pypomp.core.results.Result¶

class pypomp.core.results.Result(method: str, kind: str, panel: bool, execution_time: float | None, key: Array, theta: Any = None, config: dict[str, ~typing.Any]=<factory>, payload: Dataset = <factory>, timestamp: Timestamp = <factory>)[source]¶

Bases: object

Container for the output of a single inference run.

Parameters:

method (str) – Name of the method that produced this result ("pfilter", "mif", "train", "pmcmc", "abc", "dpop_train").
kind (str) – Payload shape: "table" (pfilter-style log-likelihood table) or "trace" (iteration-by-iteration parameter/likelihood traces).
panel (bool) – Whether this result came from a PanelPomp.
execution_time (float or None) – Total wall-clock execution time in seconds.
key (jax.Array) – The JAX random key used for the run.
theta (object, optional) – The parameter object (PompParameters / PanelParameters / list).
config (dict) – Run hyperparameters, used for summaries and merge-equality guards.
payload (xarray.Dataset) – All numeric arrays produced by the run, keyed by variable name.
timestamp (pandas.Timestamp) – When the result was created.

Attributes

acceptance_rate¶: Per-chain acceptance rate for MCMC-family results.

n_chains¶: Number of chains / parameter sets (size of the theta_idx dim).

theta: Any = None¶

method: str¶

kind: str¶

panel: bool¶

execution_time: float | None¶

key: Array¶

config: dict[str, Any]¶

payload: Dataset¶

timestamp: Timestamp¶

Methods

`CLL`([average])	Return conditional log-likelihoods as a `pandas.DataFrame`.
`ESS`([average])	Return effective sample sizes as a `pandas.DataFrame`.
`merge`(*results)	Merge results of the same type by concatenating along `theta_idx`.
`print_summary`([n])	Print a human-readable summary of this result.
`to_dataframe`([ignore_nan])	Convert this result to a tidy `pandas.DataFrame`.
`traces`()	Return the parameter/likelihood trace as a `pandas.DataFrame`.