Implementation:Evidentlyai Evidently Legacy Options

Knowledge Sources	Evidentlyai_Evidently
Domains	ML Monitoring, Configuration
Last Updated	2026-02-14 12:00 GMT

Overview

Options is the central configuration container class that aggregates all option types (color, render, data definition, and custom) for configuring Evidently legacy reports and metrics.

Description

The Options class extends BaseModel and serves as the unified options container for the Evidently legacy pipeline. It holds the following fields:

• color (Optional[ColorOptions]) -- color scheme options for data visualization
• render (Optional[RenderOptions]) -- rendering configuration options
• custom (Dict[Type[Option], Option]) -- a dictionary of custom option objects keyed by their type
• data_definition (Optional[DataDefinitionOptions]) -- data definition configuration

The class provides several convenience properties that return default instances when no explicit option is set:

• color_options -- returns the color options or a default ColorOptions()
• render_options -- returns the render options or a default RenderOptions()
• data_definition_options -- returns the data definition options or a default DataDefinitionOptions()

Key methods include:

• get(option_type) -- retrieves an option by type, checking first the known fields, then the custom dict (including subclass matching), and falling back to a default instance
• from_list(values) -- class method that creates an Options instance from a list of Option objects, automatically routing each to the appropriate field or the custom dict
• from_any_options(options) -- class method that normalizes various input formats (dict, single Option, list of Options, or Options instance) into an Options object
• override(other) -- merges another Options instance, with the other's non-None values taking precedence
• dict(...) -- overrides the Pydantic dict() method to always exclude the custom field (since it is not JSON-serializable)

The module also defines:

• _option_cls_mapping -- an auto-generated mapping from option types to field names
• AnyOptions -- a type alias representing all accepted option input formats: Union[Options, Option, dict, List[Option], None]

Usage

Use the Options class to configure reports, metrics, and renderers in the Evidently legacy API. It provides a flexible interface for passing configuration in multiple formats.

Code Reference

Source Location

• Repository: Evidentlyai_Evidently
• File: src/evidently/legacy/options/base.py

Signature

class Options(BaseModel):
    color: Optional[ColorOptions] = None
    render: Optional[RenderOptions] = None
    custom: Dict[Type[Option], Option] = {}
    data_definition: Optional[DataDefinitionOptions] = None

    @property
    def color_options(self) -> ColorOptions: ...
    @property
    def render_options(self) -> RenderOptions: ...
    @property
    def data_definition_options(self) -> DataDefinitionOptions: ...

    def get(self, option_type: Type[TypeParam]) -> TypeParam: ...

    @classmethod
    def from_list(cls, values: List[Option]) -> "Options": ...

    @classmethod
    def from_any_options(cls, options: "AnyOptions") -> "Options": ...

    def override(self, other: "Options") -> "Options": ...

    def dict(self, *, include=None, exclude=None, ...) -> "DictStrAny": ...

AnyOptions = Union[Options, Option, dict, List[Option], None]

Import

from evidently.legacy.options.base import Options, AnyOptions

I/O Contract

Inputs

Name	Type	Required	Description
color	Optional[ColorOptions]	No	Color scheme options for visualization.
render	Optional[RenderOptions]	No	Rendering configuration options.
custom	Dict[Type[Option], Option]	No	Dictionary of custom option objects. Defaults to empty dict.
data_definition	Optional[DataDefinitionOptions]	No	Data definition options.

Outputs

Name	Type	Description
Options	Options	A unified options container that can be passed to metrics, reports, and renderers.

Usage Examples

from evidently.legacy.options.base import Options
from evidently.legacy.options import ColorOptions

# Create options with custom color scheme
options = Options(color=ColorOptions(primary_color="#0000ff"))

# Create options from a list of Option objects
options = Options.from_list([ColorOptions(primary_color="#ff0000")])

# Create options from any supported format
options = Options.from_any_options({"color": {"primary_color": "#00ff00"}})

# Retrieve a specific option type
color_opts = options.get(ColorOptions)

# Override options
base_options = Options(color=ColorOptions(primary_color="#0000ff"))
override_options = Options(color=ColorOptions(primary_color="#ff0000"))
merged = base_options.override(override_options)

Related Pages

• Environment:Evidentlyai_Evidently_Python_Core_Environment