The Python Stellar Evolution Module¶
A module for interpolating among existing stellar evolution tracks. Since
generating the interpolation is quite slow, POET supports “serializing” (storing
as a file on disk) a computed interpolation. The
stellar_evolution.manager sub-module keeps track of exsiting
interpolations, reusing them whenever appropriate, so this is the suggested
interface.
Basic usage:¶
The user requests an interpolator from the manager by specifying either the
interpolator name (using the
stellar_evolution.manager.StellarEvolutionManager.get_interpolator_by_name()
method) or by specifying a configuration for the interpolator. In the former
case, it is an error to request an intorpolator with a name not previously
defined (POET comes with a pre-defined interpolator named default). In the
latter case, if the user specifies a name to assign to the new interpolator, the
manager will automatically generate an interpolator if one does not exist. For
example:
from poet.stellar_evolution.manager import StellarEvolutionManager
import numpy
%Create a manager
manager = StellarEvolutionManager(<path to serialized interpolators>)
%Find and use the default interpolator
interpolator = manager.get_interpolator_by_name('default')
% The convective zone moment of inertia for a 0.73 solar mass star with
% [Fe/H] = -0.13
iconv = interpolator('ICONV', 0.73, -0.13)
% Evaluate the moment of inertia at a list of ages.
iconv_values = iconv(numpy.linspace(0.05, 10.0, 100))
Custom interpolation using the package tracks:¶
If for some reason the standard interpolation is not optimal for your purposes, it is possible to modify the parameters for how each quantity’s age interpolation is performed (the subsequent interpolation in mass and metallicity is always done using a bi-cubic spline):
% construct an interpolator using only a subset of the tracks, as well
% as custom smoothing and number of nodes.
interpolator = manager.get_interpolator(
nodes = dict(RADIUS = <int>, %number of nodes for the radius age
%interpolation
ICONV = <int>, %convective moment of inertia nodes
LUM = <int>, %luminosity nodes
IRAD = <int>, %radiative moment of inertia nodes
MRAD = <int>, %radiative mass nodes
RRAD = <int>), %tachocline radius nodes
smoothing = dict(RADIUS = <float>, %smoothing parameter for the
%radius age interpolation
ICONV = <float>, %convective moment of inertia
%smoothig
LUM = <float>, %luminosity smoothing
IRAD = <float>, %radiative moment of inertia
%smoothing
MRAD = <float>, %radiative mass smoothing
RRAD = <float>), %tachocline radius smoothing
vs_log_age = dict(RADIUS = <bool>, %should the independent argument
%for radius interpolation be
%log(age) instead of age?
ICONV = <bool>, %Same as RADIUS but for the
%convective moment of inertia.
LUM = <bool>, %same as RADIUS but for luminosity.
IRAD = <float>, %same as RADIUS but for radiative
%moment of inertia.
MRAD = <float>, %same as RADIUS but for mass in the
%radiative core.
RRAD = <float>), %same as RADIUS but for radius in
log_quantity = dict(RADIUS = <bool>, %should the log(radius) be used
%in the interpolation instead of
%radius?
ICONV = <bool>, %Same as RADIUS but for the
%convective moment of inertia.
LUM = <bool>, %same as RADIUS but for
%luminosity.
IRAD = <float>, %same as RADIUS but for radiative
%moment of inertia.
MRAD = <float>, %same as RADIUS but for mass in
%the radiative core.
RRAD = <float>),%same as RADIUS but for radius in
%the radiative core.
track_fnames = [...], %List of filenames containing stellar evolution
%tracks to interpolate among.
masses = [<float>, ...], %list of stellar masses to use
feh = [<float>, ...] %list of stellar [Fe/H] values to use
model_suite = <str>, %The name of a preiously registered
%suite of tracks to use in the
%interpolation. POET comes with a suite
%named 'MESA'.
new_interp_name = <str> %A human-readable name to assign should an
%interpolator matching the other arguments
%does not exist.
num_threads = <int> %The number of threads to use when deriving the
%interpolation.
)
As explained above, all subsequent calls to
stellar_evolution.manager.StellarEvolutionManager.get_interpolator() using
the same parameters, will re-use the interpolator generated by the first call.
Note the new_interp_name argument. If that argument is not supplied, and an
interpolator matching the specified configuration is not found, the result is
None.
Custom stellar evolution tracks¶
In is also possible to construct interpolations based on a set of tracks not packaged with POET. The tracks must be formatted as comma separated values, with the first line containing the names of the columns. All quantities required by POET must be tabulated. See the set of tracks shipped with POET for an example.
The first step is to register the tracks with the interpolation manager. This can be done track by track:
%Register a single stellar evolution track.
manager.register_track(<filename>, %Name of the file contaning the track
mass, %The stellar mass of the track (Msun)
metallicity, %[Fe/H] of the track.
model_suite) %Name of a track collection to assign
%this track to.
If the names of the files can be parsed to extract the mass and metallicity, there is a shortcut:
%Register an entire collection of tracks at once with stellar mass and
%metallicity encoded in the filename.
manager.register_track_collection(
[<filename>,...], %List of track files
<regular expression>, %A compiled python re which defines groups
%names 'MASS' and either 'Z' or '[Fe/H]'
%The package tracks match:
%'M(?P<MASS>[0-9.E+-]+)_Z(?P<Z>[0-9.E+-]+).csv'
model_suite %Name of a track collection to assign the
%tracks to
)
After registering the tracks, new interpolators can be generated and registered with the manager, either like we did above using the custom tracks, but this time we specify our new model suite:
% construct an interpolator using only a subset of the tracks, as well
% as custom smoothing and number of nodes.
interpolator = manager.get_interpolator(
masses = [<float>, ...], %list of stellar masses to use
metallicities = [<float>, ...], %list of stellar metallicities to use
model_suite = <str>, %the model suite defined above
nodes = dict(RADIUS = <int>, %number of nodes for the radius age
%interpolation
ICONV = <int>, %convective moment of inertia nodes
LUM = <int>, %luminosity nodes
IRAD = <int>, %radiative moment of inertia nodes
MRAD = <int>, %radiative mass nodes
RRAD = <int>), %tachocline radius nodes
smoothing = dict(RADIUS = <float>, %smoothing parameter for the
%radius age interpolation
ICONV = <float>, %convective moment of inertia
%smoothig
LUM = <float>, %luminosity smoothing
IRAD = <float>, %radiative moment of inertia
%smoothing
MRAD = <float>, %radiative mass smoothing
RRAD = <float>), %tachocline radius smoothing
new_interp_name = <str> %A human-readable name to assign should an
%interpolator matching the other arguments
%not exist.
)
This only works if the new suite has only one track per mass-metallicity
combination. If for example the new tracks again use model_suite = MESA,
then which tracks to use must be specified by name:
% construct an interpolator using a custom set of tracks, as well
% as custom smoothing and number of nodes.
interpolator = manager.get_interpolator(
track_fnames = [<path>, ...], %list of filenames containing
%the stellar evolution tracks
nodes = <same as above>
smoothing = <same as above>
new_interp_name = <same as above>
)