py21cmfast.wrapper.inputs

Input parameter classes.

There are four input parameter/option classes, not all of which are required for any given function. They are SimulationOptions, CosmoParams, AstroParams and AstroOptions. Each of them defines a number of variables, and all of these have default values, to minimize the burden on the user. These defaults are accessed via the _defaults_ class attribute of each class. The available parameters for each are listed in the documentation for each class below.

Attributes

FilterOptions
IntegralMethods
Planck18
logger

Exceptions

InputCrossValidationError

Initialize self. See help(type(self)) for accurate signature.

Classes

AstroOptions	Options for the ionization routines which enable/disable certain modules.
AstroParams	Astrophysical parameters.
CosmoParams	Cosmological parameters.
CosmoTables	Class for storing interpolation tables of cosmological functions (e.g. transfer functions, growth factor).
InputParameters	A class defining a collection of InputStruct instances.
InputStruct	A convenient interface to create a C structure with defaults specified.
MatterOptions	Structure containing options which affect the matter field (ICs, perturbedfield, halos).
SimulationOptions	Structure containing broad simulation options.
Table1D	Class for setting 1D interpolation table.

Functions

between(mn, mx)	Validate that a value is between two values.
check_halomass_range(→ None)	Check that the halo mass range is sensible given the parameters.
choice_field(*[, validator])	Create an attrs.field that is a choice.
choice_transformer(→ int)	Produce a transformer that converts a string to int.
choice_validator(inst, att, val)	Validate that a value is one of the choices.
dex2exp_transformer(x, att)	Convert from dex to exponential space.
field(*[, transformer])	Define an attrs field with a 'transformer' property.
get_logspaced_redshifts(→ tuple[float, Ellipsis])	Compute a sequence of redshifts to evolve over that are log-spaced.
input_param_field(kls)	Create an attrs field that must be an InputStruct.
logtransformer(x, att)	Convert from log to linear space.

Module Contents

py21cmfast.wrapper.inputs.between(mn, mx)

Validate that a value is between two values.

py21cmfast.wrapper.inputs.check_halomass_range(inputs: InputParameters) → None

Check that the halo mass range is sensible given the parameters.

This function checks that the minimum halo mass set by the various resolutions and flags does not have any gaps. We raise an error if there is a gap, and a warning if it is above the turnover mass.

Parameters:

inputs (InputParameters)

Return type:

None

py21cmfast.wrapper.inputs.choice_field(*, validator=None, **kwargs)

Create an attrs.field that is a choice.

py21cmfast.wrapper.inputs.choice_transformer(choice: str, att: attrs.Attribute) → int

Produce a transformer that converts a string to int.

The function must be passed a list of string choices. The resulting int is the index of the choice made.

Parameters:

• choice (str)

• att (attrs.Attribute)

Return type:

int

py21cmfast.wrapper.inputs.choice_validator(inst, att: attrs.Attribute, val)

Validate that a value is one of the choices.

Parameters:

att (attrs.Attribute)

py21cmfast.wrapper.inputs.dex2exp_transformer(x, att: attrs.Attribute)

Convert from dex to exponential space.

Notes

The reason why we convert our sigma parameters (the width of the lognormal distributions) from base-10 to base-e is because the C code works with lognormal distributions in base-e, for the following reasons: 1. If HALO_SCALING_RELATIONS_MEDIAN = False, the mu parameter of the distribution has to be adjusted by a factor of the form exp(-sigma^2 /2), if the distribution and sigma are in base-e, or by a factor of the form 10^{-ln(10)*sigma^2 /2} if the distribution and sigma are in base-10. Using base-e in the C code saves us from multiplying by ln(10). 2. In the case of stellar mass (stellar fraction), there is a multiplicative term of the form exp(-M_turn/M). Numerically, it is more efficient to combine the arguments of this exponent with the normally distributed random variable (to make the distribution log-normal), which is what the C code does. Working with base-10 in the C could would require having an expression like 10^[-M_turn/M /ln(10)], if we wish to combine the arguments in the power-law. Since the C code works with base-e, it saves the need to divide by ln(10).

While the multiplication/division by ln(10) is not a major overhead for one halo, when the box contains billions of halos that are evolved through various redshift snapshots, the work with base-e in the C code saves us from executing billions of unnecessary multiplications/divisions.

Parameters:

att (attrs.Attribute)

py21cmfast.wrapper.inputs.field(*, transformer=None, **kw)

Define an attrs field with a ‘transformer’ property.

The transformer, if given, should be a function of a single variable, which will be the attribute’s value. It will be used to transform the value before usage in C-code (e.g. by transformin from log to linear space).

py21cmfast.wrapper.inputs.get_logspaced_redshifts(min_redshift: float, z_step_factor: float, max_redshift: float) → tuple[float, Ellipsis]

Compute a sequence of redshifts to evolve over that are log-spaced.

Parameters:

• min_redshift (float)

• z_step_factor (float)

• max_redshift (float)

Return type:

tuple[float, Ellipsis]

py21cmfast.wrapper.inputs.input_param_field(kls: InputStruct)

Create an attrs field that must be an InputStruct.

Parameters:

kls (InputStruct subclass) – The parameter structure which should be returned as an attrs field

Parameters:

kls (InputStruct)

py21cmfast.wrapper.inputs.logtransformer(x, att: attrs.Attribute)

Convert from log to linear space.

Parameters:

att (attrs.Attribute)

py21cmfast.wrapper.inputs.FilterOptions

py21cmfast.wrapper.inputs.IntegralMethods

py21cmfast.wrapper.inputs.Planck18

py21cmfast.wrapper.inputs.logger