codelens-engine 1.13.6

Harness-native Rust MCP server for code intelligence with generated surface governance, hybrid retrieval, and mutation-gated workflows
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

//! Semantic search using fastembed + sqlite-vec.
//! Gated behind the `semantic` feature flag.

use crate::embedding_types::{EmbeddingIndexInfo, EmbeddingRuntimeInfo, SemanticMatch};
use fastembed::TextEmbedding;
use serde::Serialize;
use std::sync::Mutex;

// ── Sub-modules ───────────────────────────────────────────────────────
mod cache;
mod chunk_ops;
mod engine_impl;
pub(super) mod ffi;
mod prompt;
mod runtime;
mod vec_store;

use cache::TextEmbeddingCache;
use vec_store::SqliteVecStore;

// ── Public re-exports ─────────────────────────────────────────────────
pub use chunk_ops::{CategoryScore, DuplicatePair, OutlierSymbol};
pub use prompt::auto_sparse_should_enable;
pub use runtime::{
    configured_embedding_model_name, configured_embedding_runtime_info,
    configured_embedding_runtime_preference, configured_embedding_threads,
    embedding_model_assets_available,
};

// ── Internal re-exports used by sibling sub-modules ───────────────────
// vec_store.rs uses embedding_to_bytes via `super::`
pub(super) use chunk_ops::embedding_to_bytes;
// engine_impl.rs uses these constants via `super::`
pub(super) use runtime::{CHANGED_FILE_QUERY_CHUNK, DEFAULT_DUPLICATE_SCAN_BATCH_SIZE};

// ── Test-only re-exports (for tests.rs via `use super::*`) ────────────
#[cfg(test)]
pub(super) use crate::project::ProjectRoot;
#[cfg(test)]
pub(super) use chunk_ops::duplicate_pair_key;
#[cfg(test)]
pub(super) use prompt::{
    auto_hint_mode_enabled, auto_hint_should_enable, build_embedding_text,
    contains_format_specifier, extract_api_calls, extract_api_calls_inner, extract_body_hint,
    extract_comment_body, extract_leading_doc, extract_nl_tokens, extract_nl_tokens_inner,
    hint_char_budget, hint_line_budget, is_nl_shaped, is_static_method_ident, is_test_only_symbol,
    language_supports_nl_stack, language_supports_sparse_weighting, looks_like_error_or_log_prefix,
    looks_like_meta_annotation, nl_tokens_enabled, should_reject_literal_strict,
    strict_comments_enabled, strict_literal_filter_enabled,
};
#[cfg(test)]
pub(super) use runtime::{
    CODESEARCH_MODEL_NAME, DEFAULT_MACOS_EMBED_BATCH_SIZE, embed_batch_size,
    recommended_embed_threads, requested_embedding_model_override, resolve_model_dir,
};

// ── Core engine struct ───────────────────────────────────────────────────

pub struct EmbeddingEngine {
    model: Mutex<TextEmbedding>,
    store: SqliteVecStore,
    model_name: String,
    runtime_info: EmbeddingRuntimeInfo,
    text_embed_cache: Mutex<TextEmbeddingCache>,
    indexing: std::sync::atomic::AtomicBool,
}

#[derive(Debug, Clone, Serialize, PartialEq, Eq)]
pub struct QueryEmbeddingCacheStats {
    pub enabled: bool,
    pub entries: usize,
    pub max_entries: usize,
}

#[derive(Debug, Clone, Default, Serialize, PartialEq, Eq)]
pub struct EmbeddingFreshnessReport {
    pub checked_files: usize,
    pub unchanged_files: usize,
    pub refreshed_files: usize,
    pub removed_files: usize,
    pub skipped_new_files: usize,
    pub indexed_symbols: usize,
}

// ── Tests ─────────────────────────────────────────────────────────────
#[cfg(test)]
mod tests;