AxisymmetricJeansLikelihoodModel

jeanspy.sampler_numpyro.AxisymmetricJeansLikelihoodModel

See methods and properties for individual lookup pages, or the alphabetical API dictionary to search all classes. The full class contract and existing member anchors are retained below.

class jeanspy.sampler_numpyro.AxisymmetricJeansLikelihoodModel(dsph_model, parameter_specs, *, fixed_params=None, **kwargs)

Bases: jeanspy.sampler_numpyro.JeansLikelihoodModel

NumPyro likelihood for signed sky coordinates and zero mean streaming.

Uses the same ParameterSpec and NumPyroSampler contracts as the spherical likelihood. fixed_params and sampled physical names must be disjoint; optional postprocessing receives their combined dictionary. All angles in that dictionary are radians. Both q and q_projected parameterizations are supported by the axisymmetric forward model.

Parameter names and a named velocity mean are checked at construction. With parameter_postprocess, its output is checked after the callback instead; construction does not execute user callbacks or sample priors.

sigma2_bounds are admissibility limits: variances outside the interval receive zero likelihood, without clipping a finite prediction to a bound.

Notes

Inputs and units. dsph_model is a JAX AxisymmetricDSphModel; parameter_specs is a sequence of ParameterSpec, and fixed_params supplies complementary physical scalars. Calling this model takes matching finite nonempty 1-D x_pc/y_pc (pc) and vlos_kms/e_vlos_kms (km/s). Signed sky coordinates and the projected center are supported. Velocity mean and static sigmalos2 options are explicit.

Returns and shape. __call__ returns None while registering NumPyro sample, deterministic and likelihood sites; sample_parameters returns the transformed parameter mapping. Variance includes measurement error squared.

Validity. Positions in pc and velocities/errors in km/s, errors>=0. Gaussian LOS closure at fixed positions. The standard class does not add membership mixtures, velocity-cut normalization or binaries.

Errors. Bad schema/shape and sampled/fixed collisions raise; inadmissible forward variances are rejected with minus-infinite density.

Backend. NumPyro with JAX forward model.

Differentiation. Supported continuous physical parameters are traceable. J/D factors are not likelihood sites. Verify derivatives for custom prior transforms.

Examples. examples/axisymmetric_inference.py

Special-members:

__call__

sample_parameters()

Draw named physical parameters inside a NumPyro model execution.

Combines fixed_params with the sampled ParameterSpec values, then applies parameter_postprocess if supplied. Returns a physical-parameter mapping; invalid parameter names raise ValueError. Run under NumPyro inference or a seeded handler, not as an unseeded standalone random-number call.

sampling_identity()

Return model attributes for deterministic sampling-target identification.

The host dictionary includes the forward model, prior specifications and fixed settings; it is metadata rather than a physical prediction.