Struct PostingBlockHeader 

pub struct PostingBlockHeader {
    pub num_entries: u16,
    pub bits_per_delta: u8,
    pub min_offset: u32,
    pub max_offset: u32,
    pub max_weight: f32,
}

Header read from the first 16 bytes of a serialized block.

Terminology: “offset”

Throughout this module, “offset” means a document’s segment offset ID — its u32 position within a compacted segment. Posting lists are sorted by offset so that block-max pruning can iterate documents in order and merge across dimensions.

Layout (little-endian, 16 bytes total)

[0..2]   num_entries   : u16  — number of (offset, weight) pairs
[2]      bits_per_delta: u8   — bits per bitpacked delta (0xFF = directory)
[3]      _reserved     : u8   — reserved for future format versioning
[4..8]   min_offset    : u32  — smallest doc offset in this block
[8..12]  max_offset    : u32  — largest doc offset in this block
[12..16] max_weight    : f32  — largest weight in this block (for pruning)

Fields

num_entries: u16

Number of (offset, weight) pairs in the block.

bits_per_delta: u8

Bits per delta for bitpacked offset decompression. Set to 0xFF for directory blocks (different body layout).

min_offset: u32

Smallest document offset id in this block.

max_offset: u32

Largest document offset id in this block.

max_weight: f32

Largest weight in this block. Used by block-max pruning to skip entire blocks whose max contribution cannot beat the threshold.

Implementations

impl PostingBlockHeader

pub fn is_directory(&self) -> bool

Trait Implementations

impl Clone for PostingBlockHeader

fn clone(&self) -> PostingBlockHeader

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Copy for PostingBlockHeader

impl Debug for PostingBlockHeader

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

Formats the value using the given formatter.