#[repr(u16)]pub enum SectionKind {
Show 16 variants
Nodes = 1,
Segs = 2,
EdgeIndex = 3,
Edges = 4,
Labels = 5,
Lexicon = 6,
Postings = 7,
Ranks = 8,
InEdgeIndex = 9,
InEdges = 10,
AttrIndex = 11,
Attrs = 12,
Blobs = 13,
LexOrder = 14,
Trigrams = 15,
TriPostings = 16,
}Expand description
Section kinds (discriminants are burned; do not renumber).
Variants§
Nodes = 1
Fixed 72-B node records sorted by url_key.
Segs = 2
Fixed 24-B segment records grouped per node in document order.
EdgeIndex = 3
CSR offsets: (node_count + 1) × u32 into Edges.
Edges = 4
Fixed 8-B edge records, per-node runs sorted by destination id.
Labels = 5
Length-prefixed UTF-8 label heap.
Lexicon = 6
Fixed 24-B token records sorted by token_hash.
Postings = 7
LEB128 delta-encoded ascending node-id postings.
Ranks = 8
Fixed 2-B per-node rank percentiles (parallel to Nodes).
InEdgeIndex = 9
CSR offsets: (node_count + 1) × u32 into InEdges.
InEdges = 10
Fixed 8-B inbound records (source id, type, flags, weight), per-node
runs sorted by source id: the reverse of Edges.
AttrIndex = 11
CSR offsets: (node_count + 1) × u32 into Attrs.
Attrs = 12
Fixed 12-B attribute records, per-node runs sorted by key.
Blobs = 13
UTF-8 attribute values, addressed by (offset, length).
LexOrder = 14
token_count × u32 lexicon indices in token string order: prefix
lookups are a binary search.
Trigrams = 15
Fixed 12-B trigram records sorted by trigram: substring lookups intersect the postings of a term’s trigrams.
TriPostings = 16
Ascending u32 lexicon indices per trigram.