Skip to content

Archetypes

rerun.archetypes

class AnnotationContext

Bases: Archetype

Archetype: The AnnotationContext provides additional information on how to display entities.

Entities can use ClassIds and KeypointIds to provide annotations, and the labels and colors will be looked up in the appropriate AnnotationContext. We use the first annotation context we find in the path-hierarchy when searching up through the ancestors of a given entity path.

See also ClassDescription.

Example
Segmentation:

import numpy as np
import rerun as rr

rr.init("rerun_example_annotation_context_segmentation", spawn=True)

# Create a simple segmentation image
image = np.zeros((200, 300), dtype=np.uint8)
image[50:100, 50:120] = 1
image[100:180, 130:280] = 2

# Log an annotation context to assign a label and color to each class
rr.log("segmentation", rr.AnnotationContext([(1, "red", (255, 0, 0)), (2, "green", (0, 255, 0))]), timeless=True)

rr.log("segmentation/image", rr.SegmentationImage(image))

def __init__(context)

Create a new instance of the AnnotationContext archetype.

PARAMETER DESCRIPTION
context

List of class descriptions, mapping class indices to class names, colors etc.

TYPE: AnnotationContextLike

class Arrows2D

Bases: Arrows2DExt, Archetype

Archetype: 2D arrows with optional colors, radii, labels, etc.

Example
Simple batch of 2D Arrows:

import rerun as rr

rr.init("rerun_example_arrow2d", spawn=True)

rr.log(
    "arrows",
    rr.Arrows2D(
        origins=[[0.25, 0.0], [0.25, 0.0], [-0.1, -0.1]],
        vectors=[[1.0, 0.0], [0.0, -1.0], [-0.7, 0.7]],
        colors=[[255, 0, 0], [0, 255, 0], [127, 0, 255]],
        labels=["right", "up", "left-down"],
        radii=0.025,
    ),
)

def __init__(*, vectors, origins=None, radii=None, colors=None, labels=None, class_ids=None, instance_keys=None)

Create a new instance of the Arrows2D archetype.

PARAMETER DESCRIPTION
vectors

All the vectors for each arrow in the batch.

TYPE: Vec2DArrayLike

origins

All the origin points for each arrow in the batch.

If no origins are set, (0, 0, 0) is used as the origin for each arrow.

TYPE: Vec2DArrayLike | None DEFAULT: None

radii

Optional radii for the arrows.

The shaft is rendered as a line with radius = 0.5 * radius. The tip is rendered with height = 2.0 * radius and radius = 1.0 * radius.

TYPE: RadiusArrayLike | None DEFAULT: None

colors

Optional colors for the points.

TYPE: Rgba32ArrayLike | None DEFAULT: None

labels

Optional text labels for the arrows.

TYPE: Utf8ArrayLike | None DEFAULT: None

class_ids

Optional class Ids for the points.

The class ID provides colors and labels if not specified explicitly.

TYPE: ClassIdArrayLike | None DEFAULT: None

instance_keys

Unique identifiers for each individual point in the batch.

TYPE: InstanceKeyArrayLike | None DEFAULT: None

class Arrows3D

Bases: Arrows3DExt, Archetype

Archetype: 3D arrows with optional colors, radii, labels, etc.

Example
Simple batch of 3D Arrows:

from math import tau

import numpy as np
import rerun as rr

rr.init("rerun_example_arrow3d", spawn=True)

lengths = np.log2(np.arange(0, 100) + 1)
angles = np.arange(start=0, stop=tau, step=tau * 0.01)
origins = np.zeros((100, 3))
vectors = np.column_stack([np.sin(angles) * lengths, np.zeros(100), np.cos(angles) * lengths])
colors = [[1.0 - c, c, 0.5, 0.5] for c in angles / tau]

rr.log("arrows", rr.Arrows3D(origins=origins, vectors=vectors, colors=colors))

def __init__(*, vectors, origins=None, radii=None, colors=None, labels=None, class_ids=None, instance_keys=None)

Create a new instance of the Arrows3D archetype.

PARAMETER DESCRIPTION
vectors

All the vectors for each arrow in the batch.

TYPE: Vec3DArrayLike

origins

All the origin points for each arrow in the batch.

If no origins are set, (0, 0, 0) is used as the origin for each arrow.

TYPE: Vec3DArrayLike | None DEFAULT: None

radii

Optional radii for the arrows.

The shaft is rendered as a line with radius = 0.5 * radius. The tip is rendered with height = 2.0 * radius and radius = 1.0 * radius.

TYPE: RadiusArrayLike | None DEFAULT: None

colors

Optional colors for the points.

TYPE: Rgba32ArrayLike | None DEFAULT: None

labels

Optional text labels for the arrows.

TYPE: Utf8ArrayLike | None DEFAULT: None

class_ids

Optional class Ids for the points.

The class ID provides colors and labels if not specified explicitly.

TYPE: ClassIdArrayLike | None DEFAULT: None

instance_keys

Unique identifiers for each individual point in the batch.

TYPE: InstanceKeyArrayLike | None DEFAULT: None

class Asset3D

Bases: Asset3DExt, Archetype

Archetype: A prepacked 3D asset (.gltf, .glb, .obj, etc.).

See also Mesh3D.

Example
Simple 3D asset:

import sys

import rerun as rr

if len(sys.argv) < 2:
    print(f"Usage: {sys.argv[0]} <path_to_asset.[gltf|glb|obj]>")
    sys.exit(1)

rr.init("rerun_example_asset3d", spawn=True)

rr.log("world", rr.ViewCoordinates.RIGHT_HAND_Z_UP, timeless=True)  # Set an up-axis
rr.log("world/asset", rr.Asset3D(path=sys.argv[1]))

def __init__(*, path=None, contents=None, media_type=None, transform=None)

Create a new instance of the Asset3D archetype.

PARAMETER DESCRIPTION
path

A path to an file stored on the local filesystem. Mutually exclusive with contents.

