Renderer

The renderer was designed with extensibility in mind and allows you to create custom renderers for different specification formats. The renderer module exposes a BaseRenderer abstract class to provide the clients with a common interface. The AltairRenderer class is a concrete implementation of the BaseRenderer. In order to constraint the input handled by AltairRenderer, strict typing is introduced in the draco.renderer.altair.types module.

BaseRenderer

class draco.renderer.BaseRenderer

Base class for all renderers. Should handle the creation of a visualization represented as an object of type T.

__init__()

abstract render(spec, data, label_mapping=None)

Render a visualization from a dictionary-based specification and data.

Parameters:

• spec (dict) – Specification of the visualization.

• data (Union[dict[Any, Any], NativeDataFrame]) – Data to render.

• label_mapping (Union[dict[str, str], Callable[[str], str], None]) – Mapping of field names to human-readable labels. If a callable is provided, it will be called with the field name and should return a human-readable label.

Return type:

TypeVar(T)

Returns:

Produced visualization object of type T.

AltairRenderer

class draco.renderer.AltairRenderer(concat_mode=None, mark_config=None)

Produces a Vega-Lite visualization represented as an Altair chart object.

__init__(concat_mode=None, mark_config=None)

Instantiates a new Altair <https://altair-viz.github.io/>-based renderer.

Parameters:

• concat_mode (Optional[Literal['hconcat', 'vconcat']]) – The concatenation mode to use when concatenating multiple views. Only the first view is returned if None.

• mark_config (dict[str, dict[str, Any]] | None) – Optional custom mark configuration. The keys are the mark types, the values are the configuration dictionaries for the mark type.

render(spec, data, label_mapping=None)

Render a visualization from a dictionary-based specification and data.

Parameters:

• spec (dict) – Specification of the visualization.

• data (Union[dict[Any, Any], NativeDataFrame]) – Data to render.

• label_mapping (Union[dict[str, str], Callable[[str], str], None]) – Mapping of field names to human-readable labels. If a callable is provided, it will be called with the field name and should return a human-readable label.

Return type:

TypeVar(VegaLiteChart, VConcatChart, HConcatChart, FacetChart, Chart, LayerChart)

Returns:

Produced visualization object of type T.

Types

Note

Introducing the types through the below-described members made it possible to traverse specifications in a type-safe manner in the AltairRenderer class. The types reflect the default specification form, as described in the Facts chapters.

draco.renderer.altair.types.DatasetNumberRows

ID set for a field to be used as the entity_id when being processed as an ASP entity.

alias of Annotated[int, Gt(gt=0)]

class draco.renderer.altair.types.Encoding(**data)

Encoding schema. Encodings define how data fields map to visual properties (channel) of the mark.

Read More.

model_config: ClassVar[ConfigDict] = {'extra': 'forbid'}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

draco.renderer.altair.types.EncodingAggregate

How the data is binned into N bins. Positive integer. Described as (encoding,binning)

alias of Literal[‘count’, ‘mean’, ‘median’, ‘min’, ‘max’, ‘stdev’, ‘sum’]

draco.renderer.altair.types.EncodingBinning

Stacking strategy. One of zero, center, or normalize. Described as (encoding,stack).

alias of Annotated[int, Gt(gt=0)]

draco.renderer.altair.types.EncodingChannel

The field that maps to the visual property of the mark. Arbitrary string. Described as (encoding,field).

alias of Literal[‘x’, ‘y’, ‘color’, ‘size’, ‘shape’, ‘text’]

draco.renderer.altair.types.EncodingField

How the data is aggregated. One of count, mean, median, min, max, stdev, or sum. Described as (encoding,aggregate).

draco.renderer.altair.types.EncodingStack

The scale channel. One of x, y, color, size, shape, or text. Same as EncodingChannel. Described as (scale,channel).

alias of Literal[‘zero’, ‘center’, ‘normalize’]

class draco.renderer.altair.types.Facet(**data)

Facet schema. With the facet operator, we can partition a dataset by a field and create a view for each field. The resulting chart is often called a small multiples chart.

Read More.

model_config: ClassVar[ConfigDict] = {'extra': 'forbid'}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

draco.renderer.altair.types.FacetChannel

The facet field. Arbitrary string. Described as (facet,field).

alias of Literal[‘col’, ‘row’]

draco.renderer.altair.types.FacetField

Binning a numeric field for faceting. Positive integer. Described as (facet,binning).

class draco.renderer.altair.types.Field(**data)

Field schema. Represents a column in the dataset. Draco can use information about the field type and field statistics.

