ebi 0.3.12

A stochastic process mining utility and library
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

use crate::{
    ebi_traits::ebi_trait_finite_stochastic_language::EbiTraitFiniteStochasticLanguage,
    math::distances::TriangularDistanceMatrix,
};
use ebi_objects::{
    FiniteLanguage,
    anyhow::{Result, anyhow},
    ebi_arithmetic::{Fraction, One, Zero, f},
};

pub fn medoid<T>(log: &T, number_of_traces: &usize) -> Result<FiniteLanguage>
where
    T: EbiTraitFiniteStochasticLanguage + ?Sized,
{
    let activity_key = log.activity_key().clone();
    let mut result = FiniteLanguage::new_hashmap();

    log::info!("Computing {} medoid traces", number_of_traces);

    let distances = TriangularDistanceMatrix::new(log);

    if number_of_traces.is_one() {
        if let Some(trace_index) = medoid_single(log, &distances) {
            result.insert(
                log.iter_traces()
                    .nth(trace_index)
                    .unwrap()
                    .to_owned(),
            );
            return Ok((activity_key, result).into());
        } else {
            return Err(anyhow!(
                "1 trace was requested, but the stochastic language contains none."
            ));
        }
    }

    if log.number_of_traces() < *number_of_traces {
        return Err(anyhow!(
            "{} traces were requested, but the stochastic language contains only {} traces.",
            number_of_traces,
            log.number_of_traces()
        ));
    }

    let mut sum_distance = sum_distances(log, &distances);

    let mut list = Vec::new();
    while list.len() < *number_of_traces {
        //find the position of the minimum value
        let mut min_pos = 0;
        for i in 1..sum_distance.len() {
            if sum_distance[i] < sum_distance[min_pos] {
                min_pos = i;
            }
        }

        //report the minimum value
        list.push(min_pos);
        sum_distance[min_pos] = f!(2);
    }
    list.sort();

    //put in the output format
    let mut list_i = 0;
    for (i1, trace1) in log.iter_traces().enumerate() {
        if list_i < list.len() && i1 == list[list_i] {
            result.insert(trace1.to_vec());
            list_i += 1;
        }
    }

    Ok((activity_key, result).into())
}

/**
 * Returns the index of the weighted medoid, if there is one.
 */
pub fn medoid_single<T>(log: &T, distances: &TriangularDistanceMatrix) -> Option<usize>
where
    T: EbiTraitFiniteStochasticLanguage + ?Sized,
{
    let sum_distance = sum_distances(log, distances);

    if distances.len() == 0 {
        return None;
    }

    //report the minimum value
    let mut min_pos = 0;
    let mut min_value = &Fraction::one();
    for (pos, value) in sum_distance.iter().enumerate() {
        if value < &min_value {
            min_pos = pos;
            min_value = value;
        }
    }

    return Some(min_pos);
}

pub fn sum_distances<T>(log: &T, distances: &TriangularDistanceMatrix) -> Vec<Fraction>
where
    T: EbiTraitFiniteStochasticLanguage + ?Sized,
{
    let mut sum_distance = vec![Fraction::zero(); log.number_of_traces()];

    for (i, j, _, distance) in distances {
        let mut distance_i = distance.as_ref().clone();
        distance_i *= log.iter_probabilities().nth(j).unwrap();
        sum_distance[i] += &distance_i;

        let mut distance_j = distance.as_ref().clone();
        distance_j *= log.iter_probabilities().nth(i).unwrap();
        sum_distance[j] += distance_j;
    }

    sum_distance
}

#[cfg(test)]
mod tests {
    use std::fs;

    use ebi_objects::FiniteStochasticLanguage;

    use crate::techniques::medoid;

    #[test]
    fn medoid() {
        let fin = fs::read_to_string("testfiles/aa-ab-ba.slang").unwrap();
        let slang = fin.parse::<FiniteStochasticLanguage>().unwrap();
        let fout = fs::read_to_string("testfiles/ba.lang").unwrap();
        let medoid = medoid::medoid(&slang, &1).unwrap();
        assert_eq!(fout, medoid.to_string())
    }
}