laurus 0.5.0

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

use laurus::PrecomputedEmbedder;
use laurus::storage::file::FileStorageConfig;
use laurus::storage::{StorageConfig, StorageFactory};

use laurus::vector::DistanceMetric;
use laurus::vector::VectorSearchRequestBuilder;
use laurus::vector::VectorStore;
use laurus::vector::{FieldOption, FlatOption};
use laurus::vector::{VectorFieldConfig, VectorIndexConfig};
use laurus::{DataValue, Document};

use std::sync::Arc;
use tempfile::tempdir;

#[tokio::test(flavor = "multi_thread")]
async fn test_mmap_mode_basic_search() {
    let dir = tempdir().unwrap();
    let storage_path = dir.path().to_owned();

    let storage_config = StorageConfig::File(FileStorageConfig::new(storage_path.clone()));
    let storage = StorageFactory::create(storage_config).unwrap();

    // Configure a fields with Mmap loading
    let field_config = VectorFieldConfig {
        vector: Some(FieldOption::Flat(FlatOption {
            dimension: 3,
            distance: DistanceMetric::Cosine,
            base_weight: 1.0,
            quantizer: None,
            embedder: None,
        })),
        lexical: None,
    };

    let config = VectorIndexConfig::builder()
        .embedder(Arc::new(PrecomputedEmbedder::new()))
        .field("mmap_field", field_config)
        .build()
        .unwrap();

    let engine = VectorStore::new(storage, config).unwrap();

    // Add vectors
    let vectors = vec![
        vec![1.0, 0.0, 0.0],
        vec![0.0, 1.0, 0.0],
        vec![0.0, 0.0, 1.0],
    ];

    for (i, vec_data) in vectors.into_iter().enumerate() {
        let doc = Document::builder()
            .add_field("mmap_field", DataValue::Vector(vec_data))
            .build();
        engine
            .upsert_document_by_internal_id((i + 1) as u64, doc)
            .await
            .unwrap();
    }
    engine.commit().await.unwrap();

    let query_vector = vec![1.0, 0.1, 0.0];
    let request = VectorSearchRequestBuilder::new()
        .add_vector("mmap_field", query_vector)
        .limit(2)
        .build();

    let results = engine.search(request).unwrap();

    assert_eq!(results.hits.len(), 2);
}

#[tokio::test(flavor = "multi_thread")]
async fn test_mmap_mode_persistence_reload() {
    let dir = tempdir().unwrap();
    let storage_path = dir.path().to_owned();

    {
        let storage_config = StorageConfig::File(FileStorageConfig::new(storage_path.clone()));
        let storage = StorageFactory::create(storage_config).unwrap();

        let field_config = VectorFieldConfig {
            vector: Some(FieldOption::Flat(FlatOption {
                dimension: 3,
                distance: DistanceMetric::Cosine,
                base_weight: 1.0,
                quantizer: None,
                embedder: None,
            })),
            lexical: None,
        };

        let config = VectorIndexConfig::builder()
            .embedder(Arc::new(PrecomputedEmbedder::new()))
            .field("mmap_field", field_config)
            .build()
            .unwrap();

        let engine = VectorStore::new(storage, config).unwrap();

        let vectors = vec![vec![1.0, 0.0, 0.0], vec![0.0, 1.0, 0.0]];

        for (i, vec_data) in vectors.into_iter().enumerate() {
            let doc = Document::builder()
                .add_field("mmap_field", DataValue::Vector(vec_data))
                .build();
            engine
                .upsert_document_by_internal_id((i + 1) as u64, doc)
                .await
                .unwrap();
        }
        engine.commit().await.unwrap();
    }

    // Re-open
    {
        let storage_config = StorageConfig::File(FileStorageConfig::new(storage_path.clone()));
        let storage = StorageFactory::create(storage_config).unwrap();

        let field_config = VectorFieldConfig {
            vector: Some(FieldOption::Flat(FlatOption {
                dimension: 3,
                distance: DistanceMetric::Cosine,
                base_weight: 1.0,
                quantizer: None,
                embedder: None,
            })),
            lexical: None,
        };

        let config = VectorIndexConfig::builder()
            .embedder(Arc::new(PrecomputedEmbedder::new()))
            .field("mmap_field", field_config)
            .build()
            .unwrap();

        let engine = VectorStore::new(storage, config).unwrap();

        // IMPORTANT: In Mmap mode, vectors are LOADED from file on demand.
        // If file persistence works, search should find them.

        let query_vector = vec![0.0, 1.0, 0.0];
        let request = VectorSearchRequestBuilder::new()
            .add_vector("mmap_field", query_vector)
            .limit(1)
            .build();

        let results = engine.search(request).unwrap();

        assert_eq!(results.hits.len(), 1);
        // We expect it to match the second vector (doc_id=2, vector [0,1,0]).
        assert_eq!(
            results.hits[0].doc_id, 2,
            "Top result should be doc_id=2 (exact match for [0,1,0])"
        );
    }
}