Rabacus¶

Installing prerequisites with pip¶

A simple way to install Python packages is using the package manager pip. To check if you have pip installed on your system, type pip at the command line,

pip

If this produces usage instructions then congratulations, you have pip installed. If not, instructions for installing pip can be found here. To check if the python packages are installed on your system, attempt to import them from the python command prompt,

>>> import scipy
>>> import numpy as np
>>> import quantities as pq

If any of these import commands produce an error message you will need to install the proper software before installing Rabacus. Once you have access to pip, you can install any missing prerequisites using the following commands,

sudo pip install scipy
sudo pip install numpy
sudo pip install quantities

If you do not have root access on your system you can pass the --user flag which will install the packages into a hidden folder called .local in your home directory,

pip install --user scipy
pip install --user numpy
pip install --user quantities

Installing prerequisites on Debian (Ubuntu)¶

On Debian based systems (such as Ubuntu) you may prefer to install these prerequisites using the APT tool,

sudo apt-get install python-scipy python-numpy python-quantities

To increase the speed of execution, much of Rabacus is written in Fortran 90 and then wrapped using the f2py tool that is part of numpy. For the installation to be successful, a fortran compiler must be in your executable path. If you don’t already have one, I recommend the gnu fortran compiler gfortran. On Debian based systems (such as Ubuntu) you can install this compiler using the APT tool,

sudo apt-get install gfortran

Setting F90 environment variable¶

Rabacus makes use of OpenMP directives in the Fortran code base and so we have to make sure the code is compiled correctly. In order to do this, you have to let the build system know what Fortran 90 compiler you are going to be using. The simplest way to do this is to set the environment variable F90 before following the installation instructions below. Rabacus has been tested with the intel compiler and the gnu gfortran compiler. For other compilers you will have to follow the Manual Install instructions below.

To use the gfortran compiler, type the following at the command line (in Bash)

export F90=gfortran

To use the intel compiler set

export F90=ifort

Single command install¶

If you have made the appropriate sacrifices to the computer gods, you should be able to install an OpenMP enabled version of Rabacus with a single comand line call to pip,

sudo -E pip install rabacus

Note that the -E flag preserves the environment variables, namely the F90 environment variable that needs to be set. If this causes problems you can try setting the F90 environment variable as the root user before giving the pip command. As was the case for the prerequisites, if you do not have root access on your system you can pass the --user flag which will install Rabacus into a hidden folder called .local in your home directory,

pip install --user rabacus

If the last two lines printed to the screen are,

Successfully installed rabacus
Cleaning up...

then congratulations you have a working copy of Rabacus. To double check, begin an ipython session and attempt an import,

import rabacus as ra

Packages installed with pip can be uninstalled in the same way,

pip uninstall rabacus

Manual install¶

If the above process fails for any reason we can always download Rabacus and manually invoke the setup script. The first step is to download and untar the Rabacus tar.gz file from the PyPI site (https://pypi.python.org/pypi/rabacus) and change into the main Rabacus directory,

gunzip rabacus-x.x.x.tar.gz
tar xvf rabacus-x.x.x.tar
cd rabacus-x.x.x

Now we have direct access to the setup.py file which gives us a lot more freedom but it comes at the cost of slightly more complexity. First it’s a good idea to see which fortran compilers are detected on your machine. The following command will list all of the fortan compilers found on your system and all the compilers available for your system but not found.

f2py -c --help-fcompiler

For example on my machine I get the following,

Fortran compilers found:
  --fcompiler=gnu95    GNU Fortran 95 compiler (4.8.1-10)
  --fcompiler=intelem  Intel Fortran Compiler for 64-bit apps (14.0.2.144)
Compilers available for this platform, but not found:
  --fcompiler=absoft   Absoft Corp Fortran Compiler
  --fcompiler=compaq   Compaq Fortran Compiler
  --fcompiler=g95      G95 Fortran Compiler
  --fcompiler=gnu      GNU Fortran 77 compiler
  --fcompiler=intel    Intel Fortran Compiler for 32-bit apps
  --fcompiler=intele   Intel Fortran Compiler for Itanium apps
  --fcompiler=lahey    Lahey/Fujitsu Fortran 95 Compiler
  --fcompiler=nag      NAGWare Fortran 95 Compiler
  --fcompiler=pathf95  PathScale Fortran Compiler
  --fcompiler=pg       Portland Group Fortran Compiler
  --fcompiler=vast     Pacific-Sierra Research Fortran 90 Compiler

Now we decide which of the fortran compilers to use and which flags to pass the build command. Suppose you wanted to use the Intel compiler. Edit the setup.py file such that the variable f90_flags is a list of compile flags and omp_lib is a list containing the linking flags. For example,

f90_flags = ["-openmp", "-fPIC", "-xHost", "-O3", "-ipo",
             "-funroll-loops", "-heap-arrays", "-mcmodel=medium"]

omp_lib = ["-liomp5"]

These variables are already defined near the top of the setup.py file and will need to be overwritten. Once this is done, we give the build command to the setup.py script,

python setup.py build --fcompiler=intelem

After the package is built, give the install command to actually install it,

sudo python setup.py install --record rabacus_install_files.txt

The last part of the command is to allow for easy uninstall. This process just involves deleting all installed files which will be listed in the file rabacus_install_files.txt. This can be accomplished using the following command,

cat rabacus_install_files.txt | xargs sudo rm -rf

The install can also be done locally for those without root permission on their system by passing the --user flag to the install command,

python setup.py install --user --record rabacus_install_files.txt

Note that if you previously did an install of Rabacus that required the sudo command you will likely need to delete the rabacus.egg-info directory and some directories inside the build directory as they will need to be modified but will be owned by root. If you are only doing a local install then this shoudn’t be necessary. This procedure should work for any fortran compiler supported by f2py (i.e. any compiler in the list returned when using the --help-fcompiler flag.

Testing install¶

Detailed examples of using rabacus are available by following the link to the users guide below. However, we present a short example with the expected output below as a way to quickly test that a new installation has basic functionality. We first import rabacus and then create an object that gives access to the meta galactic radiation background described in Haardt & Madau 2012. Finally, we ask for the photo-heating rate of He I at a redshift of 3.0.

import rabacus as ra
hm12 = ra.HM12_Photorates_Table()
z = 3.0
print hm12.He1h(z)

The expected output from a working rabacus installation is given below. Note that there may be differences in the last significant figure due to different processor architectures.

3.39163517433e-12 eV/s

Users guide¶

A detailed users guide can be found by following the link below. It includes many example use cases,

API documentation¶

Automatically generated API documentation for all packages, modules, and classes in Rabacus can be found by following the link below,

Indices and tables¶

• Index
• Module Index
• Search Page

Author¶

Rabacus was written by Gabriel Altay and any questions can be directed to gabriel.altay@gmail.com