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
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177

use std::cmp::Ordering;

pub mod error;
pub mod models;

#[cfg(target_arch = "wasm32")]
pub mod wasm;

pub use error::ClevelError;
pub use models::{Model, NormalizedModel};

trait FloatIterExt {
    fn float_min(&mut self) -> f64;
    fn float_max(&mut self) -> f64;
}

impl<T> FloatIterExt for T
where
    T: Iterator<Item = f64>,
{
    fn float_max(&mut self) -> f64 {
        self.fold(std::f64::NAN, f64::max)
    }

    fn float_min(&mut self) -> f64 {
        self.fold(std::f64::NAN, f64::min)
    }
}

lazy_static::lazy_static! {
    pub static ref VERSION: &'static str = {
        let mut version = env!("CARGO_PKG_VERSION").to_string();
        if let Some(v) = option_env!("GIT_HASH") {
            version.push_str("-");
            version.push_str(v);
        }
        if cfg!(debug_assertions) {
            version.push_str("-debug");
        }
        Box::leak(Box::new(version))
    };
}

fn phase_data(crisp: u32, a: u32, b: u32) -> Option<f64> {
    match Some(f64::default()) {
        Some(_) if crisp <= a => Some(0.0),
        Some(_) if a < crisp && crisp <= (a + b) / 2 => {
            let first = (crisp - a) as f64;
            let second = (b - a) as f64;
            return Some(2.0 * f64::powf(first / second, 2.0));
        }
        Some(_) if (a + b) / 2 < crisp && crisp < b => {
            let first = (b - crisp) as f64;
            let second = (b - a) as f64;
            return Some(1.0 - 2.0 * f64::powf(first / second, 2.0));
        }
        Some(_) if crisp >= b => Some(1.0),
        _ => None,
    }
}

fn phase_data_vec(models: &Vec<Model>) -> Vec<NormalizedModel> {
    let mut normalized_models: Vec<NormalizedModel> = vec![];
    for item in models {
        let phased_crisps = item
            .crisps
            .iter()
            .map(|num| phase_data(*num, item.a, item.b).unwrap())
            .collect::<Vec<f64>>();
        let phased_t = phase_data(item.t, item.a, item.b).unwrap();
        normalized_models.push(NormalizedModel {
            phased_crisps,
            phased_t,
        })
    }
    normalized_models
}

fn weighting_normalization(models: &Vec<Model>) -> Vec<f64> {
    let pvec = models.iter().map(|model| model.p).collect::<Vec<u32>>();
    let sum_of_pvec = pvec.iter().sum::<u32>();
    pvec.iter()
        .map(|p| *p as f64 / sum_of_pvec as f64)
        .collect::<Vec<f64>>()
}

fn calc_desired_value(crisp: f64, min_crisp: f64, max_crisp: f64, t: f64) -> f64 {
    let first = (t - crisp).abs();
    let second: Vec<f64> = vec![t - min_crisp, max_crisp - t];
    let max_second = second.iter().cloned().float_max();
    1.0 - (first / max_second)
}

fn desired_values(models: &Vec<Model>) -> Vec<Vec<f64>> {
    let mut vec_of_vec: Vec<Vec<f64>> = vec![];
    let normalized_models = phase_data_vec(models);
    for nm in normalized_models {
        let vv: Vec<f64> = nm
            .phased_crisps
            .iter()
            .map(|num: &f64| {
                let max_val = nm.phased_crisps.iter().cloned().float_max();
                let min_val = nm.phased_crisps.iter().cloned().float_min();
                calc_desired_value(*num, min_val, max_val, nm.phased_t)
            })
            .collect();
        vec_of_vec.push(vv)
    }
    vec_of_vec
}

fn simplify_desired_values(dv: Vec<Vec<f64>>) -> Vec<f64> {
    let mut new_dv: Vec<f64> = vec![];

    for v in dv {
        for i in v {
            new_dv.push(i)
        }
    }

    new_dv
}

fn initialize_matrix(models: &Vec<Model>) -> simple_matrix::Matrix<f64> {
    let dv = desired_values(&models);
    let simple_dv = simplify_desired_values(dv);

    let rows = models.len();
    let cols = models[0].crisps.len();

    let matrix: simple_matrix::Matrix<f64> =
        simple_matrix::Matrix::from_iter(rows, cols, simple_dv.iter().cloned());

    matrix
}

pub fn aggregated_score(models: &Vec<Model>) -> Vec<f64> {
    let weights = weighting_normalization(&models);
    let matrix = initialize_matrix(&models);

    let mut temp: Vec<f64> = vec![];

    for col in 0..matrix.cols() {
        let get_col = matrix.get_col(col).unwrap().collect::<Vec<&f64>>();
        let mut res: Vec<f64> = vec![];

        for row in 0..matrix.rows() {
            res.push(*get_col[row] * weights[row])
        }

        temp.push(res.iter().sum())
    }

    temp
}

pub fn find_head_from_models(models: Vec<Model>) -> (String, f64) {
    let aggregation_estimates = aggregated_score(&models);
    let maxf = aggregation_estimates.iter().cloned().float_max();
    let mut expert: String = String::new();

    for (index, item) in aggregation_estimates.iter().enumerate() {
        if *item == maxf {
            expert = format!("Expert {}", index + 1);
        }
    }

    (expert, maxf)
}

pub fn ranking_of_experts(models: &Vec<Model>) -> Vec<f64> {
    let aggregation_estimates = aggregated_score(&models);
    let mut sorted_aggregation_estimates = aggregation_estimates.clone();
    sorted_aggregation_estimates.sort_by(|a, b| a.partial_cmp(b).unwrap_or(Ordering::Greater));

    sorted_aggregation_estimates.iter().rev().cloned().collect()
}