Read More.

model_config: ClassVar[ConfigDict] = {'extra': 'forbid'}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

draco.renderer.altair.types.FieldEntropy

The minimum value. Only used for numbers. Described as (field,min)

alias of Annotated[float, Ge(ge=0)]

draco.renderer.altair.types.FieldFreq

The minimum length of the field. Only used for strings. Described as (field,min_length)

alias of Annotated[int, Gt(gt=0)]

draco.renderer.altair.types.FieldId

The name of a data field. Described as (field,name)

draco.renderer.altair.types.FieldInteresting

The coordinates system of the view. Can be one of cartesian, or polar. Described as (view,coordinates).

draco.renderer.altair.types.FieldMax

The standard deviation. Only used for numbers. Described as (field,std)

draco.renderer.altair.types.FieldMaxLength

The time span of the field in seconds. Only used for datetime fields. Described as (field,span_seconds)

alias of Annotated[int, Ge(ge=0)]

draco.renderer.altair.types.FieldMin

The maximum value. Only used for numbers. Described as (field,max)

draco.renderer.altair.types.FieldMinLength

The maximum length of a field. Only used for strings. Described as (field,max_length)

alias of Annotated[int, Ge(ge=0)]

draco.renderer.altair.types.FieldName

The type of the data in the column for this field. One of number, string, boolean, or datetime. Described as (field,type)

draco.renderer.altair.types.FieldSkew

The frequency of the most common value. Only used for strings. Described as (field,freq)

draco.renderer.altair.types.FieldSpanSeconds

When the task regards specific fields, fields can be marked as relevant to the task. Described as (field,interesting)

alias of Annotated[int, Ge(ge=0)]

draco.renderer.altair.types.FieldStd

The skew of the field. Only used for numbers. Described as (field,skew)

draco.renderer.altair.types.FieldType

The number of unique values. Described as (field,unique)

alias of Literal[‘number’, ‘string’, ‘boolean’, ‘datetime’]

draco.renderer.altair.types.FieldUnique

The entropy of the field. Described as (field,entropy)

alias of Annotated[int, Gt(gt=0)]

class draco.renderer.altair.types.Mark(**data)

Mark schema. A mark represents the graphical mark of the visualization.

Read More.

model_config: ClassVar[ConfigDict] = {'extra': 'forbid'}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

draco.renderer.altair.types.MarkType

The visual channel. One of x, y, color, size, shape, or text. Same as ScaleChannel. Described as (encoding,channel).

alias of Literal[‘point’, ‘bar’, ‘line’, ‘area’, ‘text’, ‘tick’, ‘rect’]

class draco.renderer.altair.types.Scale(**data)

Scale schema. Scales map abstract values such as time or temperature to a visual value such as x- or y-position or color.

Read More.

model_config: ClassVar[ConfigDict] = {'extra': 'forbid'}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

draco.renderer.altair.types.ScaleChannel

The scale type. One of linear, log, ordinal, or categorical. Described as (scale,type).

alias of Literal[‘x’, ‘y’, ‘color’, ‘size’, ‘shape’, ‘text’]

draco.renderer.altair.types.ScaleType

Whether to include zero in the scale domain. Described as (scale,zero).

alias of Literal[‘linear’, ‘log’, ‘ordinal’, ‘categorical’]

draco.renderer.altair.types.ScaleZero

The task type. The task that the user tries to complete when looking at a visualization helps Draco find the most appropriate visualization.

class draco.renderer.altair.types.SchemaBase(**data)

Base class for all schema classes.

model_config: ClassVar[ConfigDict] = {'extra': 'forbid'}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

class draco.renderer.altair.types.SpecificationDict(**data)

Specification schema. Describes a visualization completely.

model_config: ClassVar[ConfigDict] = {'extra': 'forbid'}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

view: list[View]

Top-level scales define shared scales across views. None means no shared scales are used.

draco.renderer.altair.types.Task

The facet channel. Can be one of col and row. Described as (facet,channel).

alias of Literal[‘value’, ‘summary’]

class draco.renderer.altair.types.View(**data)

View schema. A view can group marks and scales together.

Read More.

facet: list[Facet] | None

View-specific scales define scales on a per-channel basis in the view. None means that a top-level scale (shared scale) is present.

model_config: ClassVar[ConfigDict] = {'extra': 'forbid'}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

draco.renderer.altair.types.ViewCoordinate

The mark type. One of point, bar, line, area, text, tick, or rect. Described as (mark,type).

alias of Literal[‘cartesian’, ‘polar’]