concision-linear 0.1.14

Concision is a complete data-science toolkit written in Rust
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
130
131
132
133
134
135
136
137
138
139
140

/*
   Appellation: features <mod>
   Contrib: FL03 <jo3mccain@icloud.com>
*/
use nd::prelude::{Dimension, Ix2, ShapeBuilder};
use nd::{ErrorKind, IntoDimension, RemoveAxis, ShapeError};

pub(crate) fn features<D>(dim: D) -> Result<Features, ShapeError>
where
    D: Dimension,
{
    if dim.ndim() == 1 {
        Ok(Features::new(1, dim[0]))
    } else if dim.ndim() >= 2 {
        Ok(Features::new(dim[0], dim[1]))
    } else {
        Err(ShapeError::from_kind(ErrorKind::IncompatibleShape))
    }
}

#[derive(Clone, Copy, Debug, Default, Eq, Hash, Ord, PartialEq, PartialOrd)]
#[cfg_attr(feature = "serde", derive(serde::Deserialize, serde::Serialize))]
pub struct Features {
    pub dmodel: usize,  // inputs
    pub outputs: usize, // outputs
}

impl Features {
    pub fn new(outputs: usize, dmodel: usize) -> Self {
        Self { dmodel, outputs }
    }

    pub fn from_dim<D>(dim: D) -> Self
    where
        D: RemoveAxis,
    {
        features(dim).unwrap()
    }

    pub fn from_shape<D, Sh>(shape: Sh) -> Self
    where
        D: RemoveAxis,
        Sh: ShapeBuilder<Dim = D>,
    {
        let shape = shape.into_shape();
        let dim = shape.raw_dim().clone();
        features(dim).unwrap()
    }

    pub fn check_dim<D>(&self, dim: D) -> bool
    where
        D: Dimension,
    {
        if dim.ndim() == 1 {
            self.dmodel == dim[0]
        } else if dim.ndim() >= 2 {
            self.outputs == dim[0] && self.dmodel == dim[1]
        } else {
            false
        }
    }

    pub fn into_pattern(self) -> (usize, usize) {
        (self.outputs, self.dmodel)
    }

    pub fn neuron(d_model: usize) -> Self {
        Self::new(1, d_model)
    }

    pub fn dmodel(&self) -> usize {
        self.dmodel
    }

    pub fn features(&self) -> usize {
        self.outputs
    }

    pub fn uniform_scale<T: num::Float>(&self) -> T {
        T::from(self.dmodel()).unwrap().recip().sqrt()
    }
}

impl core::fmt::Display for Features {
    fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result {
        write!(f, "({}, {})", self.features(), self.dmodel(),)
    }
}

impl IntoDimension for Features {
    type Dim = Ix2;

    fn into_dimension(self) -> Self::Dim {
        ndarray::Ix2(self.features(), self.dmodel())
    }
}

impl PartialEq<(usize, usize)> for Features {
    fn eq(&self, other: &(usize, usize)) -> bool {
        self.features() == other.0 && self.dmodel() == other.1
    }
}

macro_rules! impl_from {
    ($($s:ty: $t:ty { $into:expr }),* $(,)?) => {
        $(impl_from!(@impl $s: $t { $into });)*
    };
    (@impl $s:ty: $t:ty { $into:expr }) => {
        impl From<$t> for $s {
            fn from(features: $t) -> Self {
                $into(features)
            }
        }
    };
}

impl_from!(
    Features: usize { |f: usize| Features::new(1, f) },
    Features: [usize; 2] {| shape: [usize; 2] | Features::new(shape[0], shape[1])},
    Features: (usize, usize) {| shape: (usize, usize) | Features::new(shape.0, shape.1)},
    Features: nd::Ix1 {| shape: nd::Ix1 | Features::from(&shape)},
    Features: nd::Ix2 {| shape: nd::Ix2 | Features::from(&shape)},
    Features: nd::IxDyn {| shape: nd::IxDyn | Features::from(&shape)},
);

impl_from!(
    nd::Ix2: Features { |f: Features| f.into_dimension() },
    nd::IxDyn: Features { |f: Features| f.into_dimension().into_dyn() },
    [usize; 2]: Features { |f: Features| [f.outputs, f.dmodel] },
    (usize, usize): Features { |f: Features| (f.outputs, f.dmodel) },
);

impl<'a, D> From<&'a D> for Features
where
    D: RemoveAxis,
{
    fn from(dim: &'a D) -> Features {
        features(dim.clone()).unwrap()
    }
}