On a linux machine with "nothing" installed, the following commands should get you to be able to run the benchmark. Explanations and alternate ways to install are detailed below. This benchmark should take around 2 minutes to run.
cd ~
git clone https://github.com/teuben/QAC
# phase 1: install CASA (skip this if "casa" is already in your path)
cd QAC/casa
./install_casa
source casa_start.sh
cd ..
# phase 2: prepare install and run benchmark
make data
make install
cd ~/.casa
ln -s ~/QAC
cd ~/QAC/test
ln -s ../data/M100_TP_CO_cube.bl.image
ln -s ../data/M100_aver_12.ms
ln -s ../data/M100_aver_7.ms
make bench
You will need to install CASA before you can use QAC. See https://casa.nrao.edu/casa_obtaining.shtml
An example is in the casa/install_casa shell script.
NOTE: CASA5 and below use init.py and prelude.py, CASA6 uses config.py and startup.py.
Although you can manually execfile() the appropriate python file(s), this can become cumbersome, and we now encourage you to use CASA's ~/.casa/init.py method. With this you will automatically have anything personal available within CASA each time you use CASA.
Within QAC we have made this one step easier, by creating a casa.init.py file that you can directly load.
For example, the following shell commands would grab QAC from github, and add an execfile() command to your ~/.casa/init.py so this is loaded into CASA each time you start CASA.
mkdir -p ~/.casa
cd ~/.casa
git clone https://github.com/teuben/QAC
cd QAC
make install
make tp2vis
The next time CASA starts, you will see the QAC routines are loaded, as well have access to the tp2vis routines.
You can look at QAC's casa.init.py file and steal some ideas how to load other interesting CASA add-ons.
An alternative to loading TP2VIS via QAC is loading it manually:
execfile("tp2vis.py")
this is the recommending procedure in the public release of tp2vis.
In the current CASA environment there are two other ways QAC could be installed:
- the "buildmytasks" method (supposedly deprecated?)
- python's "setup.py" method (e.g. ADMIT uses this)
but none are supported in this development version of QAC. They might be in some future release, depending on further development of CASA. As it stands now, using the execfile() style to load the correct files is the more pragmatic approach, well realizing this will have to be replaced when CASA switches to Python 3.
Most scripts that use QAC start with a pdir= that is created, but under certain circumstances has expensive datasets that could be re-used. For these it would be useful to be re-entrant, like the dc2019 "script4paper" examples.
E.g. pdir='"sky4a"' sdir='"sky4"' sdata='["sky4.aca.cycle6.ms","sky4.alma.cycle6.1.ms","sky4.alma.cycle6.4.ms"]'