start the Rails and Sidekiq together.

$ foreman start -f Procfile.dev

Note: Foreman will also execute the whenever gem in trial mode. This will validate that the whenever configuration is valid, but will not actually schedule any cron jobs.

Alternatively, you can start these services locally without foreman by using the following commands individually.

• Rails: bundle exec rails server or bundle exec puma
• Sidekiq: bundle exec sidekiq

Running the tests

bin/rspec

Run tests interactively quickly while developing:

bin/guard

Code

We use the standard gem as our default formatter and linter. To enable it directly in your editor, follow this.

To check all the offenses throughout the codebase:

$ bundle exec standardrb

To fix any offenses that standard can autofix, run

$ bundle exec standardrb --fix

Generating seed data

NOTE: Its highly recommended to tune your local PostgreSQL before generating new seed data, especially large seed data sets. See the docs for that under Development. To generate a full set of seed data, including facilities, users, patients with BPs, etc, run the following:

bin/rails db:seed

You can always do a full reset to get back to a working dataset locally - note that reset clears all DBs, recreates them, runs seed, and refreshes matviews.

bin/rails db:reset

Need a larger dataset? Try adding the SEED_TYPE ENV variable. Available sizes are small, medium, and large, and profiling. Large and profiling take a long time to run (20 mins to an hour), but they are very helpful for performance testing.

SEED_TYPE=medium bin/rails db:reset
# You also may want an entirely new large dataset, with more facilities and regions, and more patients per facility.
SEED_TYPE=large bin/rails db:reset

To purge the generated patient data only, run the following. Note that you usually don't want this, and a full db:reset is safer in terms of generating a valid data set.

$ bin/rails db:purge_users_data

Creating an admin user

If you need new admin users, you can run the following command from the project root. Note that the standard seed process already creates various admins for you, so you probably don't need this for typical dev.

$ bin/rails 'create_admin_user[<name>,<email>,<password>]'

View Sandbox data in your local environment

NOTE: generating seed data locally is the recommended way to get data in your env. Sandbox data is actually just generated via db:seed, so the below process really just adds SCP overhead to the process.

1. Follow the steps in the "How to add an SSH key..." section here to add your SSH key to the deployment repo
2. Ask someone from the Simple team to add you as an admin to Sandbox
3. Create a password for your Sandbox account and use that to log into the Sandbox dashboard on https://api-sandbox.simple.org
4. Run ssh [email protected] to verify that your SSH access from step 1 was completed successfully.
5. Run bundle exec cap sandbox db:pull to sync Sandbox data with your local machine.
6. Use your Sandbox email and password to log into your local environment (http://localhost:3000).

Profiling

We use the vegeta utility to run performance benchmarks. The suite and additional instructions are here.

Security audits

Security audits generally require some test data to be set up in a specific way, and account credentials and other information to be shared with the auditor. Run the following command to set up the necessary test data and print out an information sheet to be shared.

$ bin/rails 'prepare_security_environment'

This task can only be executed in development and security environments.

Documentation

API

API Documentation can be accessed at /api-docs on local server and hosted at https://api.simple.org/api-docs

To regenerate the Swagger API documentation, run the following command.

$ bundle exec rake docs

ADRs

Architecture decisions are captured in ADR format and are available in /doc/arch

Wiki

Guides, instructions and long-form maintenance documentation can go in /doc/wiki

ERD (Entity-Relationship Diagram)

These are not actively committed into the repository. But can be generated by running bundle exec erd

Deployment

Simple Server is continuously deployed from master to all environments via Semaphore Workflows as long as the build passes.

If you need to make a manual production release, run the release script from master:

bin/release

This will create a git release tag and automatically trigger a deployment to all environments through Semaphore. You can monitor the deployment progress in Semaphore via the tagged release's workflow. Please make sure to copy / paste the changelog from bin/release so you can post it in the #releases channel.

Deployment Resources

The infrastructure setup including the ansible and terraform scripts are documented in the deployment repository.

Contributing

If you're working on a project that will affect any of the indicators listed in this document, please contact the product / design team.