AraProc is a python-based analysis framework for doing ARA data analysis.
To get started, see the example in the examples folder.
You are strongly encouraged to install AraProc into a python virtual environment built on an already existing ARA software stack. The easiest way to do this is to grab the ARA software stack from cvmfs. You can also build your own via the cvmfs scripts.
You are also strongly encouraged to install it in "editable" mode so that you can easily make changes to the analysis classes, which you very likely want to do.
The instructions below show you the steps to the recommended installation method.
# activate the python in cvmfs
source /cvmfs/ara.opensciencegrid.org/trunk/RHEL_7_x86_64/setup.sh
# build a virtual environment
python3 -m venv /path/to/araproc_venv
# then activate your new virtual environment
source /path/to/araproc_venv/bin/activate
# then get, and install, AraProc
(araproc_venv) git clone git@github.com:clark2668/AraProc.git
(araproc_venv) cd AraProc
(araproc_venv) pip install -e .
To do work in the future, you will always need to start up your environment, and then your virtual env. So you will always start by doing:
source /cvmfs/ara.opensciencegrid.org/trunk/RHEL_7_x86_64/setup.sh
source /path/to/araproc_venv/bin/activateThe former is necessary so that you get all the other ARA depenencies (e.g. ROOT, GSL, etc.). The latter actually turns on your venv.
AraProc dependsd heavily on the entire ARA software stack, including AraRoot and AraSim. This means that any dependency of AraRoot and AraSim is a dependency of AraProc. See the AraRoot and the AraSim documenetation. As a result, you are strongly encouraged to either use the ARA cvmfs installation, or to you use the ARA cvmfs builder scripts in a local directory, in order to get AraProc installed. See above for more words on installation.
Because AraProc has such strong dependencies on AraRoot and AraSim, the first thing the module does upon initialization
is check that the correct library files are available. These are the lines found in araproc.__init__.py.
By default it looks for AraRoot and AraSim in $ARA_UTIL_INSTALL_DIR, and for FFTTools in $ARA_DEPS_INSTALL_DIR.
If you have installed AraRoot, AraSim, etc. into non-standard locations, then you will need to update those paths.
Please do not ignore error messages about being unable to find those libraries!!
One issue we encountered early were memory leaks. In particular, AraRoot and FFTTools like to return pointers to ROOT objects. E.g. the AraRoot getGraphFromRFChan function. This is dangerous, because pyroot doesn't know how to correctly cleanup pointers; see for example, this article about this on the ROOT forums. The solution is that you need to explicitly give python control of the objects memory when it is created. For example:
wave = usefulPtr.getGraphFromRFChan(0)
ROOT.SetOwnership(wave, True)