nabled 0.0.8

A linear algebra library written in Rust with an ndarray-first architecture.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129

//! Arrow adapters for Cholesky decomposition workflows.

use arrow_array::types::{Float32Type, Float64Type};
use arrow_array::{FixedSizeListArray, PrimitiveArray};
use arrow_schema::Field;

use super::{
    ArrowInteropError, complex64_matrix_from_owned, complex64_matrix_view,
    complex64_vector_from_owned, complex64_vector_view, fixed_size_list_from_owned,
    fixed_size_list_view, primitive_array_from_owned, primitive_array_view,
};

/// Compute `f32` Cholesky decomposition directly from an Arrow dense matrix.
///
/// The inbound Arrow -> ndarray bridge is zero-copy. The decomposition result remains in
/// `nabled`'s ndarray-native result struct.
///
/// # Errors
/// Returns an error when the matrix contains nulls, is empty, or is not SPD.
pub fn decompose_f32(
    matrix: &FixedSizeListArray,
) -> Result<crate::linalg::cholesky::NdarrayCholeskyResult<f32>, ArrowInteropError> {
    let matrix_view = fixed_size_list_view::<Float32Type>(matrix)?;
    Ok(crate::linalg::cholesky::decompose_view(&matrix_view)?)
}

/// Compute `f64` Cholesky decomposition directly from an Arrow dense matrix.
///
/// The inbound Arrow -> ndarray bridge is zero-copy. The decomposition result remains in
/// `nabled`'s ndarray-native result struct.
///
/// # Errors
/// Returns an error when the matrix contains nulls, is empty, or is not SPD.
pub fn decompose_f64(
    matrix: &FixedSizeListArray,
) -> Result<crate::linalg::cholesky::NdarrayCholeskyResult<f64>, ArrowInteropError> {
    let matrix_view = fixed_size_list_view::<Float64Type>(matrix)?;
    Ok(crate::linalg::cholesky::decompose_view(&matrix_view)?)
}

/// Solve `Ax=b` directly from `f32` Arrow dense inputs using Cholesky decomposition.
///
/// # Errors
/// Returns an error when inputs contain nulls, are empty, dimensions mismatch, or the matrix is
/// not SPD.
pub fn solve_f32(
    matrix: &FixedSizeListArray,
    rhs: &PrimitiveArray<Float32Type>,
) -> Result<PrimitiveArray<Float32Type>, ArrowInteropError> {
    let matrix_view = fixed_size_list_view::<Float32Type>(matrix)?;
    let rhs_view = primitive_array_view(rhs)?;
    let solution = crate::linalg::cholesky::solve_view(&matrix_view, &rhs_view)?;
    Ok(primitive_array_from_owned::<Float32Type>(solution))
}

/// Solve `Ax=b` directly from `f64` Arrow dense inputs using Cholesky decomposition.
///
/// # Errors
/// Returns an error when inputs contain nulls, are empty, dimensions mismatch, or the matrix is
/// not SPD.
pub fn solve_f64(
    matrix: &FixedSizeListArray,
    rhs: &PrimitiveArray<Float64Type>,
) -> Result<PrimitiveArray<Float64Type>, ArrowInteropError> {
    let matrix_view = fixed_size_list_view::<Float64Type>(matrix)?;
    let rhs_view = primitive_array_view(rhs)?;
    let solution = crate::linalg::cholesky::solve_view(&matrix_view, &rhs_view)?;
    Ok(primitive_array_from_owned::<Float64Type>(solution))
}

/// Compute the `f32` inverse directly from an Arrow SPD dense matrix.
///
/// # Errors
/// Returns an error when the matrix contains nulls, is empty, or is not SPD.
pub fn inverse_f32(matrix: &FixedSizeListArray) -> Result<FixedSizeListArray, ArrowInteropError> {
    let matrix_view = fixed_size_list_view::<Float32Type>(matrix)?;
    let output = crate::linalg::cholesky::inverse_view(&matrix_view)?;
    fixed_size_list_from_owned::<Float32Type>(output)
}

/// Compute the `f64` inverse directly from an Arrow SPD dense matrix.
///
/// # Errors
/// Returns an error when the matrix contains nulls, is empty, or is not SPD.
pub fn inverse_f64(matrix: &FixedSizeListArray) -> Result<FixedSizeListArray, ArrowInteropError> {
    let matrix_view = fixed_size_list_view::<Float64Type>(matrix)?;
    let output = crate::linalg::cholesky::inverse_view(&matrix_view)?;
    fixed_size_list_from_owned::<Float64Type>(output)
}

/// Compute complex Cholesky decomposition directly from Arrow complex dense input.
///
/// # Errors
/// Returns an error when the matrix contains nulls, is empty, or is not Hermitian positive
/// definite.
pub fn decompose_complex(
    matrix: &FixedSizeListArray,
) -> Result<crate::linalg::cholesky::NdarrayComplexCholeskyResult, ArrowInteropError> {
    let matrix_view = complex64_matrix_view(matrix)?;
    Ok(crate::linalg::cholesky::decompose_complex_view(&matrix_view)?)
}

/// Solve a complex Hermitian positive-definite system directly from Arrow complex inputs.
///
/// # Errors
/// Returns an error when inputs contain nulls, are empty, or dimensions mismatch.
pub fn solve_complex(
    matrix: &FixedSizeListArray,
    rhs_field: &Field,
    rhs: &FixedSizeListArray,
) -> Result<(Field, FixedSizeListArray), ArrowInteropError> {
    let matrix_view = complex64_matrix_view(matrix)?;
    let rhs_view = complex64_vector_view(rhs_field, rhs)?;
    let output = crate::linalg::cholesky::solve_complex_view(&matrix_view, &rhs_view)?;
    complex64_vector_from_owned("cholesky_solve_complex", output)
}

/// Compute the complex inverse of an Arrow complex Hermitian positive-definite matrix.
///
/// # Errors
/// Returns an error when the matrix contains nulls, is empty, or is not Hermitian positive
/// definite.
pub fn inverse_complex(
    matrix: &FixedSizeListArray,
) -> Result<FixedSizeListArray, ArrowInteropError> {
    let matrix_view = complex64_matrix_view(matrix)?;
    let output = crate::linalg::cholesky::inverse_complex_view(&matrix_view)?;
    complex64_matrix_from_owned(output)
}