PancakeSoftware / openHabAI

Train Neuronal networks to automate your home

OpenHabAI

Automate your home using Neuronal networks.
OpenHabAI provides a fast c++ backend (mxnet is used for computation) and frontend that run in browser.

To see api documentation look at: catflow/README.md

Install

Download install-packages from artifacts. Extract it and install the .deb package.

cd build/pack
dpkg --install OpenHabAI-0.0.0-Linux.deb
# resolve deps 
apt-get install -f                          

Build from Source

First install these packages:

• For frontend
  • nodejs
  • npm
• For trainServer
  • zlib1g-dev
  • libssl-dev
  • for mxnet
    • libopenblas-dev
    • liblapack-dev
    • cuda (optional, if you want to use gpu) see at mxnet.io
  • cmake, git, c++ build tools

Execute build command:

mkdir build
cd build
cmake ../
make 

The compiled trainSever executable can be found in build/bin.
To run frontend: make frontendRun or see in README of frontend

Development

To use the Websocket Api see backend-frontend protocol definition.

Report Bugs and Improvements

If you found a bug or have a good idea for new a feature just open a new issue at gitlab.