Struct InMemoryIndex 

pub struct InMemoryIndex {
    pub indexes: HashMap<String, IndexState>,
    pub position_encoding: PositionEncoding,
    pub dictionary: Option<DictionaryConfig>,
}

In-memory inverted index supporting exact, pinyin, and fuzzy search over documents.

Fields

indexes: HashMap<String, IndexState>

position_encoding: PositionEncoding

dictionary: Option<DictionaryConfig>

Implementations

impl InMemoryIndex

pub fn with_position_encoding(encoding: PositionEncoding) -> Self

Create an index that returns match spans in the given encoding.

pub fn with_dictionary_config(dictionary: DictionaryConfig) -> Self

Create an index that uses a custom dictionary for tokenization.

pub fn set_position_encoding(&mut self, encoding: PositionEncoding)

Set the encoding (bytes or UTF-16) used when returning match spans.

pub fn set_dictionary_config(&mut self, dictionary: Option<DictionaryConfig>)

Swap in or remove a dictionary config for future tokenization.

pub fn add_doc( &mut self, index_name: &str, doc_id: &str, text: &str, index: bool, )

Add or replace a document in an index. Set index to false to stage content without tokenization (doc will exist but not be searchable).

pub fn remove_doc(&mut self, index_name: &str, doc_id: &str)

Remove a document and its postings from an index.

pub fn get_doc(&self, index_name: &str, doc_id: &str) -> Option<String>

Fetch raw document content by id, if present.

pub fn take_dirty_and_deleted( &mut self, ) -> (Vec<(String, String, String, i64)>, HashMap<String, HashSet<String>>)

Return and clear the sets of dirty and deleted docs for persistence.

pub fn has_unpersisted_changes(&self, index_name: Option<&str>) -> bool

Returns true if the index has new changes awaiting persistence. Pass Some(name) to check a specific index or None to check all.

pub fn persist_if_dirty<E>( &mut self, index_name: &str, persist: impl FnMut(SnapshotData) -> Result<(), E>, ) -> Result<bool, E>

Persist the given index only if there are pending changes.

Returns Ok(true) if persistence was attempted (and succeeded), Ok(false) if skipped. The index is marked clean only after the provided callback returns Ok.

pub fn get_matches( &self, index_name: &str, doc_id: &str, query: &str, ) -> Vec<(u32, u32)>

Get byte/UTF-16 spans for a query’s terms within a document by re-tokenizing the query.

pub fn get_matches_for_terms( &self, index_name: &str, doc_id: &str, terms: &[String], ) -> Vec<(u32, u32)>

Get spans for specific terms within a document.

pub fn get_matches_for_matched_terms( &self, index_name: &str, doc_id: &str, terms: &[MatchedTerm], ) -> Vec<(u32, u32)>

Get spans for previously returned matched terms (e.g., from search_hits).

pub fn load_snapshot(&mut self, index_name: &str, snapshot: SnapshotData)

Load a snapshot into an index, expecting all auxiliary structures to be present.

pub fn get_snapshot_data(&self, index_name: &str) -> Option<SnapshotData>

Get a serializable snapshot of the given index, including postings.

impl InMemoryIndex

pub fn search(&self, index_name: &str, query: &str) -> Vec<(String, f64)>

Execute an auto-mode search and return doc ids with scores.

pub fn search_hits(&self, index_name: &str, query: &str) -> Vec<SearchHit>

Execute an auto-mode search and return full hits including matched terms.

pub fn search_with_mode( &self, index_name: &str, query: &str, mode: SearchMode, ) -> Vec<(String, f64)>

Execute a search in the specified mode and return doc ids with scores.

pub fn search_with_mode_hits( &self, index_name: &str, query: &str, mode: SearchMode, ) -> Vec<SearchHit>

Execute a search in the specified mode and return full hits including matched terms.

impl InMemoryIndex

pub fn snapshot_version() -> u32

Trait Implementations

impl Debug for InMemoryIndex

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

Formats the value using the given formatter.

impl Default for InMemoryIndex

fn default() -> Self

Returns the “default value” for a type.