Module search 

Search Infrastructure

Full-text search across items using RediSearch and MySQL.

Architecture

QueryBuilder (AST)
    ↓
    ├─→ RediSearchTranslator → FT.SEARCH syntax
    └─→ SqlTranslator → MySQL JSON_EXTRACT queries

Index Registration

sync-engine stays “dumb” about content but provides index lifecycle management:

// Register schema at startup using builder pattern
let index = SearchIndex::new("users", "crdt:users:")
    .text_sortable("name")
    .text("email")
    .numeric_sortable("age")
    .tag("roles");

engine.create_search_index(index).await.unwrap();

// RediSearch auto-indexes matching JSON documents
// Search using Query AST
let query = Query::field_eq("name", "Alice");
let results = engine.search("users", &query).await.unwrap();

Query Language (Lucene/RediSearch syntax)

@name:Alice               - Field equals
@age:[25 40]              - Numeric range
@tags:{rust|database}     - Tag membership (OR)
@name:Alice AND @age:[25 40]  - Boolean AND
@status:active | @status:pending  - Boolean OR
-@deleted:true            - Boolean NOT
@name:*alice*             - Wildcard contains
@name:ali*                - Prefix match
@name:%alice%             - Fuzzy match (Levenshtein)
(filter)=>[KNN 10 @embedding $blob]  - Vector KNN search

Structs

FieldQuery

Field query

IndexManager

Index manager for RediSearch

Query

Search query AST

QueryBuilder

Builder for complex queries

RediSearchTranslator

RediSearch query translator

SearchCache

Search result cache with merkle-based invalidation

SearchCacheStats

Cache statistics

SearchField

Search field definition

SearchIndex

Search index definition

SqlQuery

SQL query result with parameterized placeholders

SqlTranslator

SQL query translator for JSON column queries

TranslatedQuery

Result of translating a query that may contain vector search. Contains the query string and any binary parameters needed for vector search.

VectorParams

Parameters for RediSearch vector fields.

VectorQuery

Vector similarity search query (KNN)

Enums

DistanceMetric

Distance metric for vector similarity search.

FieldOperator

Field comparison operator

QueryNode

Query AST node

QueryValue

Query value type

SearchFieldType

Search field types supported by RediSearch

SqlParam

SQL parameter value

VectorAlgorithm

Vector search algorithm for RediSearch. HNSW provides faster queries with higher memory usage. FLAT provides exact results with O(n) query time.