ContentRetriever class

Adaptive content retrieval engine.

Selects the best retrieval strategy based on chat mode:

  • Question-driven modes use search (local → semantic → FTS → document order).
  • Sequential modes load all chunks in document order.

Retrieval cascade for question-driven modes:

  1. Local on-device vector search (Tier 3) — fastest, offline-capable
  2. Supabase pgvector semantic search (Tier 2) — requires network + embeddings
  3. Supabase full-text search (Tier 1) — keyword matching
  4. Document order fallback — load all chunks sequentially

Constructors

ContentRetriever(SupabaseDatasource _supabase)

Properties

hashCode → int
The hash code for this object.
no setterinherited
runtimeType → Type
A representation of the runtime type of the object.
no setterinherited

Methods

noSuchMethod(Invocation invocation) → dynamic
Invoked when a nonexistent method or property is accessed.
inherited
retrieveChunks({required String mode, String? topicId, String? userMessage, String? domainId}) → Future<List<ContentChunkModel>>
Retrieve the most relevant content chunks for a given mode.
setLocalEmbedding(LocalEmbeddingDatasource ds) → void
Set the local embedding datasource for on-device inference.
setLocalVectorStore(LocalVectorStore store) → void
Set the local vector store for on-device search.
toString() → String
A string representation of this object.
inherited

Operators

operator ==(Object other) → bool
The equality operator.
inherited