Struct Index 

pub struct Index { /* private fields */ }

MinHash + LSH index. Insert documents, then query with a fresh string.

Implementations

impl Index

pub fn new(cfg: Config) -> Result<Self>

Build an empty index.

pub fn config(&self) -> &Config

The active config (read-only).

pub fn len(&self) -> usize

Number of indexed documents.

pub fn is_empty(&self) -> bool

True iff no documents are indexed.

pub fn insert(&mut self, id: impl Into<String>, text: &str) -> Result<()>

Insert a document. Duplicate ids are allowed but discouraged; the index does not deduplicate them itself.

pub fn signature(&self, text: &str) -> Vec<u32>

Compute the MinHash signature for a string. Pure function of text and the index config; no insertion side effects.

pub fn jaccard(sig_a: &[u32], sig_b: &[u32]) -> f64

Estimated Jaccard similarity between two signatures of the same length.

pub fn near_duplicates(&self, text: &str, min_similarity: f64) -> Vec<Hit>

Return all indexed documents whose estimated Jaccard similarity to text is >= min_similarity. Sorted by similarity descending.

pub fn save<P: AsRef<Path>>(&self, path: P) -> Result<()>

Persist the index to a JSON file. Stores cfg, the hash family (a, b), and the per-doc signatures. The runtime bands map and band_hasher are reconstructed on load (band_hasher is keyed off cfg.seed, so band hashes round-trip identically).

pub fn load<P: AsRef<Path>>(path: P) -> Result<Self>

Reverse of Index::save. Validates the loaded cfg (rejects configs that wouldn’t construct cleanly) before rebuilding the in-memory bands map.