Installation Instructions

Dependencies

The perception module depends on the Berkeley AutoLab’s autolab_core module, which can be installed using the instructions here. The module primarily wraps OpenCV version >= 2.11 which can be installed using pip.

Furthermore, the perception module optionally depends on Tensorflow and pylibfreenect2 for Convolutional Neural Networks and Kinect2 sensor usage, repectively. Install these according to their website’s instructions if their functionality is required.

Installing our repo using pip will attempt to install these automatically.

Any other dependencies will be installed automatically when perception is installed with pip.

Cloning the Repository

You can clone or download our source code from Github.

$ git clone git@github.com:BerkeleyAutomation/perception.git

Installation

To install perception in your current Python environment, simply change directories into the perception repository and run

$ pip install -e .

or

$ pip install -r requirements.txt

Alternatively, you can run

$ pip install /path/to/perception

to install perception from anywhere.

Testing

To test your installation, run

$ python setup.py test

We highly recommend testing before using the module.

Building Documentation

Building perception’s documentation requires a few extra dependencies – specifically, sphinx and a few plugins.

To install the dependencies required, simply run

$ pip install -r docs_requirements.txt

Then, go to the docs directory and run make with the appropriate target. For example,

$ cd docs/
$ make html

will generate a set of web pages. Any documentation files generated in this manner can be found in docs/build.

Deploying Documentation

To deploy documentation to the Github Pages site for the repository, simply push any changes to the documentation source to master and then run

$ . gh_deploy.sh

from the docs folder. This script will automatically checkout the gh-pages branch, build the documentation from source, and push it to Github.