Merge pull request #976 from saona-raimundo/dev

Implement conversion methods for RowDVector
This commit is contained in:
Sébastien Crozet 2021-08-28 13:54:33 +02:00 committed by GitHub
commit 80c7064bf4
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
3 changed files with 53 additions and 24 deletions

View File

@ -4,6 +4,12 @@ documented here.
This project adheres to [Semantic Versioning](https://semver.org/). This project adheres to [Semantic Versioning](https://semver.org/).
## [0.29.1] - WIP
### Added
- The conversion trait `From<Vec<T>>` and method `from_vec_storage` for `RowDVector`. See [#975](https://github.com/dimforge/nalgebra/issues/975)
## [0.29.0] ## [0.29.0]
### Breaking changes ### Breaking changes
- We updated to the version 0.6 of `simba`. This means that the trait bounds `T: na::RealField`, `na::ComplexField`, - We updated to the version 0.6 of `simba`. This means that the trait bounds `T: na::RealField`, `na::ComplexField`,
@ -242,7 +248,7 @@ for details about this switch and its benefits.
geometric types. geometric types.
### Modified ### Modified
* Use of traits like `alga::general::{RealField, ComplexField}` have now been replaced by * Use of traits like `alga::general::{RealField, ComplexField}` have now been replaced by
`simba::scalar::{RealField, ComplexField}`. `simba::scalar::{RealField, ComplexField}`.
* The implementation of traits from the __alga__ crate (and well as the dependency to _alga__) are now * The implementation of traits from the __alga__ crate (and well as the dependency to _alga__) are now
omitted unless the `alga` cargo feature is activated. omitted unless the `alga` cargo feature is activated.
### Removed ### Removed
@ -264,16 +270,16 @@ for details about this switch and its benefits.
### Added ### Added
* `.remove_rows_at` and `remove_columns_at` which removes a set of rows or columns (specified by indices) from a matrix. * `.remove_rows_at` and `remove_columns_at` which removes a set of rows or columns (specified by indices) from a matrix.
* Several formatting traits have been implemented for all matrices/vectors: `LowerExp`, `UpperExp`, `Octal`, `LowerHex`, * Several formatting traits have been implemented for all matrices/vectors: `LowerExp`, `UpperExp`, `Octal`, `LowerHex`,
`UpperHex`, `Binary`, `Pointer`. `UpperHex`, `Binary`, `Pointer`.
* `UnitQuaternion::quaternions_mean(...)` which computes the mean rotation of a set of unit quaternions. This implements * `UnitQuaternion::quaternions_mean(...)` which computes the mean rotation of a set of unit quaternions. This implements
the algorithm from _Oshman, Yaakov, and Avishy Carmi, "Attitude estimation from vector observations using a genetic-algorithm-embedded quaternion particle filter." the algorithm from _Oshman, Yaakov, and Avishy Carmi, "Attitude estimation from vector observations using a genetic-algorithm-embedded quaternion particle filter."
### Modified ### Modified
* It is now possible to get the `min/max` element of unsigned integer matrices. * It is now possible to get the `min/max` element of unsigned integer matrices.
### Added to nalgebra-glm ### Added to nalgebra-glm
* Some infinite and reversed perspectives: `::infinite_perspective_rh_no`, `::infinite_perspective_rh_zo`, * Some infinite and reversed perspectives: `::infinite_perspective_rh_no`, `::infinite_perspective_rh_zo`,
`::reversed_perspective_rh_zo`, and `::reversed_infinite_perspective_rh_zo`. `::reversed_perspective_rh_zo`, and `::reversed_infinite_perspective_rh_zo`.
## [0.18.0] ## [0.18.0]
This release adds full complex number support to nalgebra. This includes all common vector/matrix operations as well This release adds full complex number support to nalgebra. This includes all common vector/matrix operations as well
@ -329,22 +335,22 @@ Note that all the other BLAS operation will continue to work for all fields, inc
* Add `.slerp` and `.try_slerp` to unit vectors. * Add `.slerp` and `.try_slerp` to unit vectors.
* Add `.lerp` to vectors. * Add `.lerp` to vectors.
* Add `.to_projective` and `.as_projective` to `Perspective3` and `Orthographic3` in order to * Add `.to_projective` and `.as_projective` to `Perspective3` and `Orthographic3` in order to
use them as `Projective3` structures. use them as `Projective3` structures.
* Add `From/Into` impls to allow the conversion of any transformation type to a matrix. * Add `From/Into` impls to allow the conversion of any transformation type to a matrix.
* Add `Into` impls to convert a matrix slice into an owned matrix. * Add `Into` impls to convert a matrix slice into an owned matrix.
* Add `Point*::from_slice` to create a point from a slice. * Add `Point*::from_slice` to create a point from a slice.
* Add `.map_with_location` to matrices to apply a map which passes the component indices to the user-defined closure alongside * Add `.map_with_location` to matrices to apply a map which passes the component indices to the user-defined closure alongside
the component itself. the component itself.
* Add impl `From<Vector>` for `Point`. * Add impl `From<Vector>` for `Point`.
* Add impl `From<Vector4>` for `Quaternion`. * Add impl `From<Vector4>` for `Quaternion`.
* Add impl `From<Vector>` for `Translation`. * Add impl `From<Vector>` for `Translation`.
* Add the `::from_vec` constructor to construct a matrix from a `Vec` (a `DMatrix` will reuse the original `Vec` * Add the `::from_vec` constructor to construct a matrix from a `Vec` (a `DMatrix` will reuse the original `Vec`
as-is for its storage). as-is for its storage).
