Struct MemoryBackend 

pub struct MemoryBackend { /* private fields */ }

In-memory FTS backend backed by HashMaps keyed by (tid, collection, …) tuples.

Uses RefCell for interior mutability so the FtsBackend trait can use &self uniformly (redb has its own transactional isolation).

Implementations

impl MemoryBackend

pub fn new() -> MemoryBackend

Trait Implementations

impl Debug for MemoryBackend

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

Formats the value using the given formatter.

impl Default for MemoryBackend

fn default() -> MemoryBackend

Returns the “default value” for a type.

impl FtsBackend for MemoryBackend

type Error = MemoryError

Error type for backend operations.

fn read_postings( &self, tid: u64, collection: &str, term: &str, ) -> Result<Vec<Posting>, <MemoryBackend as FtsBackend>::Error>

Read the posting list for a term in a collection.

fn write_postings( &self, tid: u64, collection: &str, term: &str, postings: &[Posting], ) -> Result<(), <MemoryBackend as FtsBackend>::Error>

Write/replace the posting list for a term in a collection.

fn remove_postings( &self, tid: u64, collection: &str, term: &str, ) -> Result<(), <MemoryBackend as FtsBackend>::Error>

Remove a term’s posting list entirely.

fn read_doc_length( &self, tid: u64, collection: &str, doc_id: Surrogate, ) -> Result<Option<u32>, <MemoryBackend as FtsBackend>::Error>

Read the document length (token count) for a document.

fn write_doc_length( &self, tid: u64, collection: &str, doc_id: Surrogate, length: u32, ) -> Result<(), <MemoryBackend as FtsBackend>::Error>

Write/replace the document length for a document.

fn remove_doc_length( &self, tid: u64, collection: &str, doc_id: Surrogate, ) -> Result<(), <MemoryBackend as FtsBackend>::Error>

Remove a document’s length entry.

fn collection_terms( &self, tid: u64, collection: &str, ) -> Result<Vec<String>, <MemoryBackend as FtsBackend>::Error>

Get all term names in a collection (for fuzzy matching).

fn collection_stats( &self, tid: u64, collection: &str, ) -> Result<(u32, u64), <MemoryBackend as FtsBackend>::Error>

Get total document count and sum of all document lengths for a collection. Returns (doc_count, total_token_sum).

fn increment_stats( &self, tid: u64, collection: &str, doc_len: u32, ) -> Result<(), <MemoryBackend as FtsBackend>::Error>

Increment collection stats after indexing a document. doc_len is the number of tokens in the newly indexed document.

fn decrement_stats( &self, tid: u64, collection: &str, doc_len: u32, ) -> Result<(), <MemoryBackend as FtsBackend>::Error>

Decrement collection stats after removing a document. doc_len is the token count of the removed document.

fn read_meta( &self, tid: u64, collection: &str, subkey: &str, ) -> Result<Option<Vec<u8>>, <MemoryBackend as FtsBackend>::Error>

Read a metadata blob by sub-key (e.g., “docmap”, “fieldnorms”, “analyzer”, “language”).

fn write_meta( &self, tid: u64, collection: &str, subkey: &str, value: &[u8], ) -> Result<(), <MemoryBackend as FtsBackend>::Error>

Write a metadata blob by sub-key.

fn write_segment( &self, tid: u64, collection: &str, segment_id: &str, data: &[u8], ) -> Result<(), <MemoryBackend as FtsBackend>::Error>

Write a segment blob. segment_id is a stable per-collection identifier (e.g., "L{level}:{id:016x}").

fn read_segment( &self, tid: u64, collection: &str, segment_id: &str, ) -> Result<Option<Vec<u8>>, <MemoryBackend as FtsBackend>::Error>

Read a segment blob. Returns None if not found.

fn list_segments( &self, tid: u64, collection: &str, ) -> Result<Vec<String>, <MemoryBackend as FtsBackend>::Error>

List all segment ids for a collection.

fn remove_segment( &self, tid: u64, collection: &str, segment_id: &str, ) -> Result<(), <MemoryBackend as FtsBackend>::Error>

Remove a segment blob.

fn purge_collection( &self, tid: u64, collection: &str, ) -> Result<usize, <MemoryBackend as FtsBackend>::Error>

Remove all entries for a collection. Returns count of removed entries.

fn purge_tenant( &self, tid: u64, ) -> Result<usize, <MemoryBackend as FtsBackend>::Error>

Remove all entries for a tenant across every collection. Returns count of removed entries. Implementations MUST use a structural drop (e.g., tuple range (tid, ..)..(tid+1, ..)) rather than a lexical-prefix scan.