Installation pylabrad Kubuntu 12.04

Install and Make Virtual Environment

In general, virtualenv and virtualenvwrapper python packages provide a good way to maintain and version-control all the python libraries. The basic idea is that one creates a virtual environment where libraries are installed and maintained. In the virtual environment, libraries don't auto-update and are isolated from the system libraries. This is particularly important in Kubuntu given the system's use of PyQt.This reference provides a good guide to using virtual environments with the virtualenvwrapper class, simplifying the usage.

Kubuntu 12.04 comes with Python 2.7.3 we will use it to create virtual environment. First, install pip:

sudo apt-get install python-pip python-dev build-essential

Then update pip and install virtual environment:

sudo pip install --upgrade pip
sudo pip install virtualenv virtualenvwrapper

Start using the virtual environment. To make permanent add this command to /etc/bash.bashrc or ~/.bashrc

source virtualenvwrapper.sh

Then make a new virtual env

mkvirtualenv labrad

Install pylabrad

There are three ways to install pylabrad described in the subsections below.

Pull the pylabrad packages from GIT (best)

first get twisted

pip install twisted

then get labrad

the link is here https://github.com/martinisgroup/pylabrad , clone the repository then cd in the folder type:

python setup.py install

Through PIP

This is the easiest way but may not provide the latest bug-fixes.

workon labrad
pip install twisted
pip install pylabrad

SVN Checkout

Gets the latest source code

workon labrad
pip install twisted
cd ~/Downloads
sudo apt-get install subversion -y
svn co https://svn.code.sf.net/p/pylabrad/svn/trunk pylabrad
cd pylabrad
python setup.py install

Additional Packages

In addition to the basic pylabrad installation, the user may want to install scientific processing or graphical python libraries, or drivers. Keep in mind these have to be installed in the virtual environment.

numpy

pip install numpy

matplotlib

sudo apt-get install libpng12-dev libjpeg8-dev libfreetype6-dev
pip install matplotlib

alternatively  aptitude may be better (this had better luck recently)
sudo apt-get install python-matplotlib

configure to use Qt4 backend

Edit configuration file found with

workon labrad
python
import matplotlib
print matplotlib.matplotlib_fname()

to state:

backend      : Qt4Agg

scipy

sudo apt-get install libatlas-base-dev gfortran
pip install scipy

or aptitude

sudo apt-get install python-scipy

pyserial

pip install pyserial

miscellaneous pip packages

pip install bitarray ipython psutil treedict

qutip

pip install cython

Download the latest qutip installation and run:

python setup.py install

treedict

pip install treedict

treedict can also be installed with

sudo easy_install treedict

PyQt

PyQt is needed for development and execution of graphical applications. To install, one needs to first build SIP. Download SIP 4.14.7 (at the time of writing this) from http://www.riverbankcomputing.com/software/sip/download.

To install SIP:

sudo apt-get install g++ byacc flex
workon labrad
python configure.py
make
sudo make install

Download the latest PyQt version (4.10.2 at the time of writing this) from http://www.riverbankcomputing.co.uk/software/pyqt/download

To install PyQt:

sudo apt-get install qt4-dev-tools libxext-dev libxext6 qt4-designer
workon labrad
python configure.py
make
sudo make install

• If you see an error about sipconfig, make sure that sipconfig.py in site-packages has the right ownership.

Opal Kelly API

First, copy and extract the FrontPanelUSB-FC15-64-4.0.3.tgz from the CD that came with the FPGA (use the 32 bit version if appropriate):

tar -xvzf FrontPanelUSB-FC15-64-4.0.3.tgz

Now we will follow a Ubuntu-adapted version of install.sh:

cp ./60-opalkelly.rules /etc/udev/rules.d
cp API/libokFrontPanel.so /usr/local/lib/
cp -r API/okFrontPanelDLL.* /usr/local/include/

Then reboot and copy the files in API/Python/ to the ok subfolder of the site-packages folder of the virtual environment.

Desktop Shortcuts

It is convenient to make a desktop launch for example, to launch the the node server. To do that, first make a bash script startNodeServer.sh as follows:

#!/bin/bash
source virtualenvwrapper.sh
workon 'labrad'
twistd -n labradnode
exit 0

and then make a desktop launcher with the command. The -i option is necessary for the script to be aware of the standard environmental variables.

bash -i /some_path/startNodeServer.sh