Struct WandData 

pub struct WandData {
    pub total_docs: u64,
    pub total_tokens: u64,
    pub avg_doc_len: f32,
    pub bm25_k1: f32,
    pub bm25_b: f32,
    /* private fields */
}

Collection-level WAND data

Contains pre-computed statistics needed for efficient WAND query processing. This data is typically computed offline using hermes-tool term-stats and loaded at index open time.

Fields

total_docs: u64

Total number of documents in the collection

total_tokens: u64

Total number of tokens across all documents

avg_doc_len: f32

Average document length (tokens per document)

bm25_k1: f32

BM25 k1 parameter used for computing upper bounds

bm25_b: f32

BM25 b parameter used for computing upper bounds

Implementations

impl WandData

pub fn new(total_docs: u64, avg_doc_len: f32) -> Self

Create empty WAND data

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

Load WAND data from a JSON file

pub fn from_json_reader<R: Read>(reader: R) -> Result<Self>

Load WAND data from a JSON reader

pub fn from_json_bytes(bytes: &[u8]) -> Result<Self>

Load WAND data from JSON bytes

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

Save WAND data to a JSON file

pub fn to_json_writer<W: Write>(&self, writer: W) -> Result<()>

Write WAND data to a JSON writer

pub fn get_idf(&self, field: &str, term: &str) -> Option<f32>

Get IDF for a term in a field

Returns None if the term is not found in the pre-computed data. In that case, you should compute IDF on-the-fly using the segment’s document count and term document frequency.

pub fn get_term_info(&self, field: &str, term: &str) -> Option<&TermWandInfo>

Get full term info for a term in a field

pub fn get_upper_bound(&self, field: &str, term: &str) -> Option<f32>

Get upper bound score for a term

pub fn compute_idf(&self, df: u32) -> f32

Compute IDF for a term given its document frequency

Uses the BM25 IDF formula: log((N - df + 0.5) / (df + 0.5))

pub fn compute_upper_bound(&self, max_tf: u32, idf: f32) -> f32

Compute upper bound score for a term given max_tf and IDF

Uses conservative length normalization (assumes shortest possible document)

pub fn add_term( &mut self, field: &str, term: &str, df: u32, total_tf: u64, max_tf: u32, )

Add or update term statistics

pub fn num_terms(&self) -> usize

Get the number of terms in the WAND data

pub fn has_term(&self, field: &str, term: &str) -> bool

Check if WAND data is available for a term

Trait Implementations

impl Clone for WandData

fn clone(&self) -> WandData

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for WandData

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for WandData

fn default() -> Self

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for WandData

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for WandData

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.