Struct SymbolCache 

pub struct SymbolCache { /* private fields */ }

Symbol cache for storing and retrieving parsed symbols

Implementations

impl SymbolCache

pub fn open(cache_dir: &Path) -> Result<Self>

Open a symbol cache at the given cache directory

pub fn get( &self, file_path: &str, file_hash: &str, ) -> Result<Option<Vec<SearchResult>>>

Get cached symbols for a file (returns None if not cached or hash mismatch)

pub fn batch_get( &self, files: &[(String, String)], ) -> Result<Vec<(String, Option<Vec<SearchResult>>)>>

Get cached symbols for multiple files in one transaction (batch read)

This is significantly faster than calling get() repeatedly because:

• Opens only ONE database connection instead of N
• Reuses ONE prepared statement instead of creating N
• Executes in ONE transaction instead of N

Returns results in the same order as input. None means cache miss or hash mismatch.

pub fn batch_get_with_kind( &self, file_ids: &[(i64, String, String)], kind_filter: Option<SymbolKind>, ) -> Result<HashMap<i64, Vec<SearchResult>>>

Get cached symbols for multiple files with optional kind filtering

Uses integer file_ids for fast batch retrieval, then filters by kind in Rust. This avoids the cache miss detection bug that occurs with SQL-level filtering.

Automatically chunks large batches to avoid SQLite parameter limits (999 max).

Parameters:

• file_ids: Vec of (file_id, file_hash, file_path) tuples
• kind_filter: Optional symbol kind to filter by (applied in Rust after retrieval)

Returns HashMap of file_id → symbols for cache hits.

pub fn set( &self, file_path: &str, file_hash: &str, symbols: &[SearchResult], ) -> Result<()>

Store symbols for a file using file_id

pub fn batch_set( &self, entries: &[(String, String, Vec<SearchResult>)], ) -> Result<()>

Batch store symbols for multiple files in a single transaction

pub fn clear(&self) -> Result<()>

Clear all cached symbols

pub fn stats(&self) -> Result<SymbolCacheStats>

Get cache statistics

pub fn cleanup_stale(&self) -> Result<usize>

Remove symbols for files that are no longer in the index

This cleanup operation removes stale symbol cache entries for files that have been deleted or are no longer indexed.

Note: With foreign key constraints (CASCADE DELETE), this should rarely find anything to clean up, but it’s useful for manual verification.