Channels¶

This page contains classes defined in hakowan.channel module.

Visual channel models for positions, glyphs, fields, and maps.

`Channel` `dataclass` ¶

Channel base class.

Source code in hakowan/grammar/channel/channel.py

@dataclass(kw_only=True, slots=True)
class Channel:
    """Channel base class."""

    pass

`Position` `dataclass` ¶

Bases: Channel

Map a geometry-dimensional attribute to rendered positions.

Without this channel, source mesh vertices supply positions. Use it for deformed or alternate position attributes.

Attributes:

Name	Type	Description
`data`	`AttributeLike`	The attribute used to encode the position field.

Source code in hakowan/grammar/channel/channel.py

@dataclass(slots=True)
class Position(Channel):
    """Map a geometry-dimensional attribute to rendered positions.

    Without this channel, source mesh vertices supply positions. Use it for
    deformed or alternate position attributes.

    Attributes:
        data (AttributeLike): The attribute used to encode the position field.

    """

    data: AttributeLike

`Normal` `dataclass` ¶

Bases: Channel

Map a geometry-dimensional attribute to surface normals.

Without this channel, Hakowan computes normals from geometry.

Attributes:

Name	Type	Description
`data`	`AttributeLike`	The attribute used to encode the normal field.

Source code in hakowan/grammar/channel/channel.py

@dataclass(slots=True)
class Normal(Channel):
    """Map a geometry-dimensional attribute to surface normals.

    Without this channel, Hakowan computes normals from geometry.

    Attributes:
        data (AttributeLike): The attribute used to encode the normal field.

    """

    data: AttributeLike

`Size` `dataclass` ¶

Bases: Channel

Map a scalar attribute or constant to point/curve glyph size.

world values are radii in input geometry units. scene values are visible diameter fractions of the compiled scene bounding-box diagonal. screen values are visible diameters in output pixels.

Attributes:

Name	Type	Description
`data`	`AttributeLike \| float`	Attribute or constant size value.
`space`	`Literal['world', 'scene', 'screen']`	World, scene-relative, or screen-pixel sizing.

Source code in hakowan/grammar/channel/channel.py

@dataclass(slots=True)
class Size(Channel):
    """Map a scalar attribute or constant to point/curve glyph size.

    ``world`` values are radii in input geometry units. ``scene`` values are
    visible diameter fractions of the compiled scene bounding-box diagonal.
    ``screen`` values are visible diameters in output pixels.

    Attributes:
        data: Attribute or constant size value.
        space: World, scene-relative, or screen-pixel sizing.
    """

    data: AttributeLike | float
    space: Literal["world", "scene", "screen"] = "world"

`VectorField` `dataclass` ¶

Bases: Channel

Render a vertex- or facet-domain vector attribute as curve glyphs.

The vector must have the same dimension as the source geometry.

Attributes:

Name	Type	Description
`data`	`AttributeLike`	The attribute used to encode the vector field.
`refinement_level`	`int`	The refinement level of the vector field. This parameter is used to control the density of the vector field. The default value is 0.
`style`	`CurveStyle \| None`	The style of the vector field. If None, the default style will be used.
`end_type`	`Literal['point', 'arrow', 'flat']`	The type of the vector field end. `"point"` tapers the tip to zero (cone/spike); `"arrow"` renders a flared arrowhead; `"flat"` keeps a constant radius at both ends (cylinder). The default value is `"point"`.
`normalize`	`bool`	If True, every vector is rescaled to unit length so that all arrows have the same length and only encode direction. The magnitude is freed up to be mapped to another channel (e.g. `size` or color via `hakowan.norm()`). Normalization is applied before any scale attached to `data`, so a uniform scale on `data` controls the common arrow length. By default (False), arrow length is proportional to the vector magnitude. The default value is `False`.

Source code in hakowan/grammar/channel/channel.py

@dataclass(slots=True)
class VectorField(Channel):
    """Render a vertex- or facet-domain vector attribute as curve glyphs.

    The vector must have the same dimension as the source geometry.

    Attributes:
        data (AttributeLike): The attribute used to encode the vector field.
        refinement_level (int): The refinement level of the vector field. This parameter is used to
            control the density of the vector field. The default value is 0.
        style (CurveStyle | None): The style of the vector field. If None, the default style will
            be used.
        end_type (Literal["point", "arrow", "flat"]): The type of the vector field end.
            ``"point"`` tapers the tip to zero (cone/spike); ``"arrow"`` renders a
            flared arrowhead; ``"flat"`` keeps a constant radius at both ends
            (cylinder). The default value is ``"point"``.
        normalize (bool): If True, every vector is rescaled to unit length so
            that all arrows have the same length and only encode direction.
            The magnitude is freed up to be mapped to another channel (e.g.
            ``size`` or color via ``hakowan.norm()``). Normalization is
            applied *before* any scale attached to ``data``, so a uniform scale
            on ``data`` controls the common arrow length. By default (False),
            arrow length is proportional to the vector magnitude. The default
            value is ``False``.