* Add `.to_homogeneous` to square matrices (and with dimensions higher than 1x1). This will increase their number of row * Add `.to_homogeneous` to square matrices (and with dimensions higher than 1x1). This will increase their number of row
and columns by 1. The new column and row are filled with 0, except for the diagonal element which is set to 1. and columns by 1. The new column and row are filled with 0, except for the diagonal element which is set to 1.
* Implement `Extend<Vec>` for matrices with a dynamic storage. The provided `Vec` is assumed to represent a column-major * Implement `Extend<Vec>` for matrices with a dynamic storage. The provided `Vec` is assumed to represent a column-major
matrix with the same number of rows as the one being extended. This will effectively append new columns on the right of matrix with the same number of rows as the one being extended. This will effectively append new columns on the right of
the matrix being extended. the matrix being extended.
* Implement `Extend<Vec>` for vectors with a dynamic storage. This will concatenate the vector with the given `Vec`. * Implement `Extend<Vec>` for vectors with a dynamic storage. This will concatenate the vector with the given `Vec`.
* Implement `Extend<Matrix<...>>` for matrices with dynamic storage. This will concatenate the columns of both matrices. * Implement `Extend<Matrix<...>>` for matrices with dynamic storage. This will concatenate the columns of both matrices.
* Implement `Into<Vec>` for the `MatrixVec` storage. * Implement `Into<Vec>` for the `MatrixVec` storage.
@ -353,10 +359,10 @@ Note that all the other BLAS operation will continue to work for all fields, inc
### Modified ### Modified
* The orthographic projection no longer require that `bottom < top`, that `left < right`, and that `znear < zfar`. The * The orthographic projection no longer require that `bottom < top`, that `left < right`, and that `znear < zfar`. The
only restriction now ith that they must not be equal (in which case the projection would be singular). only restriction now ith that they must not be equal (in which case the projection would be singular).
* The `Point::from_coordinates` methods is deprecated. Use `Point::from` instead. * The `Point::from_coordinates` methods is deprecated. Use `Point::from` instead.
* The `.transform_point` and `.transform_vector` methods are now inherent methods for matrices so that the user does not have to * The `.transform_point` and `.transform_vector` methods are now inherent methods for matrices so that the user does not have to
explicitly import the `Transform` trait from the alga crate. explicitly import the `Transform` trait from the alga crate.
* Renamed the matrix storage types: `MatrixArray` -> `ArrayStorage` and `MatrixVec` -> `VecStorage`. * Renamed the matrix storage types: `MatrixArray` -> `ArrayStorage` and `MatrixVec` -> `VecStorage`.
* Renamed `.unwrap()` to `.into_inner()` for geometric types that wrap another type. * Renamed `.unwrap()` to `.into_inner()` for geometric types that wrap another type.
This is for the case of `Unit`, `Transform`, `Orthographic3`, `Perspective3`, `Rotation`. This is for the case of `Unit`, `Transform`, `Orthographic3`, `Perspective3`, `Rotation`.
@ -552,8 +558,8 @@ overview of all the added/modified features.
This version is a major rewrite of the library. Major changes are: This version is a major rewrite of the library. Major changes are:
* Algebraic traits are now defined by the [alga](https://crates.io/crates/alga) crate. * Algebraic traits are now defined by the [alga](https://crates.io/crates/alga) crate.
All other mathematical traits, except `Axpy` have been removed from All other mathematical traits, except `Axpy` have been removed from
**nalgebra**. **nalgebra**.
* Methods are now preferred to free functions because they do not require any * Methods are now preferred to free functions because they do not require any
trait to be used anymore. trait to be used anymore.
* Most algebraic entities can be parametrized by type-level integers * Most algebraic entities can be parametrized by type-level integers
@ -746,11 +752,11 @@ you [there](https://users.nphysics.org)!
### Modified ### Modified
* Vectors are now multipliable with isometries. This will result into a pure rotation (this is how * Vectors are now multipliable with isometries. This will result into a pure rotation (this is how
vectors differ from point semantically: they design directions, so they are not translatable). vectors differ from point semantically: they design directions, so they are not translatable).
* `{Isometry3, Rotation3}::look_at` reimplemented and renamed to `::look_at_rh` and `::look_at_lh` to agree * `{Isometry3, Rotation3}::look_at` reimplemented and renamed to `::look_at_rh` and `::look_at_lh` to agree
with the computer graphics community (in particular, the GLM library). Use the `::look_at_rh` with the computer graphics community (in particular, the GLM library). Use the `::look_at_rh`
variant to build a view matrix that variant to build a view matrix that
may be successfully used with `Persp` and `Ortho`. may be successfully used with `Persp` and `Ortho`.
* The old `{Isometry3, Rotation3}::look_at` implementations are now called `::new_observer_frame`. * The old `{Isometry3, Rotation3}::look_at` implementations are now called `::new_observer_frame`.
* Rename every `fov` on `Persp` to `fovy`. * Rename every `fov` on `Persp` to `fovy`.
* Fixed the perspective and orthographic projection matrices. * Fixed the perspective and orthographic projection matrices.

