Drink-O-Matic API

An API written in rust to turn some pumps, relays, and a Raspberry PI into an automatic personal cocktail bar.

About The Project

Ingredients go here

Your cup goes under there

I was watching some YouTube videos about cool things to do with embedded systems when I came across something called a "barbot". It was an arduino powered cocktail bar which inspired me to try and make something similiar at a cost that wouldn't break the bank. Thus the "Drink-O-Matic" was born. I took to Aliexpress to find some parts, slapped them together in a wooden box I found from a craft store and then wrote this for my Raspberry PI. I'm not going to explain how I put everything together here but you can find a bill of materials below.

Bill of Materials

• 1 Raspberry PI | price varies but mine was $37.07
• 6 food grade 385 diaphragm pumps | $32.16
• 1 six channel 5v trigger relay module \w octo-coupler output | $7.36
• 10 meters of 6x9 mm transparent food grade silicone tubing | $17.85
• 1 10A 12V power supply | $15.53
• 6 1UF capacitors | $2.05
• Zip ties for style points | $3.94
• Wooden box to stuff everything in | $10.32

My Total Without Shipping: $126.28 (Not bad eh!?)

In essence, this queues up jobs for each pump and runs them serially in the background. It optionally stores settings for my user interface if you enable the "bff" feature because I was too lazy too create yet another repo for a back-end for front-end layer. I'm also not the best with electrical engineering so it may not be necessary to limit it to 1 active pump at a time but I didn't want to chance it and burn my house down 🤣.

Getting Started

Prerequisites

This utilizes GPIO character device ABI, ensure your Linux kernel is version >= v4.4

Usage

1. Create a folder named ".drink-o-matic" in your user home directory (locations by OS here)
2. Copy the example dotenv file to the folder created in the above step and rename it to ".env"
   1. My relay was inverted so make double sure you set IS_RELAY_INVERTED to 0 if yours isn't or you'll have a wet floor when it turns on
   2. Additionally spend some time tweaking the MILLISECONDS_PER_ML variable once everything is setup to ensure measurements are accurate
3. Copy the strings xml file to the folder created in step 1
4. Update the ".env" to support your current configuration
5. If desired, set the address in the rocket toml file to "0.0.0.0" so that other machines on your network can access the API
6. Run the following to start the API

cargo run -r

Note: If you are running this for use with the frontend I made (umbreon222/drink-o-matic-web-interface), use this command instead to enable the back-end for front-end feature:

cargo run -r --features bff

Development Note

I built most of this from my Windows PC which obviously doesn't support the GPIO character device ABI. For situations like this, you can build/run with the --no-default-features flag. This turns off the dependency on gpio-cdev and mocks it instead which is useful for debugging and testing.

An example from when I was testing the UI:

cargo build --no-default-features --features bff

License

Distributed under the MIT License. See LICENSE.txt for more information. (back to top)