Baby Buddy

A buddy for babies! Helps caregivers track sleep, feedings, diaper changes, and tummy time to learn about and predict baby's needs without (as much) guess work.

Table of Contents

• Demo
• Deployment
  • AWS Elastic Beanstalk
  • Docker
  • Heroku
  • Manual
• Configuration
• API
  • Authentication
  • GET Method
  • OPTIONS Method
  • POST Method
• Development
  • Installation
  • Gulp Commands
  • Pre-commit Hook

Demo

A demo of Baby Buddy is available on Heroku. The demo instance resets every hour. Login credentials are:

• Username: admin
• Password: admin

Deployment

The default user name and password for Baby Buddy is admin/admin. For any deployment, log in and change the default admin password immediately.

Many of Baby Buddy's configuration settings can be controlled using environment variables - see Configuration for detailed information.

AWS Elastic Beanstalk

A basic Elastic Beanstalk configuration is provided in .ebextensions/babybuddy.config. The steps below are a rough guide to deployment. See Working with Python for detailed information.

1. Clone/download the Baby Buddy repo

    git clone https://github.com/cdubz/babybuddy.git
   

2. Enter the cloned/downloaded directory

    cd babybuddy
   

3. Change the SECRET_KEY value to something random in .ebextensions/babybuddy.config

4. Create an IAM user in AWS with EB, EC2, RDS and S3 privileges.

5. Initialize the Elastic Bean application (using the IAM user from the previous step)

    eb init -p python-3.6
   

6. Create/deploy the environment! 🚀

    eb create -db -db.engine postgres
   

The create command will also do an initial deployment. Run eb deploy to redeploy the app (e.g. if there are errors or settings are changed).

Docker

A Docker deployment requires Docker and Docker Compose to create two containers - one for the database and one for the application.

1. Copy the docker.env.example to docker.env and set the ALLOWED_HOSTS and SECRET_KEY variables within

    cp docker.env.example docker.env
    editor docker.env
   

   See Configuration for other settings that can be controlled by environment variables added to the docker.env file.

2. Build/run the application

    docker-compose up -d
   

3. Initialize the database (first run/after updates)

    docker-compose exec app python manage.py migrate
   

The app should now be locally available at http://127.0.0.1:8000. See Get Started, Part 6: Deploy your app for detailed information about how to deployment methods with Docker.

Heroku

For manual deployments to Heroku without using the deploy button, make sure to create the following settings before pushing:

heroku config:set DJANGO_SETTINGS_MODULE=babybuddy.settings.heroku
heroku config:set SECRET_KEY=<CHANGE TO SOMETHING RANDOM>
heroku config:set DISABLE_COLLECTSTATIC=1

See Configuration for other settings that can be controlled by heroku config:set.

Manual

There are many ways to deploy Baby Buddy manually to any server/VPS. The basic requirements are Python, a web server, an application server, and a database.

Requirements

• Python 3.5+, pip, pipenv
• Web server (nginx, Apache, etc.)
• Application server (uwsgi, gunicorn, etc.)
• Database (sqlite, Postgres, MySQL, etc.)

Example deployment

This example assumes a 512MB VPS instance with Ubuntu 16.04. It uses Python 3.5+, nginx, uwsgi and sqlite and should be sufficient for a few users (e.g. two parents and 1+ child).

1. Install system packages

    sudo apt-get install python3 python3-pip nginx uwsgi uwsgi-plugin-python3 git libopenjp2-7-dev
   

2. Default python3 to python for this session

    alias python=python3
   

3. Install pipenv

    sudo -H pip3 install pipenv
   

4. Set up directories and files

    sudo mkdir /var/www/babybuddy
    sudo chown user:user /var/www/babybuddy
    mkdir -p /var/www/babybuddy/data/media
    git clone https://github.com/cdubz/babybuddy.git /var/www/babybuddy/public
   

5. Move in to the application folder

    cd /var/www/babybuddy/public
   

6. Initiate and enter the Python environment

    pipenv install --three
    pipenv shell
   

   Note: Python dependencies are locked on x86-64 architecture. Installs on other architectures (like Raspberry Pi's ARM) may result in a THESE PACKAGES DO NOT MATCH THE HASHES FROM Pipfile.lock! error. Add the --skip-lock flag to the above command to suppress this error (i.e.: pipenv install --three --dev --skip-lock).