View File

@ -20,7 +20,7 @@ use crate::base::{
MatrixSliceMut, OMatrix, Scalar, MatrixSliceMut, OMatrix, Scalar,
}; };
#[cfg(any(feature = "std", feature = "alloc"))] #[cfg(any(feature = "std", feature = "alloc"))]
use crate::base::{DVector, VecStorage}; use crate::base::{DVector, RowDVector, VecStorage};
use crate::base::{SliceStorage, SliceStorageMut}; use crate::base::{SliceStorage, SliceStorageMut};
use crate::constraint::DimEq; use crate::constraint::DimEq;
use crate::{IsNotStaticOne, RowSVector, SMatrix, SVector}; use crate::{IsNotStaticOne, RowSVector, SMatrix, SVector};
@ -453,6 +453,14 @@ impl<'a, T: Scalar> From<Vec<T>> for DVector<T> {
} }
} }
#[cfg(any(feature = "std", feature = "alloc"))]
impl<'a, T: Scalar> From<Vec<T>> for RowDVector<T> {
#[inline]
fn from(vec: Vec<T>) -> Self {
Self::from_vec(vec)
}
}
impl<'a, T: Scalar + Copy, R: Dim, C: Dim, S: RawStorage<T, R, C> + IsContiguous> impl<'a, T: Scalar + Copy, R: Dim, C: Dim, S: RawStorage<T, R, C> + IsContiguous>
From<&'a Matrix<T, R, C, S>> for &'a [T] From<&'a Matrix<T, R, C, S>> for &'a [T]
{ {

View File

@ -32,7 +32,7 @@ use crate::{ArrayStorage, SMatrix, SimdComplexField, Storage, UninitMatrix};
use crate::storage::IsContiguous; use crate::storage::IsContiguous;
use crate::uninit::{Init, InitStatus, Uninit}; use crate::uninit::{Init, InitStatus, Uninit};
#[cfg(any(feature = "std", feature = "alloc"))] #[cfg(any(feature = "std", feature = "alloc"))]
use crate::{DMatrix, DVector, Dynamic, VecStorage}; use crate::{DMatrix, DVector, Dynamic, RowDVector, VecStorage};
use std::mem::MaybeUninit; use std::mem::MaybeUninit;
/// A square matrix. /// A square matrix.
@ -411,6 +411,21 @@ impl<T> DVector<T> {
} }
} }
// TODO: Consider removing/deprecating `from_vec_storage` once we are able to make
// `from_data` const fn compatible
#[cfg(any(feature = "std", feature = "alloc"))]
impl<T> RowDVector<T> {
/// Creates a new heap-allocated matrix from the given [`VecStorage`].
///
/// This method exists primarily as a workaround for the fact that `from_data` can not
/// work in `const fn` contexts.
pub const fn from_vec_storage(storage: VecStorage<T, U1, Dynamic>) -> Self {
// This is sound because the dimensions of the matrix and the storage are guaranteed
// to be the same
unsafe { Self::from_data_statically_unchecked(storage) }
}
}
impl<T, R: Dim, C: Dim> UninitMatrix<T, R, C> impl<T, R: Dim, C: Dim> UninitMatrix<T, R, C>
where where
DefaultAllocator: Allocator<T, R, C>, DefaultAllocator: Allocator<T, R, C>,