monitor_schema.models.column_schema#

Schema definitions.

Attributes#

COLUMN_NAME_TYPE

Classes#

`Metadata`	Metadata for a top-level objects such as monitors, analyzers, and schema.
`NoExtrasBaseModel`	No extras base model.
`Segment`	A segment is a list of tags.
`ColumnDiscreteness`	Classifying the type.
`ColumnDataType`	Options for configuring data type for a column.
`ColumnSchema`	Schema configuration for a column.
`WeightConfig`	Object that specifies column weights.
`SegmentWeightConfig`	Object that specifies column weights for a segment.
`EntitySchema`	Schema definition of an entity.
`EntityWeights`	Entity weight configurations.

Module Contents#

class monitor_schema.models.column_schema.Metadata[source]#

Bases: NoExtrasBaseModel

Metadata for a top-level objects such as monitors, analyzers, and schema.

This object is managed by WhyLabs. Any user-provided values will be ignored on WhyLabs side.

version: int#

schemaVersion: int | None#

updatedTimestamp: int#

author: str#

description: str | None#

class monitor_schema.models.column_schema.NoExtrasBaseModel[source]#

Bases: pydantic.BaseModel

No extras base model.

Inherit to prevent accidental extra fields.

class monitor_schema.models.column_schema.Segment[source]#

Bases: monitor_schema.models.commons.NoExtrasBaseModel

A segment is a list of tags.

We normalize these in the backend.

tags: List[SegmentTag]#

monitor_schema.models.column_schema.COLUMN_NAME_TYPE#

class monitor_schema.models.column_schema.ColumnDiscreteness[source]#

Bases: str, enum.Enum

Classifying the type.

discrete = 'discrete'#

continuous = 'continuous'#

class monitor_schema.models.column_schema.ColumnDataType[source]#

Bases: str, enum.Enum

Options for configuring data type for a column.

integral = 'integral'#

fractional = 'fractional'#

boolean = 'bool'#

string = 'string'#

unknown = 'unknown'#

null = 'null'#

class monitor_schema.models.column_schema.ColumnSchema[source]#

Bases: monitor_schema.models.commons.NoExtrasBaseModel

Schema configuration for a column.

Should be generated by WhyLabs originally but can be overridden by users.

discreteness: ColumnDiscreteness#

dataType: ColumnDataType#

classifier: str | None#

class monitor_schema.models.column_schema.WeightConfig[source]#

Bases: monitor_schema.models.commons.NoExtrasBaseModel

Object that specifies column weights.

By default, the weight of a column is None (unspecified)
If the weight is unspecified, the column is EXCLUDED when you perform a filter/sort by weight
For sorting, unweighted column take the LEAST PRECEDENCE, meaning that weight column have higher priorities
They are not hierarchical: if a segment weight config is specified and a column does not have a weight in that

config, we will not use any hierarchy to resolve the value. It will be None - Order of unweighted column is undefined.