aprender-core 0.31.2

Next-generation machine learning library in pure Rust
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

pub(crate) use super::*;

#[test]
fn test_diarization_config_default() {
    let config = DiarizationConfig::default();
    assert!(config.validate().is_ok());
    assert_eq!(config.min_speakers, 1);
    assert!(config.max_speakers.is_none());
}

#[test]
fn test_diarization_config_with_speakers() {
    let config = DiarizationConfig::default().with_speakers(2, Some(4));
    assert_eq!(config.min_speakers, 2);
    assert_eq!(config.max_speakers, Some(4));
}

#[test]
fn test_diarization_config_validation() {
    let mut config = DiarizationConfig::default();
    config.min_speakers = 0;
    assert!(config.validate().is_err());

    config.min_speakers = 2;
    config.max_speakers = Some(1);
    assert!(config.validate().is_err());

    config.max_speakers = Some(3);
    config.clustering_threshold = 1.5;
    assert!(config.validate().is_err());
}

#[test]
fn test_speaker_new() {
    let speaker = Speaker::new(0, vec![1.0, 0.0, 0.0]);
    assert_eq!(speaker.id, 0);
    assert_eq!(speaker.label, "SPEAKER_00");
    assert_eq!(speaker.embedding.len(), 3);
}

#[test]
fn test_speaker_similarity() {
    let s1 = Speaker::new(0, vec![1.0, 0.0, 0.0]);
    let s2 = Speaker::new(1, vec![1.0, 0.0, 0.0]);
    let s3 = Speaker::new(2, vec![0.0, 1.0, 0.0]);

    assert!((s1.similarity(&s2) - 1.0).abs() < 0.001); // Identical
    assert!(s1.similarity(&s3).abs() < 0.001); // Orthogonal
}

#[test]
fn test_speaker_segment_new() {
    let seg = SpeakerSegment::new(0, 1000, 2000);
    assert_eq!(seg.speaker_id, 0);
    assert_eq!(seg.speaker_label, "SPEAKER_00");
    assert_eq!(seg.duration_ms(), 1000);
}

#[test]
fn test_diarization_result_empty() {
    let result = DiarizationResult::new();
    assert!(result.is_empty());
    assert_eq!(result.speaker_count(), 0);
}

#[test]
fn test_cosine_similarity() {
    let a = vec![1.0, 0.0, 0.0];
    let b = vec![1.0, 0.0, 0.0];
    let c = vec![0.0, 1.0, 0.0];
    let d = vec![-1.0, 0.0, 0.0];

    assert!((cosine_similarity(&a, &b) - 1.0).abs() < 0.001);
    assert!(cosine_similarity(&a, &c).abs() < 0.001);
    assert!((cosine_similarity(&a, &d) + 1.0).abs() < 0.001);
}

#[test]
fn test_cosine_similarity_edge_cases() {
    assert_eq!(cosine_similarity(&[], &[]), 0.0);
    assert_eq!(cosine_similarity(&[1.0], &[1.0, 2.0]), 0.0); // Different lengths
    assert_eq!(cosine_similarity(&[0.0, 0.0], &[1.0, 0.0]), 0.0); // Zero vector
}

#[test]
fn test_average_embeddings() {
    let e1 = vec![1.0, 0.0];
    let e2 = vec![0.0, 1.0];
    let embeddings: Vec<&Vec<f32>> = vec![&e1, &e2];

    let avg = average_embeddings(&embeddings, 2);
    assert!((avg[0] - 0.5).abs() < 0.001);
    assert!((avg[1] - 0.5).abs() < 0.001);
}

#[test]
fn test_diarize_empty() {
    let config = DiarizationConfig::default();
    let result = diarize(&[], &[], &config);
    assert!(result.is_ok());
    assert!(result.unwrap().is_empty());
}

#[test]
fn test_diarize_single_speaker() {
    let config = DiarizationConfig {
        embedding_dim: 3,
        clustering_threshold: 0.9,
        ..Default::default()
    };

    // Two very similar embeddings = same speaker
    let embeddings = vec![vec![1.0, 0.0, 0.0], vec![0.99, 0.1, 0.0]];
    let times = vec![(0, 1000), (1000, 2000)];

    let result = diarize(&embeddings, &times, &config).unwrap();
    assert_eq!(result.speaker_count(), 1);
    assert_eq!(result.segments.len(), 2);
}

#[test]
fn test_diarize_two_speakers() {
    let config = DiarizationConfig {
        embedding_dim: 3,
        clustering_threshold: 0.5,
        ..Default::default()
    };

    // Two very different embeddings = different speakers
    let embeddings = vec![vec![1.0, 0.0, 0.0], vec![0.0, 1.0, 0.0]];
    let times = vec![(0, 1000), (1000, 2000)];

    let result = diarize(&embeddings, &times, &config).unwrap();
    assert_eq!(result.speaker_count(), 2);
}

#[test]
fn test_diarize_speaking_time() {
    let config = DiarizationConfig {
        embedding_dim: 3,
        clustering_threshold: 0.9,
        ..Default::default()
    };

    let embeddings = vec![vec![1.0, 0.0, 0.0], vec![1.0, 0.0, 0.0]];
    let times = vec![(0, 1000), (1000, 3000)];

    let result = diarize(&embeddings, &times, &config).unwrap();
    assert_eq!(result.speaking_time_ms(0), 3000);
}

#[test]
fn test_diarize_mismatched_lengths() {
    let config = DiarizationConfig::default();
    let embeddings = vec![vec![0.0; 192]];
    let times = vec![(0, 1000), (1000, 2000)]; // One extra

    let result = diarize(&embeddings, &times, &config);
    assert!(result.is_err());
}

#[test]
fn test_diarize_wrong_embedding_dim() {
    let config = DiarizationConfig {
        embedding_dim: 192,
        ..Default::default()
    };

    let embeddings = vec![vec![0.0; 100]]; // Wrong dim
    let times = vec![(0, 1000)];

    let result = diarize(&embeddings, &times, &config);
    assert!(result.is_err());
}