Struct DefaultScalars

Source

pub struct DefaultScalars { /* private fields */ }

Expand description

Dynamic host tensor over the crate’s preset scalar set.

The payload is a private inline value: a host tensor is one move and a copy allocates nothing, and the preset variants are not part of the public surface. Construct through DefaultScalars::from_vec_col_major and read through DefaultScalars::as_slice, DefaultScalars::as_mut_slice, or DefaultScalars::into_vec_col_major.

§Examples

use tenferro_tensor::{DefaultScalars, DType};
use tenferro_tensor::ScalarSet;

let value = DefaultScalars::from_vec_col_major(vec![1], vec![7_i32])?;
assert_eq!(value.tag(), DType::I32);
assert_eq!(value.as_slice::<i32>()?, &[7]);

Implementations§

Source §

impl DefaultScalars

Source

pub fn from_vec_col_major<T: TensorScalar>( shape: impl Into<Vec<usize>>, data: Vec<T>, ) -> Result<Self>

Create a dynamic tensor from a column-major host buffer.

§Examples

use tenferro_tensor::{DefaultScalars, DType};

let tensor = DefaultScalars::from_vec_col_major(vec![1], vec![2.0_f32])?;
assert_eq!(tensor.dtype(), DType::F32);

§Errors

Returns a validation error carrying tenferro_tensor_core::ValidationError::ShapeDataLengthMismatch when the shape product differs from data.len(), or tenferro_tensor_core::ValidationError::IntegerOverflow when validating the shape overflows.

Source

pub fn dtype(&self) -> DType

Return the tensor dtype tag.

§Examples

use tenferro_tensor::{DefaultScalars, DType};

let tensor = DefaultScalars::from_vec_col_major(vec![1], vec![false])?;
assert_eq!(tensor.dtype(), DType::Bool);

Source

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

Borrow the tensor shape.

§Examples

use tenferro_tensor::DefaultScalars;

let tensor = DefaultScalars::from_vec_col_major(vec![2], vec![1_i32, 2])?;
assert_eq!(tensor.shape(), &[2]);

Source

pub fn rank(&self) -> usize

Return the tensor rank.

§Examples

use tenferro_tensor::DefaultScalars;

let tensor = DefaultScalars::from_vec_col_major(vec![1, 1], vec![1_i64])?;
assert_eq!(tensor.rank(), 2);

Source

pub fn is_empty(&self) -> bool

Return whether the tensor has zero elements.

§Examples

use tenferro_tensor::DefaultScalars;

let tensor = DefaultScalars::from_vec_col_major(vec![0], Vec::<f64>::new())?;
assert!(tensor.is_empty());

Source

pub fn as_slice<T: TensorScalar>(&self) -> Result<&[T]>

Borrow the typed host slice when the dtype matches.

§Examples

use tenferro_tensor::DefaultScalars;

let tensor = DefaultScalars::from_vec_col_major(vec![1], vec![3.0_f64])?;
assert_eq!(tensor.as_slice::<f64>()?, &[3.0]);
assert!(tensor.as_slice::<f32>().is_err());

§Errors

Returns crate::Error::Validation with a dtype mismatch when T does not match the tensor’s runtime dtype.

Source

pub fn as_mut_slice<T: TensorScalar>(&mut self) -> Result<&mut [T]>

Mutably borrow the typed host slice when the dtype matches.

§Examples

use tenferro_tensor::DefaultScalars;

let mut tensor = DefaultScalars::from_vec_col_major(vec![1], vec![3.0_f64])?;
tensor.as_mut_slice::<f64>()?[0] = 4.0;
assert_eq!(tensor.as_slice::<f64>()?, &[4.0]);

§Errors

Returns crate::Error::Validation with a dtype mismatch when T does not match the tensor’s runtime dtype.

Source

pub fn as_view(&self) -> DefaultScalarsView<'_>

Borrow this tensor as a dynamic zero-offset view.

§Examples

use tenferro_tensor::{DefaultScalars, DType};

let tensor = DefaultScalars::from_vec_col_major(vec![1], vec![1.0_f32])?;
assert_eq!(tensor.as_view().dtype(), DType::F32);

Source

pub fn into_vec_col_major<T: TensorScalar>(self) -> Result<(Vec<usize>, Vec<T>)>

Consume this tensor and return typed column-major data when the dtype matches.

§Examples

use tenferro_tensor::DefaultScalars;

let tensor = DefaultScalars::from_vec_col_major(vec![2], vec![1.0_f64, 2.0])?;
let (shape, data) = tensor.into_vec_col_major::<f64>()?;
assert_eq!(shape, vec![2]);
assert_eq!(data, vec![1.0, 2.0]);