A Python package for calculations with the MRP parameterisation of the Halo Mass Function.
See Murray, Robotham, Power 2016 (in prep.) for more details on what the MRP is.
Required packages are numpy, scipy, mpmath (for incomplete gamma functions), and cached_property. These should be automatically installed when installing mrpy.
To use the MCMC fitting features, emcee and pystan are needed. These are not installed automatically.
The simplest way to install is
pip install mrpy. This should install the required
To get the bleeding edge, use
pip install git+git://github.com/steven-murray/mrpy.git.
If for some reason
pip is not an option, manually download the github
repo and use
python setup.py install.
There’s a lot of things that you can do with mrpy. What you require will depend on the problem at hand. We recommend looking at some of the examples, and the API itself for how to use the code.
With mrpy you can:
- Calculate basic statistics of the truncated generalised gamma distribution (TGGD) with the TGGD class: mean, mode, variance, skewness, pdf, cdf, generate random variates etc.
- Generate MRP quantities with the MRP class: differential number counts, cumulative number counts, various methods for generating normalisations.
- Generate the MRP-based halo mass function as a function of physical parameters via the mrp_b13 function.
- Fit MRP parameters to data in the form of arbitrary curves with the get_fit_curve function.
- Fit MRP parameters to data in the form of a sample of variates with the SimFit class: simulation data is supported with extra efficiency, simulation suites fitted simultaneously is also supported, arbitrary priors on parameters, log-normal uncertainties on variates supported.
- Calculate analytic hessians, jacobians at any point (including the solution of a fit).
- Use alternate parameterisations of the same form via the reparameterise module.
- Work with a special entirely analytic model to understand the effects of various parameters in the analytic_model module.
There are several examples featured in the
docs/examples directory of the github repository. These can also be found
in the official documentation.
- How to get started
- Fit MRP parameters to model data
- Fit MRP parameters to a suite of simulation data simultaneously
- Explore an entirely analytic model which includes no Poisson error
- Determine relationship of MRP parameters to physical ones
- Use different parameterisations of the MRP and check how well they perform
- Use the MRP to explore the relation between halo mass and stellar mass
- Construct a custom fit for an extension of the MRP: a double-Schechter function.
- API Summary