laurus 0.9.0

Unified search library for lexical, vector, and semantic retrieval
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
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

use laurus::storage::file::{FileStorage, FileStorageConfig};
use laurus::vector::{
    DistanceMetric, HnswIndexConfig, HnswIndexReader, HnswIndexWriter, HnswSearcher, Vector,
    VectorIndexQuery, VectorIndexSearcher, VectorIndexWriter, VectorIndexWriterConfig,
};
use std::sync::Arc;
use tempfile::tempdir;

#[test]
fn test_hnsw_append_persistence() -> Result<(), Box<dyn std::error::Error>> {
    let dir = tempdir()?;
    let path = dir.path();
    let index_name = "test_index";

    let index_config = HnswIndexConfig {
        dimension: 3,
        m: 16,
        ef_construction: 100,
        normalize_vectors: false,
        distance_metric: DistanceMetric::Euclidean,
        ..Default::default()
    };

    let writer_config = VectorIndexWriterConfig {
        parallel_build: true,
        ..Default::default()
    };

    // 1. Initial Build
    {
        let storage_config = FileStorageConfig::new(path);
        let storage = Arc::new(FileStorage::new(path, storage_config)?);

        // Use with_storage to ensure we can write
        let mut writer = HnswIndexWriter::with_storage(
            index_config.clone(),
            writer_config.clone(),
            index_name,
            storage,
        )?;

        let vectors = vec![
            (1, "doc1".to_string(), Vector::new(vec![1.0f32, 0.0, 0.0])),
            (2, "doc2".to_string(), Vector::new(vec![0.0f32, 1.0, 0.0])),
        ];

        writer.add_vectors(vectors)?;
        writer.finalize()?;
        writer.write()?;
    }

    // 2. Load and Append
    {
        let storage_config = FileStorageConfig::new(path);
        let storage = Arc::new(FileStorage::new(path, storage_config)?);

        let mut writer = HnswIndexWriter::load(
            index_config.clone(),
            writer_config.clone(),
            storage,
            index_name,
        )?;

        let new_vectors = vec![
            (3, "doc3".to_string(), Vector::new(vec![0.0f32, 0.0, 1.0])),
            (4, "doc4".to_string(), Vector::new(vec![1.0f32, 1.0, 0.0])),
        ];

        writer.add_vectors(new_vectors)?;
        writer.finalize()?;
        writer.write()?;
    }

    Ok(())
}

#[test]
fn test_hnsw_append_search_verification() -> Result<(), Box<dyn std::error::Error>> {
    let dir = tempdir()?;
    let path = dir.path();
    let index_name = "search_test";

    let index_config = HnswIndexConfig {
        dimension: 2,
        m: 16,
        ef_construction: 100,
        normalize_vectors: false,
        distance_metric: DistanceMetric::Euclidean,
        ..Default::default()
    };
    let writer_config = VectorIndexWriterConfig {
        parallel_build: true,
        ..Default::default()
    };

    // 1. Initial Build
    {
        let storage_config = FileStorageConfig::new(path);
        let storage = Arc::new(FileStorage::new(path, storage_config)?);

        let mut writer = HnswIndexWriter::with_storage(
            index_config.clone(),
            writer_config.clone(),
            index_name,
            storage,
        )?;

        writer.add_vectors(vec![(
            1,
            "doc1".to_string(),
            Vector::new(vec![1.0f32, 0.0]),
        )])?;
        writer.finalize()?;
        writer.write()?;
    }

    // 2. Load and Append
    {
        let storage_config = FileStorageConfig::new(path);
        let storage = Arc::new(FileStorage::new(path, storage_config)?);

        let mut writer = HnswIndexWriter::load(
            index_config.clone(),
            writer_config.clone(),
            storage,
            index_name,
        )?;

        writer.add_vectors(vec![(
            2,
            "doc2".to_string(),
            Vector::new(vec![0.0f32, 1.0]),
        )])?;
        writer.finalize()?;
        writer.write()?;
    }

    // 3. Search
    {
        let storage_config = FileStorageConfig::new(path);
        let storage = Arc::new(FileStorage::new(path, storage_config)?);

        let reader = Arc::new(HnswIndexReader::load(
            storage.clone(),
            index_name,
            DistanceMetric::Euclidean,
        )?);

        // Create Searcher
        let searcher = HnswSearcher::new(reader)?;

        // Search for closest to [1.0, 0.0] -> should be doc1
        let req1 = VectorIndexQuery::new(Vector::new(vec![1.0f32, 0.0])).top_k(1);
        let results1 = searcher.search(&req1)?;
        assert_eq!(results1.len(), 1, "Expected 1 result for doc1");
        assert_eq!(results1.results[0].doc_id, 1, "Expected doc1");

        // Search for closest to [0.0, 1.0] -> should be doc2
        // If append works, doc2 is in the graph.
        let req2 = VectorIndexQuery::new(Vector::new(vec![0.0f32, 1.0])).top_k(1);
        let results2 = searcher.search(&req2)?;
        assert_eq!(results2.len(), 1, "Expected 1 result for doc2");
        assert_eq!(results2.results[0].doc_id, 2, "Expected doc2");
    }

    Ok(())
}