TYPE: str | Path | None DEFAULT: None

contents

The contents of the file. Can be a BufferedReader, BytesIO, or bytes. Mutually exclusive with path.

TYPE: BlobLike | None DEFAULT: None

media_type

The Media Type of the asset.

For instance: * model/gltf-binary * model/obj

If omitted, it will be guessed from the path (if any), or the viewer will try to guess from the contents (magic header). If the media type cannot be guessed, the viewer won't be able to render the asset.

TYPE: Utf8Like | None DEFAULT: None

transform

An out-of-tree transform.

Applies a transformation to the asset itself without impacting its children.

TYPE: Transform3DLike | None DEFAULT: None

class BarChart

Bases: BarChartExt, Archetype

Archetype: A bar chart.

The x values will be the indices of the array, and the bar heights will be the provided values.

Example
Simple bar chart:

import rerun as rr

rr.init("rerun_example_bar_chart", spawn=True)
rr.log("bar_chart", rr.BarChart([8, 4, 0, 9, 1, 4, 1, 6, 9, 0]))

def __init__(values, *, color=None)

Create a new instance of the BarChart archetype.

PARAMETER DESCRIPTION
values

The values. Should always be a rank-1 tensor.

TYPE: TensorDataLike

color

The color of the bar chart

TYPE: Rgba32Like | None DEFAULT: None

class Boxes2D

Bases: Boxes2DExt, Archetype

Archetype: 2D boxes with half-extents and optional center, rotations, rotations, colors etc.

Example
Simple 2D boxes:

import rerun as rr

rr.init("rerun_example_box2d", spawn=True)

rr.log("simple", rr.Boxes2D(mins=[-1, -1], sizes=[2, 2]))

# Log an extra rect to set the view bounds
rr.log("bounds", rr.Boxes2D(sizes=[4.0, 3.0]))

def __init__(*, sizes=None, mins=None, half_sizes=None, centers=None, array=None, array_format=None, radii=None, colors=None, labels=None, draw_order=None, class_ids=None, instance_keys=None)

Create a new instance of the Boxes2D archetype.

PARAMETER DESCRIPTION
sizes

Full extents in x/y. Incompatible with array and half_sizes.

TYPE: Vec2DArrayLike | None DEFAULT: None

half_sizes

All half-extents that make up the batch of boxes. Specify this instead of sizes Incompatible with array and sizes.

TYPE: Vec2DArrayLike | None DEFAULT: None

mins

Minimum coordinates of the boxes. Specify this instead of centers. Incompatible with array. Only valid when used together with either sizes or half_sizes.

TYPE: Vec2DArrayLike | None DEFAULT: None

array

An array of boxes in the format specified by array_format. Requires specifying array_format. Incompatible with sizes, half_sizes, mins and centers.

TYPE: ArrayLike | None DEFAULT: None

array_format

How to interpret the data in array.

TYPE: Box2DFormat | None DEFAULT: None

centers

Optional center positions of the boxes.

TYPE: Vec2DArrayLike | None DEFAULT: None

colors

Optional colors for the boxes.

TYPE: Rgba32ArrayLike | None DEFAULT: None

radii

Optional radii for the lines that make up the boxes.

TYPE: RadiusArrayLike | None DEFAULT: None

labels

Optional text labels for the boxes.

TYPE: Utf8ArrayLike | None DEFAULT: None

draw_order

An optional floating point value that specifies the 2D drawing order. Objects with higher values are drawn on top of those with lower values.

The default for 2D boxes is 10.0.

TYPE: DrawOrderLike | None DEFAULT: None

class_ids

Optional ClassIds for the boxes.

The class ID provides colors and labels if not specified explicitly.

TYPE: ClassIdArrayLike | None DEFAULT: None

instance_keys

Unique identifiers for each individual boxes in the batch.

TYPE: InstanceKeyArrayLike | None DEFAULT: None

class Boxes3D

Bases: Boxes3DExt, Archetype

Archetype: 3D boxes with half-extents and optional center, rotations, rotations, colors etc.

Example
Batch of 3D boxes:

import rerun as rr
from rerun.datatypes import Angle, Quaternion, Rotation3D, RotationAxisAngle

rr.init("rerun_example_box3d_batch", spawn=True)

rr.log(
    "batch",
    rr.Boxes3D(
        centers=[[2, 0, 0], [-2, 0, 0], [0, 0, 2]],
        half_sizes=[[2.0, 2.0, 1.0], [1.0, 1.0, 0.5], [2.0, 0.5, 1.0]],
        rotations=[
            Rotation3D.identity(),
            Quaternion(xyzw=[0.0, 0.0, 0.382683, 0.923880]),  # 45 degrees around Z
            RotationAxisAngle(axis=[0, 1, 0], angle=Angle(deg=30)),
        ],
        radii=0.025,
        colors=[(255, 0, 0), (0, 255, 0), (0, 0, 255)],
        labels=["red", "green", "blue"],
    ),
)

def __init__(*, sizes=None, mins=None, half_sizes=None, centers=None, rotations=None, colors=None, radii=None, labels=None, class_ids=None, instance_keys=None)

Create a new instance of the Boxes3D archetype.

PARAMETER DESCRIPTION
sizes

Full extents in x/y/z. Specify this instead of half_sizes

TYPE: Vec3DArrayLike | None DEFAULT: None

half_sizes

All half-extents that make up the batch of boxes. Specify this instead of sizes

TYPE: Vec3DArrayLike | None DEFAULT: None

mins

Minimum coordinates of the boxes. Specify this instead of centers.

Only valid when used together with either sizes or half_sizes.

TYPE: Vec3DArrayLike | None DEFAULT: None

centers

Optional center positions of the boxes.

TYPE: Vec3DArrayLike | None DEFAULT: None

rotations

Optional rotations of the boxes.

TYPE: Rotation3DArrayLike | None DEFAULT: None

colors

Optional colors for the boxes.

TYPE: Rgba32ArrayLike | None DEFAULT: None

radii

Optional radii for the lines that make up the boxes.

