Struct IndexFields 

pub struct IndexFields {
    pub schema: Schema,
    pub id: Field,
    pub mem: Field,
    pub entity_type: Field,
    pub title: Field,
    pub sections: BTreeMap<String, Field>,
    pub metadata: BTreeMap<String, Field>,
    pub metadata_text: Field,
}

Tantivy schema + pre-resolved field handles for a single mem index.

Fields

schema: Schema

id: Field

mem: Field

entity_type: Field

title: Field

sections: BTreeMap<String, Field>

Section-key → tantivy text field. Keys mirror the section keys the parser writes into Entity.sections.

metadata: BTreeMap<String, Field>

Metadata field key → tantivy STRING field. Only filterable fields are added so the index doesn’t carry fields no caller will query.

metadata_text: Field

ONE tokenized field carrying every entity’s metadata KEYS and VALUES (“key value” lines), built from the entity — no filterable declaration needed, present even for a mem whose schema declares no metadata fields. Participates in the free-text query at a weight below title/sections so identifier-shaped values are findable without enum/date tokens swamping prose ranking. The untokenized meta_<key> fields above keep their exact-match filter role untouched.

Implementations

impl IndexFields

pub fn build(mem_schema: Option<&Arc<Schema>>) -> Self

Build the tantivy schema for a mem given the resolved mem schema.

Falls back to a minimal fixed-field schema when mem_schema is None — useful for read-mems whose pinned schema failed to resolve (we still index id/mem/title so structural filters work).

Trait Implementations

impl Clone for IndexFields

fn clone(&self) -> IndexFields

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for IndexFields

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

Formats the value using the given formatter.