💿Linux

Installing Bluetang.AI on Linux

This installation guide was tested on Antergos (Arch-based) but should work on all other distros with minimal changes.

System Requirements

• X11-Based Desktop Environment required (Wayland not supported)

• Developed and tested on Cinnamon DE

• xwininfo command required

For Arch-based distros, install xwininfo:

sudo pacman -S xorg-xwininfo

Python Environment Setup

Python Version Requirements

Bluetang.AI requires Python 3.6 or higher.

While you can use your system Python if it meets the version requirement, we recommend using pyenv with a virtual environment.

Install Python 3.6

pyenv install 3.6.4

For Debian-based systems, if the build fails:

sudo apt-get install libexpat1-dev libssl-dev zlib1g-dev libncurses5-dev libbz2-dev liblzma-dev libsqlite3-dev libffi-dev tcl-dev libgdbm-dev libreadline-dev tk tk-dev openssl

Create Virtual Environment

pyenv virtualenv 3.6.4 serpent
mkdir BluetangAI && cd BluetangAI 
pyenv local bluetang

Dependencies

Redis Setup

Redis (minimum version 3.0.0) is used for frame buffer storage and analytics events.

Installation options:

• Package manager (Arch: redis, Debian: redis-server)

• Docker container

• Source installation

Native installations are recommended over Docker for better performance.

Verify installation:

redis-cli

Installing BluetangAI

1. Navigate to your BluetangAI directory

2. Run:

pip install BluetangAI
serpent setup

Optional Modules

OCR Module

Requirements

Install Tesseract:

• Arch: tesseract tesseract-data-eng

• Debian: tesseract-ocr tesseract-ocr-eng

For Debian, also install:

sudo apt-get install libtesseract-dev libleptonica-dev

Installation

serpent setup ocr

GUI Module

Requirements

Requires Kivy framework. See Kivy Linux installation guide for setup.

Installation

serpent setup gui

Test with:

serpent visual_debugger

Machine Learning Module

GPU Requirements

• NVIDIA GPU with CUDA 3.0+ support (GTX 600 series and up)

• NVIDIA proprietary drivers

• CUDA toolkit

• cuDNN

Arch Installation

sudo pacman -S nvidia nvidia-installer cuda cudnn

Debian Installation

Drivers:

sudo apt-get purge nvidia-*
sudo apt-add-repository ppa:graphics-drivers/ppa
sudo apt-get update
sudo apt-get install nvidia-384

CUDA:

sudo apt-get install nvidia-cuda-dev nvidia-cuda-toolkit nvidia-nsight

cuDNN setup detailed in documentation.

Installation

serpent setup ml

For distro-specific tips or improvements, contributions to this guide are welcome!