TYPE: RadiusArrayLike | None DEFAULT: None

labels

Optional text labels for the boxes.

TYPE: Utf8ArrayLike | None DEFAULT: None

class_ids

Optional ClassIds for the boxes.

The class ID provides colors and labels if not specified explicitly.

TYPE: ClassIdArrayLike | None DEFAULT: None

instance_keys

Unique identifiers for each individual boxes in the batch.

TYPE: InstanceKeyArrayLike | None DEFAULT: None

class Clear

Bases: ClearExt, Archetype

Archetype: Empties all the components of an entity.

The presence of a clear means that a latest-at query of components at a given path(s) will not return any components that were logged at those paths before the clear. Any logged components after the clear are unaffected by the clear.

This implies that a range query that includes time points that are before the clear, still returns all components at the given path(s). Meaning that in practice clears are ineffective when making use of visible time ranges. Scalar plots are an exception: they track clears and use them to represent holes in the data (i.e. discontinuous lines).

Example
Flat:

import rerun as rr

rr.init("rerun_example_clear", spawn=True)

vectors = [(1.0, 0.0, 0.0), (0.0, -1.0, 0.0), (-1.0, 0.0, 0.0), (0.0, 1.0, 0.0)]
origins = [(-0.5, 0.5, 0.0), (0.5, 0.5, 0.0), (0.5, -0.5, 0.0), (-0.5, -0.5, 0.0)]
colors = [(200, 0, 0), (0, 200, 0), (0, 0, 200), (200, 0, 200)]

# Log a handful of arrows.
for i, (vector, origin, color) in enumerate(zip(vectors, origins, colors)):
    rr.log(f"arrows/{i}", rr.Arrows3D(vectors=vector, origins=origin, colors=color))

# Now clear them, one by one on each tick.
for i in range(len(vectors)):
    rr.log(f"arrows/{i}", rr.Clear(recursive=False))  # or `rr.Clear.flat()`

def __init__(*, recursive)

Create a new instance of the Clear archetype.

PARAMETER DESCRIPTION
recursive

Whether to recursively clear all children.

TYPE: bool

def flat() staticmethod

Returns a non-recursive clear archetype.

This will empty all components of the associated entity at the logged timepoint. Children will be left untouched.

def recursive() staticmethod

Returns a recursive clear archetype.

This will empty all components of the associated entity at the logged timepoint, as well as all components of all its recursive children.

class DepthImage

Bases: DepthImageExt, Archetype

Archetype: A depth image.

The shape of the TensorData must be mappable to an HxW tensor. Each pixel corresponds to a depth value in units specified by meter.

Example
Depth to 3D example:

import numpy as np
import rerun as rr

depth_image = 65535 * np.ones((200, 300), dtype=np.uint16)
depth_image[50:150, 50:150] = 20000
depth_image[130:180, 100:280] = 45000

rr.init("rerun_example_depth_image_3d", spawn=True)

# If we log a pinhole camera model, the depth gets automatically back-projected to 3D
rr.log(
    "world/camera",
    rr.Pinhole(
        width=depth_image.shape[1],
        height=depth_image.shape[0],
        focal_length=200,
    ),
)

# Log the tensor.
rr.log("world/camera/depth", rr.DepthImage(depth_image, meter=10_000.0))

def __init__(data, *, meter=None, draw_order=None)

Create a new instance of the DepthImage archetype.

PARAMETER DESCRIPTION
data

The depth-image data. Should always be a rank-2 tensor.

TYPE: TensorDataLike

meter

An optional floating point value that specifies how long a meter is in the native depth units.

For instance: with uint16, perhaps meter=1000 which would mean you have millimeter precision and a range of up to ~65 meters (2^16 / 1000).

TYPE: DepthMeterLike | None DEFAULT: None

draw_order

An optional floating point value that specifies the 2D drawing order.

Objects with higher values are drawn on top of those with lower values.

TYPE: DrawOrderLike | None DEFAULT: None

class DisconnectedSpace

Bases: DisconnectedSpaceExt, Archetype

Archetype: Spatially disconnect this entity from its parent.

Specifies that the entity path at which this is logged is spatially disconnected from its parent, making it impossible to transform the entity path into its parent's space and vice versa. It only applies to space views that work with spatial transformations, i.e. 2D & 3D space views. This is useful for specifying that a subgraph is independent of the rest of the scene.

Example
Disconnected Space:

import rerun as rr

rr.init("rerun_example_disconnected_space", spawn=True)

# These two points can be projected into the same space..
rr.log("world/room1/point", rr.Points3D([[0, 0, 0]]))
rr.log("world/room2/point", rr.Points3D([[1, 1, 1]]))

# ..but this one lives in a completely separate space!
rr.log("world/wormhole", rr.DisconnectedSpace())
rr.log("world/wormhole/point", rr.Points3D([[2, 2, 2]]))

def __init__(is_disconnected=True)

Disconnect an entity from its parent.

PARAMETER DESCRIPTION
is_disconnected

Whether or not the entity should be disconnected from the rest of the scene. Set to True to disconnect the entity from its parent. Set to False to disable the effects of this archetype, (re-)connecting the entity to its parent again.

TYPE: DisconnectedSpaceLike DEFAULT: True

class Image

Bases: ImageExt, Archetype

Archetype: A monochrome or color image.

The shape of the TensorData must be mappable to: - A HxW tensor, treated as a grayscale image. - A HxWx3 tensor, treated as an RGB image. - A HxWx4 tensor, treated as an RGBA image.

Leading and trailing unit-dimensions are ignored, so that 1x640x480x3x1 is treated as a 640x480x3 RGB image.

For an easy way to pass in image formats or encoded images, see rerun.ImageEncoded.

Example
image_simple:

import numpy as np
import rerun as rr

# Create an image with numpy
image = np.zeros((200, 300, 3), dtype=np.uint8)
image[:, :, 0] = 255
image[50:150, 50:150] = (0, 255, 0)

rr.init("rerun_example_image", spawn=True)

rr.log("image", rr.Image(image))

def __init__(data, *, draw_order=None)

Create a new instance of the Image archetype.

