bayesline.api.equity.MetricDescriptor

bayesline.api.equity.MetricDescriptor#

model MetricDescriptor#

What a metric column means and how it may be aggregated and combined.

Fields#
Field	Type	Required	Default
`additive_over`	`list`[`str`]	No	factory
`domain`	`tuple`[`float`, `float`] \| `None`	No	`None`
`family`	`MetricFamily` \| `None`	No	`None`
`label`	`str`	Yes
`leaf_only_over`	`list`[`str`]	No	factory
`name`	`str`	Yes
`precision`	`int` \| `None`	No	`None`
`qualifiers`	`UnitQualifiers`	No	factory
`sign`	`Literal`[‘diverging’, ‘sequential’, ‘none’]	No	`'none'`
`time_aggregation`	`Literal`[‘compound’, ‘sum’, ‘last’, ‘mean’, ‘none’]	Yes
`unit`	`Unit`	Yes

field additive_over: list[str] [Optional]#: Dimensions over which a group’s value equals the sum of its members’ values, so members may be stacked to reproduce the group.

field domain: tuple[float, float] | None [Optional]#: A fixed natural domain, such as [-1, 1] for a correlation.

field family: MetricFamily | None [Optional]#: The portfolio / benchmark / active family the metric belongs to.

field label: str [Required]#: A display name for the metric.

field leaf_only_over: list[str] [Optional]#: Dimensions over which the metric is defined for leaf members only; groups carry no value.

field name: str [Required]#: The metric column name.

field precision: int | None [Optional]#: Preferred number of fraction digits.

field qualifiers: UnitQualifiers [Optional]#: Refinements of the unit that also govern axis sharing.

field sign: Literal['diverging', 'sequential', 'none'] [Optional]#: Colour semantics: diverging around zero, sequential from zero, or neither.

field time_aggregation: Literal['compound', 'sum', 'last', 'mean', 'none'] [Required]#: How the metric aggregates over the date dimension: compounded, summed, the last value, the mean, or not at all (valid at a single date only).

field unit: Unit [Required]#: The semantic unit of the metric.