phago-distributed 1.0.0

Distributed colony implementation for horizontal scaling
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

//! Distributed query implementation.
//!
//! This module handles queries that span multiple shards using a
//! two-phase TF-IDF approach:
//!
//! 1. **Scatter (Phase 1)**: Get local term frequencies from all shards
//! 2. **Gather (Phase 2)**: Aggregate to global document frequencies
//! 3. **Scatter (Phase 3)**: Compute TF-IDF with global DF
//! 4. **Gather (Phase 4)**: Merge top-k results

mod distributed;

pub use distributed::{DistributedHybridConfig, DistributedQueryEngine};

use crate::types::*;

/// Tokenizer matching phago-rag's tokenizer.
///
/// Performs case-insensitive tokenization with stopword removal
/// and minimum token length filtering.
pub fn tokenize(text: &str) -> Vec<String> {
    let stopwords: std::collections::HashSet<&str> = [
        "the", "a", "an", "is", "are", "was", "were", "be", "been", "being", "have", "has", "had",
        "do", "does", "did", "will", "would", "could", "should", "may", "might", "shall", "can",
        "need", "to", "of", "in", "for", "on", "with", "at", "by", "from", "as", "into", "through",
        "during", "before", "after", "above", "below", "between", "out", "off", "over", "under",
        "again", "further", "then", "once", "and", "but", "or", "if", "while", "what", "which",
        "who", "this", "that", "these", "those", "it", "its", "how",
    ]
    .iter()
    .cloned()
    .collect();

    text.to_lowercase()
        .split_whitespace()
        .filter(|w| w.len() >= 3 && !stopwords.contains(w))
        .map(|w| w.trim_matches(|c: char| !c.is_alphanumeric()).to_string())
        .filter(|w| w.len() >= 3)
        .collect()
}

/// Merge scored results from multiple shards.
///
/// Combines results from multiple shards, sorts by score (descending),
/// and truncates to the specified maximum.
pub fn merge_results(results: Vec<Vec<ScoredNode>>, max_results: usize) -> Vec<ScoredNode> {
    let mut all: Vec<ScoredNode> = results.into_iter().flatten().collect();
    all.sort_by(|a, b| {
        b.score
            .partial_cmp(&a.score)
            .unwrap_or(std::cmp::Ordering::Equal)
    });
    all.truncate(max_results);
    all
}

#[cfg(test)]
mod tests {
    use super::*;
    use phago_core::types::NodeId;

    #[test]
    fn test_tokenize_basic() {
        let tokens = tokenize("The cell membrane");
        assert!(tokens.contains(&"cell".to_string()));
        assert!(tokens.contains(&"membrane".to_string()));
        // "The" should be filtered as stopword
        assert!(!tokens.contains(&"the".to_string()));
    }

    #[test]
    fn test_tokenize_filters_short_words() {
        let tokens = tokenize("a is the on by");
        assert!(tokens.is_empty());
    }

    #[test]
    fn test_tokenize_trims_punctuation() {
        let tokens = tokenize("cell, membrane.");
        assert!(tokens.contains(&"cell".to_string()));
        assert!(tokens.contains(&"membrane".to_string()));
    }

    #[test]
    fn test_tokenize_lowercase() {
        let tokens = tokenize("CELL Membrane");
        assert!(tokens.contains(&"cell".to_string()));
        assert!(tokens.contains(&"membrane".to_string()));
    }

    #[test]
    fn test_merge_results_empty() {
        let results: Vec<Vec<ScoredNode>> = vec![];
        let merged = merge_results(results, 10);
        assert!(merged.is_empty());
    }

    #[test]
    fn test_merge_results_sorting() {
        let results = vec![
            vec![ScoredNode {
                node_id: NodeId::from_seed(1),
                label: "low".to_string(),
                score: 0.3,
                shard_id: ShardId::new(0),
            }],
            vec![ScoredNode {
                node_id: NodeId::from_seed(2),
                label: "high".to_string(),
                score: 0.9,
                shard_id: ShardId::new(1),
            }],
        ];
        let merged = merge_results(results, 10);
        assert_eq!(merged.len(), 2);
        assert_eq!(merged[0].label, "high");
        assert_eq!(merged[1].label, "low");
    }

    #[test]
    fn test_merge_results_truncates() {
        let results = vec![vec![
            ScoredNode {
                node_id: NodeId::from_seed(1),
                label: "a".to_string(),
                score: 0.9,
                shard_id: ShardId::new(0),
            },
            ScoredNode {
                node_id: NodeId::from_seed(2),
                label: "b".to_string(),
                score: 0.8,
                shard_id: ShardId::new(0),
            },
            ScoredNode {
                node_id: NodeId::from_seed(3),
                label: "c".to_string(),
                score: 0.7,
                shard_id: ShardId::new(0),
            },
        ]];
        let merged = merge_results(results, 2);
        assert_eq!(merged.len(), 2);
        assert_eq!(merged[0].label, "a");
        assert_eq!(merged[1].label, "b");
    }
}