PARAMETER DESCRIPTION
data

The image data. Should always be a rank-2 or rank-3 tensor.

TYPE: TensorDataLike

draw_order

An optional floating point value that specifies the 2D drawing order.

Objects with higher values are drawn on top of those with lower values.

TYPE: DrawOrderLike | None DEFAULT: None

def compress(*, jpeg_quality=95)

Converts an Image to an rerun.ImageEncoded using JPEG compression.

JPEG compression works best for photographs. Only RGB or Mono images are supported, not RGBA. Note that compressing to JPEG costs a bit of CPU time, both when logging and later when viewing them.

PARAMETER DESCRIPTION
jpeg_quality

Higher quality = larger file size. A quality of 95 still saves a lot of space, but is visually very similar.

TYPE: int DEFAULT: 95

class LineStrips2D

Bases: Archetype

Archetype: 2D line strips with positions and optional colors, radii, labels, etc.

Example
line_strip2d_batch:

import rerun as rr

rr.init("rerun_example_line_strip2d_batch", spawn=True)

rr.log(
    "strips",
    rr.LineStrips2D(
        [
            [[0, 0], [2, 1], [4, -1], [6, 0]],
            [[0, 3], [1, 4], [2, 2], [3, 4], [4, 2], [5, 4], [6, 3]],
        ],
        colors=[[255, 0, 0], [0, 255, 0]],
        radii=[0.025, 0.005],
        labels=["one strip here", "and one strip there"],
    ),
)

# Log an extra rect to set the view bounds
rr.log("bounds", rr.Boxes2D(centers=[3, 1.5], half_sizes=[4.0, 4.5]))

def __init__(strips, *, radii=None, colors=None, labels=None, draw_order=None, class_ids=None, instance_keys=None)

Create a new instance of the LineStrips2D archetype.

PARAMETER DESCRIPTION
strips

All the actual 2D line strips that make up the batch.

TYPE: LineStrip2DArrayLike

radii

Optional radii for the line strips.

TYPE: RadiusArrayLike | None DEFAULT: None

colors

Optional colors for the line strips.

TYPE: Rgba32ArrayLike | None DEFAULT: None

labels

Optional text labels for the line strips.

TYPE: Utf8ArrayLike | None DEFAULT: None

draw_order

An optional floating point value that specifies the 2D drawing order of each line strip.

Objects with higher values are drawn on top of those with lower values.

TYPE: DrawOrderLike | None DEFAULT: None

class_ids

Optional ClassIds for the lines.

The class ID provides colors and labels if not specified explicitly.

TYPE: ClassIdArrayLike | None DEFAULT: None

instance_keys

Unique identifiers for each individual line strip in the batch.

TYPE: InstanceKeyArrayLike | None DEFAULT: None

class LineStrips3D

Bases: Archetype

Archetype: 3D line strips with positions and optional colors, radii, labels, etc.

Example
Many strips:

import rerun as rr

rr.init("rerun_example_line_strip3d_batch", spawn=True)

rr.log(
    "strips",
    rr.LineStrips3D(
        [
            [
                [0, 0, 2],
                [1, 0, 2],
                [1, 1, 2],
                [0, 1, 2],
            ],
            [
                [0, 0, 0],
                [0, 0, 1],
                [1, 0, 0],
                [1, 0, 1],
                [1, 1, 0],
                [1, 1, 1],
                [0, 1, 0],
                [0, 1, 1],
            ],
        ],
        colors=[[255, 0, 0], [0, 255, 0]],
        radii=[0.025, 0.005],
        labels=["one strip here", "and one strip there"],
    ),
)

def __init__(strips, *, radii=None, colors=None, labels=None, class_ids=None, instance_keys=None)

Create a new instance of the LineStrips3D archetype.

PARAMETER DESCRIPTION
strips

All the actual 3D line strips that make up the batch.

TYPE: LineStrip3DArrayLike

radii

Optional radii for the line strips.

TYPE: RadiusArrayLike | None DEFAULT: None

colors

Optional colors for the line strips.

TYPE: Rgba32ArrayLike | None DEFAULT: None

labels

Optional text labels for the line strips.

TYPE: Utf8ArrayLike | None DEFAULT: None

class_ids

Optional ClassIds for the lines.

The class ID provides colors and labels if not specified explicitly.

TYPE: ClassIdArrayLike | None DEFAULT: None

instance_keys

Unique identifiers for each individual line strip in the batch.

TYPE: InstanceKeyArrayLike | None DEFAULT: None

class Mesh3D

Bases: Mesh3DExt, Archetype

Archetype: A 3D triangle mesh as specified by its per-mesh and per-vertex properties.

See also Asset3D.

Example
Simple indexed 3D mesh:

import rerun as rr

rr.init("rerun_example_mesh3d_indexed", spawn=True)

rr.log(
    "triangle",
    rr.Mesh3D(
        vertex_positions=[[0.0, 1.0, 0.0], [1.0, 0.0, 0.0], [0.0, 0.0, 0.0]],
        vertex_normals=[0.0, 0.0, 1.0],
        vertex_colors=[[0, 0, 255], [0, 255, 0], [255, 0, 0]],
        indices=[2, 1, 0],
    ),
)

def __init__(*, vertex_positions, indices=None, mesh_properties=None, vertex_normals=None, vertex_colors=None, vertex_texcoords=None, albedo_texture=None, mesh_material=None, class_ids=None, instance_keys=None)

Create a new instance of the Mesh3D archetype.

PARAMETER DESCRIPTION
vertex_positions

The positions of each vertex. If no indices are specified, then each triplet of positions is interpreted as a triangle.

TYPE: Vec3DArrayLike

indices

If specified, a flattened array of indices that describe the mesh's triangles, i.e. its length must be divisible by 3. Mutually exclusive with mesh_properties.