    """

    data: AttributeLike
    refinement_level: int = 0
    style: CurveStyle | None = None
    end_type: Literal["point", "arrow", "flat"] = "point"
    normalize: bool = False

`Covariance` `dataclass` ¶

Bases: Channel

Map per-point 3x3 covariance data to anisotropic point glyphs.

full=True interprets the attribute as covariance; otherwise values are its square-root transform M where covariance is M @ M.T.

Attributes:

Name	Type	Description
`data`	`AttributeLike`	The attribute used to encode the covariance matrix.
`full`	`bool`	(bool): If True, the full covariance matrix is stored in the attribute. If False, its "square root", M, is stored. The full covariance matrix is ∑ := M @ M^T. The matrix M represenst the stretch and rotation transform applied on each mark.

Source code in hakowan/grammar/channel/channel.py

@dataclass(slots=True)
class Covariance(Channel):
    """Map per-point 3x3 covariance data to anisotropic point glyphs.

    ``full=True`` interprets the attribute as covariance; otherwise values are
    its square-root transform ``M`` where covariance is ``M @ M.T``.

    Attributes:
        data (AttributeLike): The attribute used to encode the covariance matrix.
        full: (bool): If True, the full covariance matrix is stored in the attribute.
            If False, its "square root", M, is stored. The full covariance matrix is ∑ := M @ M^T.
            The matrix M represenst the stretch and rotation transform applied on each mark.

    """

    data: AttributeLike
    full: bool = False

`Shape` `dataclass` ¶

Bases: Channel

Select and optionally orient sphere, disk, or cube point glyphs.

Attributes:

Name	Type	Description
`base_shape`	`Literal['sphere', 'disk', 'cube']`	The base shape used to represent a point. The default value is `"sphere"`.
`orientation`	`AttributeLike \| None`	The attribute used to encode the normal orientation of the shape. If None, orientation will be identity (i.e. normal along z-axis).

Source code in hakowan/grammar/channel/channel.py

@dataclass(slots=True)
class Shape(Channel):
    """Select and optionally orient sphere, disk, or cube point glyphs.

    Attributes:
        base_shape (Literal["sphere", "disk", "cube"]): The base shape used to represent a point.
            The default value is ``"sphere"``.
        orientation (AttributeLike | None): The attribute used to encode the normal orientation
            of the shape. If None, orientation will be identity (i.e. normal along z-axis).

    """

    base_shape: Literal["sphere", "disk", "cube"] = "sphere"
    orientation: Optional[AttributeLike] = None

`BumpMap` `dataclass` ¶

Bases: Channel

Perturb surface shading normals with a scalar bump texture.

Attributes:

Name	Type	Description
`texture`	`TextureLike`	The texture used to encode the bump map.
`scale`	`float`	The scale of the bump map. The default value is 1.0.

Source code in hakowan/grammar/channel/channel.py

@dataclass(slots=True)
class BumpMap(Channel):
    """Perturb surface shading normals with a scalar bump texture.

    Attributes:
        texture (TextureLike): The texture used to encode the bump map.
        scale (float): The scale of the bump map. The default value is 1.0.

    """

    texture: TextureLike
    scale: float = 1.0

`NormalMap` `dataclass` ¶

Bases: Channel

Perturb surface shading normals with a tangent-space normal texture.

Attributes:

Name	Type	Description
`texture`	`TextureLike`	The texture used to encode the normal map.

Source code in hakowan/grammar/channel/channel.py

@dataclass(slots=True)
class NormalMap(Channel):
    """Perturb surface shading normals with a tangent-space normal texture.

    Attributes:
        texture (TextureLike): The texture used to encode the normal map.

    """

    texture: TextureLike

Channels¶

Channel dataclass ¶

Position dataclass ¶

Normal dataclass ¶

Size dataclass ¶

VectorField dataclass ¶

Covariance dataclass ¶

Shape dataclass ¶

BumpMap dataclass ¶

NormalMap dataclass ¶

`Channel` `dataclass` ¶

`Position` `dataclass` ¶

`Normal` `dataclass` ¶

`Size` `dataclass` ¶

`VectorField` `dataclass` ¶

`Covariance` `dataclass` ¶

`Shape` `dataclass` ¶

`BumpMap` `dataclass` ¶

`NormalMap` `dataclass` ¶