7. Create a production settings file and set the SECRET_KEY and ALLOWED_HOSTS values

    cp babybuddy/settings/production.example.py babybuddy/settings/production.py
    editor babybuddy/settings/production.py
   

8. Initiate the application

    export DJANGO_SETTINGS_MODULE=babybuddy.settings.production
    python manage.py migrate
   

9. Set appropriate permissions on the database and data folder

    sudo chown -R www-data:www-data /var/www/babybuddy/data
    sudo chmod 640 /var/www/babybuddy/data/db.sqlite3
    sudo chmod 750 /var/www/babybuddy/data
   

10. Create and configure the uwsgi app

     sudo editor /etc/uwsgi/apps-available/babybuddy.ini
    

    Example config:

     [uwsgi]
     plugins = python3
     project = babybuddy
     base_dir = /var/www/babybuddy
    
     virtualenv = /path/to/venv
     chdir = %(base_dir)/public
     module =  %(project).wsgi:application
     env = DJANGO_SETTINGS_MODULE=%(project).settings.production
     master = True
     vacuum = True
    

    See the uWSGI documentation for more advanced configuration details.

    Note: Find the location of the pipenv virtual environment for the virtualenv parameter with the command pipenv --venv.

11. Symlink config and restart uWSGI:

     sudo ln -s /etc/uwsgi/apps-available/babybuddy.ini /etc/uwsgi/apps-enabled/babybuddy.ini
     sudo service uwsgi restart
    

12. Create and configure the nginx server

     sudo editor /etc/nginx/sites-available/babybuddy
    

    Example config:

     upstream babybuddy {
         server unix:///var/run/uwsgi/app/babybuddy/socket;
     }
    
     server {
         listen 80;
         server_name babybuddy.example.com;
    
         location / {
             uwsgi_pass babybuddy;
             include uwsgi_params;
         }
    
         location /media {
             alias /var/www/babybuddy/data/media;
         }
     }
    

    See the nginx documentation for more advanced configuration details.

13. Symlink config and restart NGINX:

     sudo ln -s /etc/nginx/sites-available/babybuddy /etc/nginx/sites-enabled/babybuddy
     sudo service nginx restart
    

14. That's it (hopefully)! 🎉

Configuration

Environment variables can be used to define a number of configuration settings. Baby Buddy will check the application directory structure for an .env file or take these variables from the system environment. System environment variables take precedence over the contents of an .env file.

• ALLOWED_HOSTS
• ALLOW_UPLOADS
• AWS_ACCESS_KEY_ID
• AWS_SECRET_ACCESS_KEY
• AWS_STORAGE_BUCKET_NAME
• DEBUG
• NAP_START_MAX
• NAP_START_MIN
• SECRET_KEY
• TIME_ZONE

ALLOWED_HOSTS

Default: * (any)

This option may be set to a single host or comma-separated list of hosts (without spaces). This should always be set to a specific host or hosts in production deployments.

See also: Django's documentation on the ALLOWED_HOSTS setting

ALLOW_UPLOADS

Default: True

Whether or not to allow uploads (e.g. of Child photos). For some deployments (AWS, Heroku, Nanobox) this setting will default to False due to the lack of available persistent storage.

AWS_ACCESS_KEY_ID

Default: None

Required to access your AWS S3 bucket, should be uniquely generated per bucket for security.

See also: AWS_STORAGE_BUCKET_NAME

AWS_SECRET_ACCESS_KEY

Default: None

Required to access your AWS S3 bucket, should be uniquely generated per bucket for security.