TYPE: ArrayLike | None DEFAULT: None

mesh_properties

Optional properties for the mesh as a whole (including indexed drawing). Mutually exclusive with indices.

TYPE: MeshPropertiesLike | None DEFAULT: None

vertex_normals

An optional normal for each vertex. If specified, this must have as many elements as vertex_positions.

TYPE: Vec3DArrayLike | None DEFAULT: None

vertex_texcoords

An optional texture coordinate for each vertex. If specified, this must have as many elements as vertex_positions.

TYPE: Vec2DArrayLike | None DEFAULT: None

vertex_colors

An optional color for each vertex.

TYPE: Rgba32ArrayLike | None DEFAULT: None

mesh_material

Optional material properties for the mesh as a whole.

TYPE: MaterialLike | None DEFAULT: None

albedo_texture

Optional albedo texture. Used with vertex_texcoords on Mesh3D. Currently supports only sRGB(A) textures, ignoring alpha. (meaning that the tensor must have 3 or 4 channels and use the u8 format)

TYPE: TensorDataLike | None DEFAULT: None

class_ids

Optional class Ids for the vertices. The class ID provides colors and labels if not specified explicitly.

TYPE: ClassIdArrayLike | None DEFAULT: None

instance_keys

Unique identifiers for each individual vertex in the mesh.

TYPE: InstanceKeyArrayLike | None DEFAULT: None

class Pinhole

Bases: PinholeExt, Archetype

Archetype: Camera perspective projection (a.k.a. intrinsics).

Example
Simple Pinhole Camera:

import numpy as np
import rerun as rr

rr.init("rerun_example_pinhole", spawn=True)
rng = np.random.default_rng(12345)

image = rng.uniform(0, 255, size=[3, 3, 3])
rr.log("world/image", rr.Pinhole(focal_length=3, width=3, height=3))
rr.log("world/image", rr.Image(image))

def __init__(*, image_from_camera=None, resolution=None, camera_xyz=None, width=None, height=None, focal_length=None, principal_point=None)

Create a new instance of the Pinhole archetype.

PARAMETER DESCRIPTION
image_from_camera

Row-major intrinsics matrix for projecting from camera space to image space. The first two axes are X=Right and Y=Down, respectively. Projection is done along the positive third (Z=Forward) axis. This can be specified instead of focal_length and principal_point.

TYPE: Mat3x3Like | None DEFAULT: None

resolution

Pixel resolution (usually integers) of child image space. Width and height. image_from_camera projects onto the space spanned by (0,0) and resolution - 1.

TYPE: Vec2DLike | None DEFAULT: None

camera_xyz

Sets the view coordinates for the camera.

All common values are available as constants on the components.ViewCoordinates class.

The default is ViewCoordinates.RDF, i.e. X=Right, Y=Down, Z=Forward, and this is also the recommended setting. This means that the camera frustum will point along the positive Z axis of the parent space, and the cameras "up" direction will be along the negative Y axis of the parent space.

The camera frustum will point whichever axis is set to F (or the opposite of B). When logging a depth image under this entity, this is the direction the point cloud will be projected. With RDF, the default forward is +Z.

The frustum's "up" direction will be whichever axis is set to U (or the opposite of D). This will match the negative Y direction of pixel space (all images are assumed to have xyz=RDF). With RDF, the default is up is -Y.

The frustum's "right" direction will be whichever axis is set to R (or the opposite of L). This will match the positive X direction of pixel space (all images are assumed to have xyz=RDF). With RDF, the default right is +x.

Other common formats are RUB (X=Right, Y=Up, Z=Back) and FLU (X=Forward, Y=Left, Z=Up).

NOTE: setting this to something else than RDF (the default) will change the orientation of the camera frustum, and make the pinhole matrix not match up with the coordinate system of the pinhole entity.

The pinhole matrix (the image_from_camera argument) always project along the third (Z) axis, but will be re-oriented to project along the forward axis of the camera_xyz argument.

TYPE: ViewCoordinatesLike | None DEFAULT: None

focal_length

The focal length of the camera in pixels. This is the diagonal of the projection matrix. Set one value for symmetric cameras, or two values (X=Right, Y=Down) for anamorphic cameras.

TYPE: float | ArrayLike | None DEFAULT: None

principal_point

The center of the camera in pixels. The default is half the width and height. This is the last column of the projection matrix. Expects two values along the dimensions Right and Down

TYPE: ArrayLike | None DEFAULT: None

width

Width of the image in pixels.

TYPE: int | float | None DEFAULT: None

height

Height of the image in pixels.

TYPE: int | float | None DEFAULT: None

class Points2D

Bases: Points2DExt, Archetype

Archetype: A 2D point cloud with positions and optional colors, radii, labels, etc.

Example
Randomly distributed 2D points with varying color and radius:

import rerun as rr
from numpy.random import default_rng

rr.init("rerun_example_points2d_random", spawn=True)
rng = default_rng(12345)

positions = rng.uniform(-3, 3, size=[10, 2])
colors = rng.uniform(0, 255, size=[10, 4])
radii = rng.uniform(0, 1, size=[10])

rr.log("random", rr.Points2D(positions, colors=colors, radii=radii))

# Log an extra rect to set the view bounds
rr.log("bounds", rr.Boxes2D(half_sizes=[4, 3]))

def __init__(positions, *, radii=None, colors=None, labels=None, draw_order=None, class_ids=None, keypoint_ids=None, instance_keys=None)

Create a new instance of the Points2D archetype.

PARAMETER DESCRIPTION
positions

All the 2D positions at which the point cloud shows points.

TYPE: Vec2DArrayLike

radii

Optional radii for the points, effectively turning them into circles.

TYPE: RadiusArrayLike | None DEFAULT: None

colors

Optional colors for the points.

The colors are interpreted as RGB or RGBA in sRGB gamma-space, As either 0-1 floats or 0-255 integers, with separate alpha.

