MIT License
Copyright (c) 2017 Teespring
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
San Francisco has a problem. There is but one stall for the many throngs of people, and there is no true way of telling when it is in use (aside from getting out of your seat and checking!)
Well now, no more. You can visit canigoyet.teespring.com and get a simple status update of its use.
This project consists of the following aspects:
- An S3 Bucket, configured to serve static pages
- AWS Lambda functions linked by AWS API Gateway to act as endpoints
- The Raspberry Pi attached to the stall.
- CloudFront and CloudFlare
The S3 Bucket canigoyet.teespring.com contains multiple files, the most important two of which are index.html and "state".
- index.html -- The file served to the user when they hit https://canigoyet.teespring.com. Pulls in data from "state" and displays it to the user.
- state -- File containing a JSON blob of the state of the stall. Currently only has "State" and "UpdatedDate".
These files are housed within the /static/ directory of this git repository.
Rather than starting yet another Heroku server to run pretty minimal code, we opted to use Serverless.
Serverless is responsible for taking the code in js/handler.js and turning it into a GET endpoint that can be hit to update the "state" file in the S3 bucket. In short, Serverless creates an S3 bucket to house extra code, an API Gateway endpoint, then attaches that API Gateway to the Lambda function via a Lambda event.
This provides two endpoints hosted at https://1r2pwixx9j.execute-api.us-east-1.amazonaws.com:
- /dev/poop/free
- /dev/poop/busy
Free, naturally, updates the state object to be free, while the busy endpoint updates the state object to be busy. A sample state object has been provided for this purpose.
Setup involves performing the following steps:
- Ensure that you have nvm, nodejs, and npm
- Run
nvm installwithin yourgame-of-throneproject directory. - Run
npm install serverless -gto install the Serverless client. - Run
npm installwithin the yourgame-of-throneproject directory. - Copy the
sample.envfile to.envand fill in your AWS credentials. If you don't have these, request them from EE Pod.
That's it! You should be up and running.
Lambda works by hosting your code on a dynamically resizing blob of servers that give you control over only your code. The rest is handled by AWS. The functions defined in serverless.yml are set as handlers, and events definitions tie API Gateway endpoints to these functions.
When these are tied together, requests go to the functions in handler.js, and supply an event, context, and callback. event contains everything about the request, including its headers, body, and type. context contains functions and members that pertain to the current function context. callback is the function that needs to respond to calls to the client (e.g. the thing that gives a 200 response back, along with headers and a body, if you should so desire).
When adding a new function, ensure that you update the serverless.yml to tie a new API Gateway endpoint to it. Refer to the ones that exist for potential options, as well as the Serverless documentation for even more.
Deploying is as simple as calling the deploy script at scripts/deploy.sh, which utilizes your AWS credentials within .env and calls the serverless executable.
The RPi is our way of interacting with the physical world. It's mounted to the bathroom stall door, with one wire attached to a +3.3V GPIO port, and another attached to a GPIO input pin. The wires are affixed with a gap between them, opposite the door lock rod. When the lock rod makes contact, it completes a circuit, letting the RPi know that the door lock mechanism has been engaged.
When the lock rod makes contact, a nodejs application running locally on the RPi detects whether or not the GPIO input pin detects a high voltage. If it does, the nodejs app performs an HTTP GET on the /dev/poop/free endpoint. Otherwise, it performs an HTTP GET on /dev/poop/busy.
This will all need to be performed on the RPi.
The file structure on the local RPi filesystem is as follows:
- nodejs directory:
/usr/local/poop - startup script:
/etc/init.d/start_poop.sh - autorun script:
/etc/rc.local
The RPi is currently pointed to the game-of-throne-raspberry repo, which no longer exists. This repo has been merged into the game-of-throne repo. You'll need to first ensure that this step is done before moving forward.
On the RPi, run npm install. You can then run nodejs index.js to start the application. This will start polling pin 23 for a +3.3V input.
Simply edit index.js and you'll be up and running!
S3 doesn't give the ability to serve statically from a bucket with SSL enabled. The best practice approach for this is to place it behind a Cloudfront endpoint, which can then serve HTTPS based content.
Under normal circumstances, you would simply point your DNS provider at the Cloudfront endpoint and be finished. However, both Cloudfront and CloudFlare are CDNs, and both try to enable caching accordingly. This causes issues in two important ways:
- The
stateobject will become stale - CORs issues become a problem if everything is not served from the same origin.
To fix these issues, Cloudfront was put in front of the S3 bucket with very small TTLs, and the ability to serve the OPTIONS request. Then, a CNAME was added to CloudFlare for canigoyet pointing to the Cloudfront URL, with a Page rule that prevented CloudFlare from performing any caching. From there, all content is served from CloudFlare under https://canigoyet.teespring.com
During the middle of the hack-a-thon that spawned this beauty, Mike Kavouras created a MacOS app that provides an icon in the notification bar of Macs that allows for users to be updated in real time without needing to have a web browser open.
The app provides a bonus mode called Gotta go!. When the restroom is occupied, enabling Gotta go! mode will cause the poo emoji to alternate through the various Teespring branded colors. Once the bathrrom is available again, a desktop notification will show and a sound will play.
