radiate_extensions/architects/node_collections/
node_factory.rs

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
use crate::architects::node_collections::node::Node;
use crate::{
    architects::schema::node_types::NodeType,
    operations::op::{self, Ops},
};
use radiate::random_provider;
use std::collections::HashMap;

#[derive(Clone, Default, PartialEq, Debug)]
pub struct NodeFactory<T>
where
    T: Clone + PartialEq + Default,
{
    pub node_values: HashMap<NodeType, Vec<Ops<T>>>,
}

impl<T> NodeFactory<T>
where
    T: Clone + PartialEq + Default,
{
    pub fn new() -> Self {
        NodeFactory {
            node_values: HashMap::new(),
        }
    }

    pub fn leafs(mut self, values: Vec<Ops<T>>) -> NodeFactory<T> {
        self.add_node_values(NodeType::Leaf, values);
        self
    }

    pub fn inputs(mut self, values: Vec<Ops<T>>) -> NodeFactory<T> {
        self.add_node_values(NodeType::Input, values);
        self
    }

    pub fn outputs(mut self, values: Vec<Ops<T>>) -> NodeFactory<T> {
        self.add_node_values(NodeType::Output, values);
        self
    }

    pub fn gates(mut self, values: Vec<Ops<T>>) -> NodeFactory<T> {
        self.add_node_values(NodeType::Gate, values);
        self
    }

    pub fn aggregates(mut self, values: Vec<Ops<T>>) -> NodeFactory<T> {
        self.add_node_values(NodeType::Aggregate, values);
        self
    }

    pub fn weights(mut self, values: Vec<Ops<T>>) -> NodeFactory<T> {
        self.add_node_values(NodeType::Weight, values);
        self
    }

    pub fn set_values(mut self, node_type: NodeType, values: Vec<Ops<T>>) -> NodeFactory<T> {
        self.add_node_values(node_type, values);
        self
    }

    pub fn add_node_values(&mut self, node_type: NodeType, values: Vec<Ops<T>>) {
        self.node_values.insert(node_type, values);
    }

    pub fn new_node(&self, index: usize, node_type: NodeType) -> Node<T> {
        if let Some(values) = self.node_values.get(&node_type) {
            return match node_type {
                NodeType::Input => {
                    let value = values[index % values.len()].clone();
                    Node::new(index, node_type, value)
                }
                _ => {
                    let value = random_provider::choose(values);
                    Node::new(index, node_type, value.new_instance())
                }
            };
        }

        Node::new(index, node_type, Ops::default())
    }

    pub fn regression(input_size: usize) -> NodeFactory<f32> {
        let inputs = (0..input_size).map(op::var).collect::<Vec<Ops<f32>>>();
        NodeFactory::new()
            .inputs(inputs.clone())
            .leafs(inputs.clone())
            .gates(vec![
                op::add(),
                op::sub(),
                op::mul(),
                op::div(),
                op::pow(),
                op::sqrt(),
                op::exp(),
                op::abs(),
                op::log(),
                op::sin(),
                op::cos(),
                op::tan(),
                op::sum(),
                op::prod(),
                op::max(),
                op::min(),
                op::ceil(),
                op::floor(),
                op::gt(),
                op::lt(),
            ])
            .aggregates(vec![
                op::sigmoid(),
                op::tanh(),
                op::relu(),
                op::linear(),
                op::sum(),
                op::prod(),
                op::max(),
                op::min(),
                op::mish(),
                op::leaky_relu(),
                op::softplus(),
                op::sum(),
                op::prod(),
            ])
            .weights(vec![op::weight()])
            .outputs(vec![op::linear()])
    }
}