TYPE: Rgba32ArrayLike | None DEFAULT: None

labels

Optional text labels for the points.

TYPE: Utf8ArrayLike | None DEFAULT: None

draw_order

An optional floating point value that specifies the 2D drawing order. Objects with higher values are drawn on top of those with lower values.

TYPE: DrawOrderLike | None DEFAULT: None

class_ids

Optional class Ids for the points.

The class ID provides colors and labels if not specified explicitly.

TYPE: ClassIdArrayLike | None DEFAULT: None

keypoint_ids

Optional keypoint IDs for the points, identifying them within a class.

If keypoint IDs are passed in but no class IDs were specified, the class ID will default to 0. This is useful to identify points within a single classification (which is identified with class_id). E.g. the classification might be 'Person' and the keypoints refer to joints on a detected skeleton.

TYPE: KeypointIdArrayLike | None DEFAULT: None

instance_keys

Unique identifiers for each individual point in the batch.

TYPE: InstanceKeyArrayLike | None DEFAULT: None

class Points3D

Bases: Points3DExt, Archetype

Archetype: A 3D point cloud with positions and optional colors, radii, labels, etc.

Example
Randomly distributed 3D points with varying color and radius:

import rerun as rr
from numpy.random import default_rng

rr.init("rerun_example_points3d_random", spawn=True)
rng = default_rng(12345)

positions = rng.uniform(-5, 5, size=[10, 3])
colors = rng.uniform(0, 255, size=[10, 3])
radii = rng.uniform(0, 1, size=[10])

rr.log("random", rr.Points3D(positions, colors=colors, radii=radii))

def __init__(positions, *, radii=None, colors=None, labels=None, class_ids=None, keypoint_ids=None, instance_keys=None)

Create a new instance of the Points3D archetype.

PARAMETER DESCRIPTION
positions

All the 3D positions at which the point cloud shows points.

TYPE: Vec3DArrayLike

radii

Optional radii for the points, effectively turning them into circles.

TYPE: RadiusArrayLike | None DEFAULT: None

colors

Optional colors for the points.

The colors are interpreted as RGB or RGBA in sRGB gamma-space, As either 0-1 floats or 0-255 integers, with separate alpha.

TYPE: Rgba32ArrayLike | None DEFAULT: None

labels

Optional text labels for the points.

TYPE: Utf8ArrayLike | None DEFAULT: None

class_ids

Optional class Ids for the points.

The class ID provides colors and labels if not specified explicitly.

TYPE: ClassIdArrayLike | None DEFAULT: None

keypoint_ids

Optional keypoint IDs for the points, identifying them within a class.

If keypoint IDs are passed in but no class IDs were specified, the class ID will default to 0. This is useful to identify points within a single classification (which is identified with class_id). E.g. the classification might be 'Person' and the keypoints refer to joints on a detected skeleton.

TYPE: KeypointIdArrayLike | None DEFAULT: None

instance_keys

Unique identifiers for each individual point in the batch.

TYPE: InstanceKeyArrayLike | None DEFAULT: None

class Scalar

Bases: Archetype

Archetype: Log a double-precision scalar.

The current timeline value will be used for the time/X-axis, hence scalars cannot be timeless.

When used to produce a plot, this archetype is used to provide the data that is referenced by the SeriesLine or SeriesPoint archetypes. You can do this by logging both archetypes to the same path, or alternatively configuring the plot-specific archetypes through the blueprint.

See also SeriesPoint, SeriesLine.

Example
Simple line plot:

import math

import rerun as rr

rr.init("rerun_example_scalar", spawn=True)

# Log the data on a timeline called "step".
for step in range(0, 64):
    rr.set_time_sequence("step", step)
    rr.log("scalar", rr.Scalar(math.sin(step / 10.0)))

def __init__(scalar)

Create a new instance of the Scalar archetype.

PARAMETER DESCRIPTION
scalar

The scalar value to log.

TYPE: ScalarLike

class SegmentationImage

Bases: SegmentationImageExt, Archetype

Archetype: An image made up of integer class-ids.

The shape of the TensorData must be mappable to an HxW tensor. Each pixel corresponds to a class-id that will be mapped to a color based on annotation context.

In the case of floating point images, the label will be looked up based on rounding to the nearest integer value.

Leading and trailing unit-dimensions are ignored, so that 1x640x480x1 is treated as a 640x480 image.

See also AnnotationContext to associate each class with a color and a label.

Example
Simple segmentation image:

import numpy as np
import rerun as rr

# Create a segmentation image
image = np.zeros((8, 12), dtype=np.uint8)
image[0:4, 0:6] = 1
image[4:8, 6:12] = 2

rr.init("rerun_example_segmentation_image", spawn=True)

# Assign a label and color to each class
rr.log("/", rr.AnnotationContext([(1, "red", (255, 0, 0)), (2, "green", (0, 255, 0))]), timeless=True)

rr.log("image", rr.SegmentationImage(image))

def __init__(data, *, draw_order=None)

Create a new instance of the SegmentationImage archetype.

PARAMETER DESCRIPTION
data

The image data. Should always be a rank-2 tensor.

TYPE: TensorDataLike

draw_order

An optional floating point value that specifies the 2D drawing order.

Objects with higher values are drawn on top of those with lower values.

TYPE: DrawOrderLike | None DEFAULT: None

class SeriesLine

Bases: Archetype

Archetype: Define the style properties for a line series in a chart.

This archetype only provides styling information and should be logged as timeless when possible. The underlying data needs to be logged to the same entity-path using the Scalar archetype.

