Neale Gibson ([email protected], [email protected])
PlanetFuncs python module - collection of python functions for exoplanet calculations, written in C for fast implementation where necessary.
This code is a streamlined and updated version of MyFuncs containing only planet funcs, and removed inference functions which are now in inferno. I have also simplified installation from depending on the libhyp hack into scipy code - now all the relevant code is contained in the TransitFlux submodule.
While there is support for computing RV curves, eccentric orbits etc, I don't use these regularly and they are neither well tested nor optimised.
This module first needs the GSL scientific C library installed.
The simple version is to install via conda
$ conda install gsl
Then the module can be installed as normal
$ python setup.py build
$ python setup.py install
or
$ python setup.py install --prefix=<INSTALL_DIR>
This was a little painful and may or may not work in general...
Microsoft Visual C++ 14.0 or greater is required to get the C compilers Get it with "Microsoft C++ Build Tools": https://visualstudio.microsoft.com/visual-cpp-build-tools/
- Choose "Desktop Development with C++" and install
First open an Anaconda Powershell (which has conda initiated by default)
> conda init powershell # within anaconda prompt - initalises conda in standard shell
Then I can open a windows powershell, via the Developer Command Prompt installed via VSCode. This depends on the exact version installed, but search for "developer command" in the search box and you should find it and open. Anaconda should be initialised now (indicated by the environment listed in the command prompt). In addition, this shell will have all of the associated C libraries linked.
You also need to point to the correct GSL libraries. Within the powershell run (including the initial dollar signs!):
> $env:INCLUDE="$env:CONDA_PREFIX\Library\include"
> $env:LIB="$env:CONDA_PREFIX\Library\lib"
> $env:LIBPATH="$env:CONDA_PREFIX\Library\lib"
Finally, you should be able to build and install the package
pip install .
# or
python -m pip install .
Note, to create a wheel distribution, run:
python -m build
Alternatively you can install GSL from source, and make sure it's available on your CPATH and LIBRARY_PATH. It's quite easy, in principle could speed up the code if it's better tuned to your machine.
You can either download the latest version of GSL from http://www.gnu.org/software/gsl/ or use one of the versions included in this distribution.
It is easy to install, just download, unpack, enter directory:
$ tar -xzvf gsl-1.16.tar.gz
$ cd gsl-1.16
Then run configure
$ ./configure
#or first configure to a different directory? - see gsl install instructions
$ ./configure --prefix=<INSTALL_DIR>
#or for your home directory
$ ./configure --prefix=$HOME
Then need to build and install with make
$ make
$ make install
Finally, to complete the installation your C compilers need to know where to look for the headers and code. Depending on the path, they may already be automatically found. If in doubt, you'll need to add to your CPATH and LIBRARY_PATH:
#e.g. in bash:
$ export CPATH="${HOME}/include:$CPATH"
$ export LIBRARY_PATH="${HOME}/lib:$LIBRARY_PATH"
#or for a non-standard location:
$ export CPATH="<INSTALL_DIR>/include:$CPATH"
$ export LIBRARY_PATH="<INSTALL_DIR>/lib:$LIBRARY_PATH"
Then you can run python setup.py install as above.
from PlanetFuncs import transit_quad, transit_nonlin
import matplotlib.pyplot as plt
#define time array (units just need to be consistent with T0 and P)
t = np.linspace(-0.1,0.1,1000)
#define parameter vector [T0,P,a/Rs,Rp/Rs,b,limb_darkening...,f_oot,T_grad]
p_quad = [0,3,10,0.1,0.,0.1,0.1,1.,0]
p_nonlin = [0,3,10,0.1,0.,0.1,0.1,0.1,0.1,1.,0]
flux_quad = transit_quad(p_quad,t)
flux_nonlin = transit_nonlin(p_nonlin,t)
ax = plt.subplot(xlabel='time',ylabel='rel flux')
ax.plot(t,flux_quad,'r-')
ax.plot(t,flux_nonlin,'k-')