crow.cluster_modules.projection_effects.richness_bias module

class crow.cluster_modules.projection_effects.richness_bias.CostanziRichnessBias(tau: ndarray[tuple[Any, ...], dtype[float64]] | None = None, delta_mu: ndarray[tuple[Any, ...], dtype[float64]] | None = None, fractional_sig_pure: ndarray[tuple[Any, ...], dtype[float64]] | None = None, fprj: ndarray[tuple[Any, ...], dtype[float64]] | None = None, fmsk: ndarray[tuple[Any, ...], dtype[float64]] | None = None)[source]

Bases: object

Projection effects model from Costanzi+19.

Each instance owns one calibration in parameters. Parameter values may be scalars or one-dimensional arrays broadcastable to the true-richness grid. fractional_sig_pure is the dimensionless fractional scatter, such that sig_pure = fractional_sig_pure * rich_tru.

Implemented as a CROW module.

Sprob_at_richtru(rich_obs_eds: ndarray[tuple[Any, ...], dtype[float64]], rich_obs_res: ndarray[tuple[Any, ...], dtype[float64]], rich_tru: ndarray[tuple[Any, ...], dtype[float64]]) → ndarray[tuple[Any, ...], dtype[floating]][source]

Integrate the observed richness over an interval defined by (rich_obs_low, rich_obs_hgh, rich_obs_res)

Parameters:

  • rich_obs_eds (ndarray) -- The boundaries defining the richness intervals. It has a dimension of 1 and length at least 2.

  • rich_obs_res (ndarray) -- The resolution used in the integration over each richness integral. If a constant is provided, all richness integrals use the same resolution. If an array provided, it has a dimension of 1 and length exactly of len(rich_obs_eds) - 1. The resolution is defined as the fractional increase of 1 + rich_obs_res. Smaller rich_obs_res means higher resolutions.

  • rich_tru (ndarray) -- The true richness

Returns:

Sprob_at_richtru -- The probability of int drich_obs P(rich_obs | rich_tru). The shape is (len(rich_obs_eds) - 1, len(rich_tru))

Return type:

ndarray

prob_richobs_at_richtru(rich_obs: ndarray[tuple[Any, ...], dtype[float64]], rich_tru: ndarray[tuple[Any, ...], dtype[float64]]) → ndarray[tuple[Any, ...], dtype[floating]][source]

Calculate the probability of observing the observed richness given the true richness. Equation 15 in Costanzi+19.

Parameters:

  • rich_obs (ndarray) -- The observed richness

  • rich_tru (ndarray) -- The true richness

Notes

This calculation uses tau, delta_mu, fractional_sig_pure, fprj, and fmsk from parameters.

Returns:

prob -- The probability of P(rich_obs | rich_tru). The shape is (len(rich_obs), len(rich_tru))

Return type:

ndarray