See Scalar

Example
Series Line:

from math import cos, sin, tau

import rerun as rr

rr.init("rerun_example_series_line_style", spawn=True)

# Set up plot styling:
# They are logged timeless as they don't change over time and apply to all timelines.
# Log two lines series under a shared root so that they show in the same plot by default.
rr.log("trig/sin", rr.SeriesLine(color=[255, 0, 0], name="sin(0.01t)", width=2), timeless=True)
rr.log("trig/cos", rr.SeriesLine(color=[0, 255, 0], name="cos(0.01t)", width=4), timeless=True)

# Log the data on a timeline called "step".
for t in range(0, int(tau * 2 * 100.0)):
    rr.set_time_sequence("step", t)

    rr.log("trig/sin", rr.Scalar(sin(float(t) / 100.0)))
    rr.log("trig/cos", rr.Scalar(cos(float(t) / 100.0)))

def __init__(*, color=None, width=None, name=None)

Create a new instance of the SeriesLine archetype.

PARAMETER DESCRIPTION
color

Color for the corresponding series.

TYPE: Rgba32Like | None DEFAULT: None

width

Stroke width for the corresponding series.

TYPE: StrokeWidthLike | None DEFAULT: None

name

Display name of the series.

Used in the legend.

TYPE: Utf8Like | None DEFAULT: None

class SeriesPoint

Bases: Archetype

Archetype: Define the style properties for a point series in a chart.

This archetype only provides styling information and should be logged as timeless when possible. The underlying data needs to be logged to the same entity-path using the Scalar archetype.

See Scalar

Example
Series Point:

from math import cos, sin, tau

import rerun as rr

rr.init("rerun_example_series_point_style", spawn=True)

# Set up plot styling:
# They are logged timeless as they don't change over time and apply to all timelines.
# Log two point series under a shared root so that they show in the same plot by default.
rr.log(
    "trig/sin",
    rr.SeriesPoint(
        color=[255, 0, 0],
        name="sin(0.01t)",
        marker="circle",
        marker_size=4,
    ),
    timeless=True,
)
rr.log(
    "trig/cos",
    rr.SeriesPoint(
        color=[0, 255, 0],
        name="cos(0.01t)",
        marker="cross",
        marker_size=2,
    ),
    timeless=True,
)

# Log the data on a timeline called "step".
for t in range(0, int(tau * 2 * 10.0)):
    rr.set_time_sequence("step", t)

    rr.log("trig/sin", rr.Scalar(sin(float(t) / 10.0)))
    rr.log("trig/cos", rr.Scalar(cos(float(t) / 10.0)))

def __init__(*, color=None, marker=None, name=None, marker_size=None)

Create a new instance of the SeriesPoint archetype.

PARAMETER DESCRIPTION
color

Color for the corresponding series.

TYPE: Rgba32Like | None DEFAULT: None

marker

What shape to use to represent the point

TYPE: MarkerShapeLike | None DEFAULT: None

name

Display name of the series.

Used in the legend.

TYPE: Utf8Like | None DEFAULT: None

marker_size

Size of the marker.

TYPE: MarkerSizeLike | None DEFAULT: None

class Tensor

Bases: TensorExt, Archetype

Archetype: A generic n-dimensional Tensor.

Example
Simple Tensor:

import numpy as np
import rerun as rr

tensor = np.random.randint(0, 256, (8, 6, 3, 5), dtype=np.uint8)  # 4-dimensional tensor

rr.init("rerun_example_tensor", spawn=True)

# Log the tensor, assigning names to each dimension
rr.log("tensor", rr.Tensor(tensor, dim_names=("width", "height", "channel", "batch")))

def __init__(data=None, *, dim_names=None)

Construct a Tensor archetype.

The Tensor archetype internally contains a single component: TensorData.

See the TensorData constructor for more advanced options to interpret buffers as TensorData of varying shapes.

For simple cases, you can pass array objects and optionally specify the names of the dimensions. The shape of the TensorData will be inferred from the array.

PARAMETER DESCRIPTION
self

The TensorData object to construct.

TYPE: Any

data

A TensorData object, or type that can be converted to a numpy array.

TYPE: TensorDataLike | TensorLike | None DEFAULT: None

dim_names

The names of the tensor dimensions when generating the shape from an array.

TYPE: Sequence[str | None] | None DEFAULT: None

class TextDocument

Bases: Archetype

Archetype: A text element intended to be displayed in its own text-box.

Supports raw text and markdown.

Example
Markdown text document:

import rerun as rr

rr.init("rerun_example_text_document", spawn=True)

rr.log("text_document", rr.TextDocument("Hello, TextDocument!"))

rr.log(
    "markdown",
    rr.TextDocument(
        '''
# Hello Markdown!
[Click here to see the raw text](recording://markdown:Text).

Basic formatting:

| **Feature**       | **Alternative** |
| ----------------- | --------------- |
| Plain             |                 |
| *italics*         | _italics_       |
| **bold**          | __bold__        |
| ~~strikethrough~~ |                 |
| `inline code`     |                 |

----------------------------------

## Support
- [x] [Commonmark](https://commonmark.org/help/) support
- [x] GitHub-style strikethrough, tables, and checkboxes
- Basic syntax highlighting for:
  - [x] C and C++
  - [x] Python
  - [x] Rust
  - [ ] Other languages

## Links
You can link to [an entity](recording://markdown),
a [specific instance of an entity](recording://markdown[#0]),
or a [specific component](recording://markdown:Text).

Of course you can also have [normal https links](https://github.com/rerun-io/rerun), e.g. <https://rerun.io>.

## Image
![A random image](https://picsum.photos/640/480)
'''.strip(),
        media_type=rr.MediaType.MARKDOWN,
    ),
)

def __init__(text, *, media_type=None)

Create a new instance of the TextDocument archetype.

PARAMETER DESCRIPTION
text

Contents of the text document.

TYPE: Utf8Like

media_type

The Media Type of the text.

For instance: * text/plain * text/markdown

If omitted, text/plain is assumed.

TYPE: Utf8Like | None DEFAULT: None

class TextLog

Bases: Archetype

Archetype: A log entry in a text log, comprised of a text body and its log level.

