An inexpensive, open-source system for automated, high-throughput, realtime observation and conditioning experiments.
Most people will not need to use the following instructions. They are provided for anyone who wants to set up the ATLeS box software environment themselves. The easiest way to install and use the software is to install a provided OS image. The following installation steps require that you are comfortable working in a Linux development environment.
A common setup will install atles_web.py on a server (it has been tested
under Linux and Cygwin thus far), while atles_remote.py will run on one or
more Raspberry Pi nodes in the experiment boxes. When installing on a
Raspberry Pi, start with a clean install of
Raspbian Lite.
The following instructions cover installing the software on a Raspberry Pi, and they were developed and tested on a clean install of Raspbian Stretch Lite (2018-11-13).
Before installing the software, install the required tools and libraries (many may already be installed depending on your base system):
sudo apt install build-essential pkg-config python3-dev python3-pip git rsync
Clone the ATLeS repository:
git clone https://github.com/liffiton/ATLeS.git
The ATLeS software will now be in a folder named ATLeS. The following
commands assume you have changed into that directory.
The software that runs on each box to control experiments and connects back to the server depends on several Python packages that can be installed via apt and pip.
First install numpy and smbus (for reading sensor values) with apt:
sudo apt install python3-numpy python3-smbus
OpenCV can be installed via pip with dependencies installed first with apt:
sudo apt install libatlas3-base libwebp6 libtiff5 libjasper1 libilmbase12 libopenexr22 libilmbase12 libgstreamer1.0-0 libavcodec57 libavformat57 libavutil55 libswscale4 libgtk-3-0 libpangocairo-1.0-0 libpango-1.0-0 libatk1.0-0 libcairo-gobject2 libcairo2 libgdk-pixbuf2.0-0
sudo pip3 install opencv-python-headless
Install the remaining python packages by running:
sudo pip3 install -r src/requirements_remote.txt
We install using sudo here so that the libraries are available systemwide, as
atles_remote.py will be run as root (to be able to control the PWM output).
The server software is normally run on a system separate from the Raspberry
Pi(s) (see setup). However, it can be installed on a Raspberry Pi
alongside the atles-remote.py/atles_box.py programs for testing or
prototyping purposes.
Install the Python packages required for atles_web.py by running:
pip3 install --user -r src/requirements_web.txt
The camera, SSH, and i2c need to be enabled using raspi-config. Run sudo
raspi-config and enable all three under “Interfacing Options.”
To use the camera via v4l, the bcm2835-v4l2 module must be loaded. For the
i2c interface (for sensors), the i2c-dev module must be loaded. To have the
modules loaded automatically on boot, add each to /etc/modules, one per line:
bcm2835-v4l2
i2c-dev
Then reboot the Pi.
Alternatively, modprobe can be used to load the modules manually at any time:
sudo modprobe bcm2835-v4l2
sudo modprobe i2c-dev