See also: AWS_STORAGE_BUCKET_NAME

AWS_STORAGE_BUCKET_NAME

Default: None

If you would like to use AWS S3 for storage on ephemeral storage platforms like Heroku you will need to create a bucket and add it's name. See django-storages' [Amazon S3 documentation] (http://django-storages.readthedocs.io/en/latest/backends/amazon-S3.html).

DEBUG

Default: False

When in debug mode, Baby Buddy will print much more detailed error information for exceptions. This setting should be False in production deployments.

See also Django's documentation on the DEBUG setting.

NAP_START_MAX

Default: 18:00

The maximum start time (in the instance's time zone) before which a sleep entry is consider a nap. Expects the format %H:%M.

NAP_START_MIN

Default: 06:00

The minimum start time (in the instance's time zone) after which a sleep entry is considered a nap. Expects the format %H:%M.

SECRET_KEY

Default: None

A random, unique string must be set as the "secret key" before Baby Buddy can be deployed and run.

See also Django's documentation on the SECRET_KEY setting.

TIME_ZONE

Default: Etc/UTC

The time zone to use for the instance. See List of tz database time zones for all possible values.

API

Baby Buddy uses the Django REST Framework (DRF) to provide a REST API.

The only requirement for (most) requests is that the Authorization header is set as described in the Authentication section. The one exception is the /api endpoint, which lists all available endpoints.

Currently, the following endpoints are available for GET, OPTIONS, and POST requests:

• /api/children/
• /api/changes/ (Diaper Changes)
• /api/feedings/
• /api/notes/
• /api/sleep/
• /api/timers/
• /api/tummy-times/
• /api/weight/

Authentication

By default, the TokenAuthentication and SessionAuthentication classes are enabled. Session authentication covers local API requests made by the application itself. Token authentication allows external requests to be made.

❗ In a production environment, token authentication should only be used for API calls to an https endpoint. ❗

Each user is automatically assigned an API key that can be used for token authentication. This key can be found on the User Settings page for the logged in the user. To use a key for an API request, set the request Authorization header to Token <user-key>. E.g.

Authorization: Token 2h23807gd72h7hop382p98hd823dw3g665g56

If the Authorization header is not set or the key is not valid, the API will return 403 Forbidden with additional details in the response body.

GET Method

Request

The limit and offset request parameters can be used to limit and offset the results set respectively. For example, the following request will return five diaper changes starting from the 10th diaper change entry:

curl -X GET 'https://[...]/api/changes/?limit=5&offset=10' -H 'Authorization: Token [...]'
{
    "count": <int>,
    "next": "https://[...]/api/changes/?limit=5&offset=15",
    "previous": "https://[...]/api/changes/?limit=5&offset=5",
    "results": [...]
}

Field-based filters for specific endpoints can be found the in the filters field of the OPTIONS response for specific endpoints.

Response

Returns JSON data in the response body in the following format:

{
    "count":<int>,
    "next":<url>,
    "previous":<url>,
    "results":[{...}]
}

• count: Total number of records (in the database, not just the response).
• next: URL for the next set of results.
• previous: URL for the previous set of results.
• results: An array of the results of the request.

OPTIONS Method

Request

All endpoints will respond to an OPTIONS request with detailed information about the endpoint's purpose, parameters, filters, etc.

Response

Returns JSON data in the response body describing the endpoint, available options for POST requests, and available filters for GET requests. The following example describes the /api/children endpoint:

{
    "name": "Child List",
    "renders": [
        "application/json",
        "text/html"
    ],
    "parses": [
        "application/json",
        "application/x-www-form-urlencoded",
        "multipart/form-data"
    ],
    "actions": {
        "POST": {
            "id": {
                "type": "integer",
                "required": false,
                "read_only": true,
                "label": "ID"
            },
            [...]
        }
    },
    "filters": [
        "first_name",
        "last_name",
        "slug"
    ]
}

POST Method

Request

To add new entries for a particular endpoint, send a POST request with the entry data in JSON format in the request body. The Content-Type header for POST request must be set to application/json.

