Struct rerun::components::TensorData

#[repr(transparent)]
pub struct TensorData(pub TensorData);

Component: A multi-dimensional Tensor of data.

The number of dimensions and their respective lengths is specified by the shape field. The dimensions are ordered from outermost to innermost. For example, in the common case of a 2D RGB Image, the shape would be [height, width, channel].

These dimensions are combined with an index to look up values from the buffer field, which stores a contiguous array of typed values.

Note that the buffer may be encoded in a compressed format such as jpeg or in a format with downsampled chroma, such as NV12 or YUY2. For file formats, the shape is used as a hint, for chroma downsampled format the shape has to be the shape of the decoded image.

Tuple Fields

0: TensorData

Methods from Deref<Target = TensorData>

pub fn shape(&self) -> &[TensorDimension]

The shape of the tensor, including optional dimension names.

pub fn shape_short(&self) -> &[TensorDimension]

Returns the shape of the tensor with all leading & trailing dimensions of size 1 ignored.

If all dimension sizes are one, this returns only the first dimension.

pub fn num_dim(&self) -> usize

The number of dimensions of the tensor.

An image tensor will usually have two (height, width) or three (height, width, channels) dimensions.

pub fn image_height_width_channels(&self) -> Option<[u64; 3]>

If the tensor can be interpreted as an image, return the height, width, and channels/depth of it.

pub fn is_shaped_like_an_image(&self) -> bool

Returns true if the tensor can be interpreted as an image.

pub fn is_vector(&self) -> bool

Returns true if either all dimensions have size 1 or only a single dimension has a size larger than 1.

Empty tensors return false.

pub fn get_with_image_coords( &self, x: u64, y: u64, channel: u64 ) -> Option<TensorElement>

Query with x, y, channel indices.

Allows to query values for any image-like tensor even if it has more or less dimensions than 3. (useful for sampling e.g. N x M x C x 1 tensor which is a valid image)

pub fn get(&self, index: &[u64]) -> Option<TensorElement>

Get the value of the element at the given index.

Return None if out-of-bounds, or if the tensor is encoded (e.g. TensorBuffer::Jpeg).

pub fn get_nv12_pixel(&self, x: u64, y: u64) -> Option<[u8; 3]>

Returns decoded RGB8 value at the given image coordinates if this tensor is a NV12 image.

If the tensor is not TensorBuffer::Nv12, None is returned.

It is undefined what happens if the coordinate is out-of-bounds.

pub fn get_yuy2_pixel(&self, x: u64, y: u64) -> Option<[u8; 3]>

Returns decoded RGB8 value at the given image coordinates if this tensor is a YUY2 image.

If the tensor is not TensorBuffer::Yuy2, None is returned.

It is undefined what happens if the coordinate is out-of-bounds.

pub fn dtype(&self) -> TensorDataType

The datatype of the tensor.

pub fn size_in_bytes(&self) -> usize

The size of the tensor data, in bytes.

Trait Implementations

impl Borrow<TensorData> for TensorData

fn borrow(&self) -> &TensorData

Immutably borrows from an owned value.

impl Clone for TensorData

fn clone(&self) -> TensorData

Returns a copy of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for TensorData

fn fmt(&self, f: &mut Formatter<'_>) -> Result<(), Error>

Formats the value using the given formatter.

impl Deref for TensorData

type Target = TensorData

The resulting type after dereferencing.

fn deref(&self) -> &TensorData

Dereferences the value.

impl<T> From<T> for TensorData

where T: Into<TensorData>,

fn from(v: T) -> TensorData

Converts to this type from the input type.

impl Loggable for TensorData

type Name = ComponentName

fn name() -> <TensorData as Loggable>::Name

The fully-qualified name of this loggable, e.g. rerun.datatypes.Vec2D.

fn arrow_datatype() -> DataType

The underlying arrow2::datatypes::DataType, excluding datatype extensions.

fn to_arrow_opt<'a>( data: impl IntoIterator<Item = Option<impl Into<Cow<'a, TensorData>>>> ) -> Result<Box<dyn Array>, SerializationError>

where TensorData: Clone + 'a,

Given an iterator of options of owned or reference values to the current Loggable, serializes them into an Arrow array. The Arrow array’s datatype will match Loggable::arrow_field.

fn from_arrow_opt( arrow_data: &(dyn Array + 'static) ) -> Result<Vec<Option<TensorData>>, DeserializationError>

where TensorData: Sized,

Given an Arrow array, deserializes it into a collection of optional Loggables.

fn extended_arrow_datatype() -> DataType

The underlying arrow2::datatypes::DataType, including datatype extensions.

fn arrow_field() -> Field

The underlying arrow2::datatypes::Field, including datatype extensions.

fn to_arrow<'a>( data: impl IntoIterator<Item = impl Into<Cow<'a, Self>>> ) -> Result<Box<dyn Array>, SerializationError>

where Self: 'a,

Given an iterator of owned or reference values to the current Loggable, serializes them into an Arrow array. The Arrow array’s datatype will match Loggable::arrow_field.

fn from_arrow( data: &(dyn Array + 'static) ) -> Result<Vec<Self>, DeserializationError>

Given an Arrow array, deserializes it into a collection of Loggables.

impl PartialEq for TensorData

fn eq(&self, other: &TensorData) -> bool

This method tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

This method tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl SizeBytes for TensorData

fn heap_size_bytes(&self) -> u64

Returns the total size of self on the heap, in bytes.

fn is_pod() -> bool

Is Self just plain old data?

fn total_size_bytes(&self) -> u64

Returns the total size of self in bytes, accounting for both stack and heap space.

fn stack_size_bytes(&self) -> u64

Returns the total size of self on the stack, in bytes.

impl StructuralPartialEq for TensorData