Azure beers 🍻

Requirements

1. Amd64 Linux server with Docker installed (we recommend Debian or RHEL based system). In our example we use Ubuntu 20.04 on Azure (we used B1s for 7.39€/month).
2. You will need some kind of IP camera that is visible to your server (you can use public IP, Cloudflare Tunnels or VPN). We have used ESP32-CAM with this code.
3. Optionally you can use a domain or if you use Azure you can use their domain.
4. Optionally you can use some kind of reverse proxy (we used Nginx) and put SSL or password over it.

Production

1. install docker

2. docker run -p 3000:3000 -e CAM_IP=https://example.com/beers.jpg -e LIMIT=0.5 --restart always -d --memory 500m --memory-swap 500m czm1k3/azurebeers

   or

   docker run -p 3000:3000 -e CAM_IP=https://example.com/beers.jpg -e LIMIT=0.5 --restart always -d --memory 500m --memory-swap 500m docker.pkg.github.com/czm1k3/azurebeers/beers

Build:

1. docker build . -t azurebeers

2. docker run -p 3000:3000 -e CAM_IP=https://example.com/beers.jpg -e LIMIT=0.5 azurebeers

Development:

1. Install C++ development kit (required by TensorFlowJS)

   Debian/Ubuntu:

   sudo apt install build-essential

   Arch/Manjaro

   sudo pacman -S base-devel

2. Rename .env.local.example to .env.local and insert your values

3. Install dependencies

   yarn install

4. Run dev

   yarn dev

5. Open http://localhost:3000
   
    

About:

How it all began?

The story of starting this project isn't something exciting, we have no fairy tale as the background of doing this particular thing. We all love to learn and discuss new technologies and there is no better drink like a beer for these long talks over all the stuff that is happening in tech. So this project is a kind of mixture combining both tech and beer. For Us, this project was a great but also fun way to learn how to use all the Cloud and Machine Learning tools, which may be useful knowledge for our future projects.

What is the goal?

That is a great question! It may look like it's just a weird mistake, a project that has no sense in making and developing. We don't live in illusion, this won't be the next Nobel prize thing... It's more like a maker project.

The reason for our effort is, that AI is most of the time shown as something complex, something that is meant only for nerds and experts. We want to change that view, new technologies, which AI definitely is, are super exciting. But applications like self-driving cars or data analysis are for most people too distant topics. We believe that AI is excellent to be used in highly complex cases as well as in some small and fun hobby projects.

With AzureBeers, we want to show off that AI isn't a thing for a few chosen, it's a tool that could be used for a wide variety of purposes, that anyone could benefit out of it and Our ultimate goal would be if it will help anyone, even a single person, to get more used to working with AI and get comfortable about this another industrial revolution tool.