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

//! implementation of two neuron type: brain neuron and block neuron

use std::fmt;

use ndarray::prelude::*;
use rand::prelude::*;
use serde::{Serialize, Deserialize};

use crate::consts::*;

use super::{
    nn::BaseNN,
    signal::{BrainSignal, InwardNNInputSignal, OutwardNNInputSignal},
};

const CL: usize = INWARD_NN_CHILDREN_INPUT_LEN;
const DL: usize = OUTWARD_NN_PARENT_INPUT_LEN;

#[derive(Debug, Clone, Serialize, Deserialize)]
pub enum GenericNN {
    BLOCKNN(BlockNN),
    BRAINNN(BrainNN),
}

#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct InwardNN {
    pub nn: BaseNN,
}

impl Default for InwardNN {
    fn default() -> Self {
        Self {
            nn: BaseNN::new_rand(
                Vec::from_iter(INWARD_NN_SHAPE.into_iter().clone()),
                ACTIVATION_FUNCTION,
            ),
        }
    }
}

#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct OutwardNN {
    pub nn: BaseNN,
}

impl Default for OutwardNN {
    fn default() -> Self {
        Self {
            nn: BaseNN::new_rand(
                Vec::from_iter(OUTWARD_NN_SHAPE.into_iter().clone()),
                ACTIVATION_FUNCTION,
            ),
        }
    }
}

/// neuron for blocks.
///
/// Each block should have two independent neurons:
/// InwardNN and OutwardNN
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct BlockNN {
    pub inward_nn: InwardNN,
    pub outward_nn: OutwardNN,
    pub outward_signal: OutwardNNInputSignal,
}

impl Default for BlockNN {
    fn default() -> Self {
        Self {
            inward_nn: InwardNN::default(),
            outward_nn: OutwardNN::default(),
            outward_signal: OutwardNNInputSignal::default(),
        }
    }
}

impl BlockNN {
    /// forward function for inward nn
    ///
    /// also update the `inherited` element in outward nn
    fn inward_forward(&mut self, signal: &InwardNNInputSignal) -> Array1<f32> {
        let array_signal = signal.to_array();
        // save duplicate signals for ourward usage
        self.outward_signal.inherit(&array_signal);
        self.inward_nn.nn.forward(array_signal.clone())
    }

    /// output inward signal that passing to next layer
    /// Takes input layer's singal
    pub fn get_inward_output(&mut self, signal: &InwardNNInputSignal) -> Array1<f32> {
        self.inward_forward(signal)
    }

    pub fn get_rand_inward_output(&self) -> Array1<f32> {
        let mut rng = thread_rng();
        Array1::from_shape_fn((4,), |_| rng.gen::<f32>())
    }

    pub fn get_outward_output(&mut self, parent_signal: &Array1<f32>) -> Array1<f32> {
        assert_eq!(parent_signal.len(), DL);
        self.outward_signal.parent_input = parent_signal.clone();
        self.outward_nn.nn.forward(self.outward_signal.to_array())
    }
}

/// NN for centeral brain
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct BrainNN {
    pub nn: BaseNN,
}

impl Default for BrainNN {
    fn default() -> Self {
        Self {
            nn: BaseNN::new_rand(
                Vec::from_iter(BRAIN_NN_SHAPE.into_iter().clone()),
                ACTIVATION_FUNCTION,
            ),
        }
    }
}

impl BrainNN {
    pub fn forward(&self, signal: &BrainSignal) -> Array1<f32> {
        self.nn.forward(signal.to_array())
    }

    pub fn get_rand_brain_output(&self) -> Array1<f32> {
        let mut rng = thread_rng();
        Array1::from_shape_fn((4,), |_| rng.gen::<f32>())
    }
}

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