Regular sanity checks will be performed on relevant data. See the OPTIONS response for a particular endpoint for details on required fields and data formats.

Response

Returns JSON data in the response body describing the added/updated instance or error details if errors exist. Errors are keyed by either the field in error or the general string "non_field_errors" (usually when validation incorporates multiple fields).

Development

Requirements

• Python 3.5+, pip, pipenv
• NodeJS 8.x and NPM 5.x
• Gulp

Installation

1. Install pipenv

    pip install pipenv
   

2. Install required Python packages, including dev packages

    pipenv install --dev
   

3. Install Gulp CLI

    npm install -g gulp-cli
   

4. Install required Node packages

    npm install
   

5. Set, at least, the DJANGO_SETTINGS_MODULE environment variable

    export DJANGO_SETTINGS_MODULE=babybuddy.settings.development
   

   This process will differ based on the host OS. The above example is for Linux-based systems. See Configuration for other settings and methods for defining them.

6. Migrate the database

    gulp migrate
   

7. Build assets and run the server

    gulp
   

   This command will also watch for file system changes to rebuild assets and restart the server as needed.

Open http://127.0.0.1:8000 and log in with the default user name and password (admin/admin).

Gulp commands

Baby Buddy's Gulp commands are defined and configured by files in the gulpfile.js folder. Django's management commands are defined in the babybuddy/management/commands folder.

• gulp
• gulp build
• gulp clean
• gulp collectstatic
• gulp coverage
• gulp extras
• gulp fake
• gulp lint
• gulp makemigrations
• gulp migrate
• gulp reset
• gulp runserver
• gulp scripts
• gulp styles
• gulp test

gulp

Executes the build and watch commands and runs Django's development server.

build

Creates all script, style and "extra" assets and places them in the babybuddy/static folder.

clean

Deletes all build folders and the root static folder. Generally this should be run before a gulp build to remove previous build files and the generated static assets.

collectstatic

Executes Django's collectstatic management task. This task relies on files in the babybuddy/static folder, so generally gulp build should be run before this command for production deployments. Gulp also passes along non-overlapping arguments for this command, e.g. --no-input.

coverage

Create a test coverage report. See .coveragerc for default settings information.

extras

Copies "extra" files (fonts, images and server root contents) to the build folder.

fake

Adds some fake data to the database. By default, fake creates one child and 31 days of random data. Use the --children and --days flags to change the default values, e.g. gulp fake --children 5 --days 7 to generate five fake children and seven days of data for each.

lint

Executes Python and SASS linting for all relevant source files.

makemigrations

Executes Django's makemigrations management task. Gulp also passes along non-overlapping arguments for this command.

migrate

Executes Django's migrate management task. In addition to migrating the database, this command creates the default admin user. Gulp also passes along non-overlapping arguments for this command.

reset

Resets the database to a default state with one fake child and 31 days of fake data.

runserver

Executes Django's runserver management task. Gulp also passes along non-overlapping arguments for this command.

scripts

Builds and combines relevant application scripts. Generally this command does not need to be executed independently - see the build command.

styles

Builds and combines SASS styles in to CSS files. Generally this command does not need to be executed independently - see the build command.

test

Executes Baby Buddy's suite of tests.

Gulp also passes along non-overlapping arguments for this command, however individual tests cannot be run with this command. python manage.py test can be used for individual test execution.

Pre-commit Hook

A pre-commit hook is recommended for all commits in order to make sure that static assets are correctly committed. Here is an example working script for bash:

#!/bin/bash

if [ $(git diff --cached --name-only | grep static_src -c) -ne 0 ]
then
    gulp clean && gulp build && gulp collectstatic
    if [ $? -ne 0 ]; then exit $?; fi
    git add ./static
fi

gulp lint && gulp test

exit $?

This script will build and add static assets to the commit only if changes have been made to files in a static_src directory of the project. It will always run the gulp lint and gulp test commands. If any of these processes result in an error, the commit will be rejected.