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

use crate::{Categorical, CategoricalParams, DistributionError, SampleableDistribution};
use crate::{DependentJoint, Distribution, IndependentJoint, RandomVariable};
use opensrdk_linear_algebra::*;
use rand::prelude::*;
use rayon::prelude::*;
use std::collections::HashMap;
use std::hash::Hash;
use std::{
    fmt::Debug,
    ops::{BitAnd, Mul},
};

#[derive(Clone, Debug)]
pub struct DiscreteSamplesDistribution<T>
where
    T: RandomVariable + Eq + Hash,
{
    n: usize,
    n_map: HashMap<T, usize>,
}

#[derive(thiserror::Error, Debug)]
pub enum DiscreteSamplesError {
    #[error("Samples are empty")]
    SamplesAreEmpty,
    #[error("TransformVec info mismatch")]
    TransformVecInfoMismatch,
}

impl<T> DiscreteSamplesDistribution<T>
where
    T: RandomVariable + Eq + Hash,
{
    pub fn new(samples: Vec<T>) -> Self {
        let n = samples.len();
        let mut n_map = HashMap::new();
        for sample in samples {
            *n_map.entry(sample).or_insert(0) += 1;
        }

        Self { n, n_map }
    }

    pub fn push(&mut self, v: T) {
        self.n += 1;
        *self.n_map.entry(v).or_insert(0) += 1;
    }

    pub fn mode<'a>(&'a self) -> Result<&'a T, DistributionError> {
        if self.n == 0 {
            return Err(DistributionError::InvalidParameters(
                DiscreteSamplesError::SamplesAreEmpty.into(),
            ));
        }
        Ok(self
            .n_map
            .par_iter()
            .max_by_key(|&(_, &count)| count)
            .map(|(val, _)| val)
            .unwrap_or(
                self.n_map
                    .iter()
                    .take(1)
                    .map(|(k, _)| k)
                    .collect::<Vec<_>>()[0],
            ))
    }
}

impl<T> DiscreteSamplesDistribution<T>
where
    T: RandomVariable + Eq + Hash,
{
    pub fn sum(&self) -> Result<T, DistributionError> {
        let n = self.n;
        if n == 0 {
            return Err(DistributionError::InvalidParameters(
                DiscreteSamplesError::SamplesAreEmpty.into(),
            ));
        }
        let vec = self.n_map.iter().collect::<Vec<_>>();
        let (sum, info) = vec[0].0.clone().transform_vec();
        let mut sum = sum.col_mat();
        for i in 1..n {
            let (v, info_i) = vec[i].0.clone().transform_vec();
            if info != info_i {
                return Err(DistributionError::Others(
                    DiscreteSamplesError::TransformVecInfoMismatch.into(),
                ));
            }
            sum = sum + (*vec[i].1 as f64) * v.col_mat();
        }

        T::restore(sum.elems(), &info)
    }

    pub fn mean(&mut self) -> Result<T, DistributionError> {
        let (sum, info) = self.sum().unwrap().transform_vec();
        let elems = sum
            .iter()
            .map(|elem| elem / self.n_map.len() as f64)
            .collect::<Vec<f64>>();
        T::restore(&elems, &info)
    }
}

impl<T> Distribution for DiscreteSamplesDistribution<T>
where
    T: RandomVariable + Eq + Hash,
{
    type Value = T;
    type Condition = ();

    fn p_kernel(&self, x: &Self::Value, _: &Self::Condition) -> Result<f64, DistributionError> {
        Ok(*self.n_map.get(x).unwrap_or(&0) as f64 / self.n as f64)
    }
}

impl<T, Rhs, TRhs> Mul<Rhs> for DiscreteSamplesDistribution<T>
where
    T: RandomVariable + Eq + Hash,
    Rhs: Distribution<Value = TRhs, Condition = ()>,
    TRhs: RandomVariable,
{
    type Output = IndependentJoint<Self, Rhs, T, TRhs, ()>;

    fn mul(self, rhs: Rhs) -> Self::Output {
        IndependentJoint::new(self, rhs)
    }
}

impl<T, Rhs, URhs> BitAnd<Rhs> for DiscreteSamplesDistribution<T>
where
    T: RandomVariable + Eq + Hash,
    Rhs: Distribution<Value = (), Condition = URhs>,
    URhs: RandomVariable,
{
    type Output = DependentJoint<Self, Rhs, T, (), URhs>;

    fn bitand(self, rhs: Rhs) -> Self::Output {
        DependentJoint::new(self, rhs)
    }
}

impl<T> SampleableDistribution for DiscreteSamplesDistribution<T>
where
    T: RandomVariable + Eq + Hash,
{
    fn sample(
        &self,
        _theta: &Self::Condition,
        rng: &mut dyn RngCore,
    ) -> Result<Self::Value, DistributionError> {
        let keys = self.n_map.keys().collect::<Vec<_>>();
        let pi = keys
            .iter()
            .map(|&k| *self.n_map.get(k).unwrap())
            .map(|ni| ni as f64 / self.n as f64)
            .collect();
        let params = CategoricalParams::new(pi)?;
        let sampled = Categorical.sample(&params, rng)?;

        Ok(keys[sampled].clone())
    }
}