prepare_bayesian_subjects

prepare_bayesian_subjects(
    estimates,
    as_of,
    *,
    forecast_source,
    settings=BAYESIAN_DEFAULTS,
    consensus_source=CONSENSUS_SOURCE,
)

Read the panel as of one moment and attach every history the posterior needs.

The history reads take their minimum observations, winsorisation and bias filter from the same settings the posterior is applied under, so the two cannot disagree.

Parameters

Name	Type	Description	Default
estimates	pd.DataFrame	Long input panel.	required
as_of	pd.Timestamp	Explicit exclusive UTC reading moment.	required
forecast_source	str	The source blended alongside consensus.	required
settings	BayesianSettings	The settings the posterior will be applied under.	`BAYESIAN_DEFAULTS`
consensus_source	str	The source treated as consensus.	`CONSENSUS_SOURCE`

Returns

Name	Type	Description
	pd.DataFrame	One row per subject with each source’s value and age, the growth prior, each source’s error variance and, when the bias applies, the consensus bias.

Raises

Name	Type	Description
	PanelError	The settings were fitted on data reaching as_of.