LIGHTNING_CONFIGURE_DEFAULTS#

Name#

LIGHTNING_CONFIGURE

Purpose#

Generates the Lightning configuration structure. This structure contains all of the possible selections users can make when configuring Lightning (e.g., models, fitting algorithms, etc.). The user can change the values for each tag to their desired values. Small guiding comments are included. These comments give basic details and indicate the expected values and type. Further details can be found at Configuration Settings.

Calling Sequence#

config = lightning_configure()

Output#

configstructure

A Lightning configuration structure.

Notes#

  • Parameters that take a flag (value of 0 or 1) are set if the value is 1 and not set if the value is 0.

  • Options that take arrays are indicated by brackets (i.e., []). These options can have multiple values. Options without brackets must contain a single value. Removing brackets may result in errors when the configuration structure is checked for errors.

  • All free parameters of the models have an associated prior structure, where the distribution type, prior distribution shape arguments, and initialization range are given. The prior distribution type options are: 'fixed', 'uniform', 'normal', and 'tabulated'. The number of values in the distribution shape argument array depend on the chosen distribution type.

    • 'fixed': takes a single value (Narg = 1), the value at which to fix the parameter. (If the prior is 'fixed', then the initialization range is ignored, since a fixed value does not need initialization.)

    • 'uniform': takes two values (Narg = 2), the minimum and maximum bounds of the distribution in that order.

    • 'normal': takes four values (Narg = 4), the minimum bound, maximum bound, distribution peak, and distribution standard deviation in that order.

    • 'tabulated': takes one value (Narg = 1), a string containing the path to the directory containing the user tabulated prior file (see Tabulated Priors for more details).

    The initialization range always has two values per prior, indicating the minimum and maximum bounds for the random initialization of the fitting algorithm.