Example
text_log_integration:

import logging

import rerun as rr

rr.init("rerun_example_text_log_integration", spawn=True)

# Log a text entry directly
rr.log("logs", rr.TextLog("this entry has loglevel TRACE", level=rr.TextLogLevel.TRACE))

# Or log via a logging handler
logging.getLogger().addHandler(rr.LoggingHandler("logs/handler"))
logging.getLogger().setLevel(-1)
logging.info("This INFO log got added through the standard logging interface")

def __init__(text, *, level=None, color=None)

Create a new instance of the TextLog archetype.

PARAMETER DESCRIPTION
text

The body of the message.

TYPE: Utf8Like

level

The verbosity level of the message.

This can be used to filter the log messages in the Rerun Viewer.

TYPE: Utf8Like | None DEFAULT: None

color

Optional color to use for the log line in the Rerun Viewer.

TYPE: Rgba32Like | None DEFAULT: None

class TimeSeriesScalar

Bases: Archetype

Archetype: Log a double-precision scalar that will be visualized as a time-series plot.

The current simulation time will be used for the time/X-axis, hence scalars cannot be timeless!

This archetype is in the process of being deprecated. Prefer usage of Scalar, SeriesLine, and SeriesPoint instead.

See also Scalar, SeriesPoint, SeriesLine.

Example
Simple line plot:

import math

import rerun as rr

rr.init("rerun_example_scalar", spawn=True)

# Log the data on a timeline called "step".
for step in range(0, 64):
    rr.set_time_sequence("step", step)
    rr.log("scalar", rr.Scalar(math.sin(step / 10.0)))

def __init__(scalar, *, radius=None, color=None, label=None, scattered=None)

Create a new instance of the TimeSeriesScalar archetype.

PARAMETER DESCRIPTION
scalar

The scalar value to log.

TYPE: ScalarLike

radius

An optional radius for the point.

Points within a single line do not have to share the same radius, the line will have differently sized segments as appropriate.

If all points within a single entity path (i.e. a line) share the same radius, then this radius will be used as the line width too. Otherwise, the line will use the default width of 1.0.

TYPE: RadiusLike | None DEFAULT: None

color

Optional color for the scalar entry.

If left unspecified, a pseudo-random color will be used instead. That same color will apply to all points residing in the same entity path that don't have a color specified.

Points within a single line do not have to share the same color, the line will have differently colored segments as appropriate. If all points within a single entity path (i.e. a line) share the same color, then this color will be used as the line color in the plot legend. Otherwise, the line will appear gray in the legend.

TYPE: Rgba32Like | None DEFAULT: None

label

An optional label for the point.

TODO(#1289): This won't show up on points at the moment, as our plots don't yet support displaying labels for individual points. If all points within a single entity path (i.e. a line) share the same label, then this label will be used as the label for the line itself. Otherwise, the line will be named after the entity path. The plot itself is named after the space it's in.

TYPE: Utf8Like | None DEFAULT: None

scattered

Specifies whether a point in a scatter plot should form a continuous line.

If set to true, this scalar will be drawn as a point, akin to a scatterplot. Otherwise, it will form a continuous line with its neighbors. Points within a single line do not have to all share the same scatteredness: the line will switch between a scattered and a continuous representation as required.

TYPE: ScalarScatteringLike | None DEFAULT: None

class Transform3D

Bases: Transform3DExt, Archetype

Archetype: A 3D transform.

Example
Variety of 3D transforms:

from math import pi

import rerun as rr
from rerun.datatypes import Angle, RotationAxisAngle

rr.init("rerun_example_transform3d", spawn=True)

arrow = rr.Arrows3D(origins=[0, 0, 0], vectors=[0, 1, 0])

rr.log("base", arrow)

rr.log("base/translated", rr.Transform3D(translation=[1, 0, 0]))
rr.log("base/translated", arrow)

rr.log(
    "base/rotated_scaled",
    rr.Transform3D(
        rotation=RotationAxisAngle(axis=[0, 0, 1], angle=Angle(rad=pi / 4)),
        scale=2,
    ),
)
rr.log("base/rotated_scaled", arrow)

def __init__(transform=None, *, translation=None, rotation=None, scale=None, mat3x3=None, from_parent=False)

Create a new instance of the Transform3D archetype.

PARAMETER DESCRIPTION
transform

Transform using an existing Transform3D datatype object. If not provided, none of the other named parameters must be set.

TYPE: Transform3DLike | None DEFAULT: None

translation

3D translation vector, applied last. Not compatible with transform.

TYPE: Vec3DLike | None DEFAULT: None

rotation

3D rotation, applied second. Not compatible with transform and mat3x3 parameters.

TYPE: Rotation3DLike | None DEFAULT: None

scale

3D scale, applied last. Not compatible with transform and mat3x3 parameters.

TYPE: Scale3DLike | None DEFAULT: None

mat3x3

3x3 matrix representing scale and rotation, applied after translation. Not compatible with rotation and scale parameters. TODO(#3559): Support 4x4 and 4x3 matrices.

TYPE: Mat3x3Like | None DEFAULT: None

from_parent

If true, the transform maps from the parent space to the space where the transform was logged. Otherwise, the transform maps from the space to its parent.

TYPE: bool DEFAULT: False

class ViewCoordinates

Bases: ViewCoordinatesExt, Archetype

Archetype: How we interpret the coordinate system of an entity/space.

For instance: What is "up"? What does the Z axis mean? Is this right-handed or left-handed?

The three coordinates are always ordered as [x, y, z].

For example [Right, Down, Forward] means that the X axis points to the right, the Y axis points down, and the Z axis points forward.

Example
View coordinates for adjusting the eye camera:

import rerun as rr

rr.init("rerun_example_view_coordinates", spawn=True)

rr.log("world", rr.ViewCoordinates.RIGHT_HAND_Z_UP, timeless=True)  # Set an up-axis
rr.log(
    "world/xyz",
    rr.Arrows3D(
        vectors=[[1, 0, 0], [0, 1, 0], [0, 0, 1]],
        colors=[[255, 0, 0], [0, 255, 0], [0, 0, 255]],
    ),
)

def __init__(xyz)

Create a new instance of the ViewCoordinates archetype.