Crate dakera_client

Expand description

Dakera Rust Client SDK

A high-level Rust client for interacting with Dakera AI Agent Memory Platform.

§Quick Start (HTTP)

use dakera_client::{DakeraClient, UpsertRequest, QueryRequest};

#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    // Create a client
    let client = DakeraClient::new("http://localhost:3000")?;

    // Check health
    let health = client.health().await?;
    println!("Server healthy: {}", health.healthy);

    // Upsert vectors
    let request = UpsertRequest {
        vectors: vec![
            dakera_client::Vector {
                id: "vec1".to_string(),
                values: vec![0.1, 0.2, 0.3, 0.4],
                metadata: None,
            },
        ],
    };
    client.upsert("my-namespace", request).await?;

    // Query for similar vectors
    let query = QueryRequest {
        vector: vec![0.1, 0.2, 0.3, 0.4],
        top_k: 10,
        filter: None,
        include_metadata: true,
        include_vectors: false,
        distance_metric: Default::default(),
        consistency: Default::default(),
        staleness_config: None,
    };
    let results = client.query("my-namespace", query).await?;

    for match_ in results.matches {
        println!("ID: {}, Score: {}", match_.id, match_.score);
    }

    Ok(())
}

§gRPC Client with Connection Pooling

Enable the grpc feature for high-performance gRPC communication:

use dakera_client::grpc::{GrpcClient, GrpcClientConfig, GrpcConnectionPool};

#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    // Single client with HTTP/2 multiplexing
    let config = GrpcClientConfig::default()
        .with_endpoint("http://localhost:50051")
        .with_concurrency_limit(100);
    let client = GrpcClient::connect(config).await?;

    // Or use a connection pool for even higher throughput
    let pool = GrpcConnectionPool::new(GrpcClientConfig::default(), 4).await?;
    let client = pool.get();

    Ok(())
}

Re-exports§

pub use events::DakeraEvent;
pub use events::MemoryEvent;
pub use events::OpStatus;
pub use events::VectorMutationOp;
pub use admin::BackupInfo;
pub use admin::BackupListResponse;
pub use admin::CacheStats;
pub use admin::ClusterStatus;
pub use admin::CreateBackupRequest;
pub use admin::CreateBackupResponse;
pub use admin::IndexStats;
pub use admin::IndexStatsResponse;
pub use admin::NodeInfo;
pub use admin::NodeListResponse;
pub use admin::QuotaConfig;
pub use admin::QuotaListResponse;
pub use admin::QuotaStatus;
pub use admin::RestoreBackupRequest;
pub use admin::RestoreBackupResponse;
pub use admin::RuntimeConfig;
pub use admin::SlowQueryListResponse;
pub use agents::AgentStats;
pub use agents::AgentSummary;
pub use analytics::AnalyticsOverview;
pub use analytics::LatencyAnalytics;
pub use analytics::StorageAnalytics;
pub use analytics::ThroughputAnalytics;
pub use keys::ApiKeyUsageResponse;
pub use keys::CreateKeyRequest;
pub use keys::CreateKeyResponse;
pub use keys::KeyInfo;
pub use keys::ListKeysResponse;
pub use keys::RotateKeyResponse;
pub use knowledge::AgentNetworkEdge;
pub use knowledge::AgentNetworkInfo;
pub use knowledge::AgentNetworkNode;
pub use knowledge::AgentNetworkStats;
pub use knowledge::CrossAgentNetworkRequest;
pub use knowledge::CrossAgentNetworkResponse;
pub use knowledge::DeduplicateRequest;
pub use knowledge::DeduplicateResponse;
pub use knowledge::FullKnowledgeGraphRequest;
pub use knowledge::KnowledgeEdge;
pub use knowledge::KnowledgeGraphRequest;
pub use knowledge::KnowledgeGraphResponse;
pub use knowledge::KnowledgeNode;
pub use knowledge::SummarizeRequest;
pub use knowledge::SummarizeResponse;
pub use reqwest;

Modules§

admin: Admin operations for the Dakera client.
agents: Agent management for the Dakera client.
analytics: Analytics operations for the Dakera client.
events: SSE streaming event types and client support (CE-1).
keys: API Key management for the Dakera client.
knowledge: Knowledge graph operations for the Dakera client.
memory: Memory-oriented client methods for Dakera AI Agent Memory Platform

Structs§

ActualStats: Actual execution statistics (when execute=true)
AggregationGroup: Single group in aggregation results
AggregationRequest: Request for aggregation query (Turbopuffer-inspired)
AggregationResponse: Response for aggregation query
BatchQueryItem: A single query within a batch request
BatchQueryRequest: Batch query request - execute multiple queries in parallel
BatchQueryResponse: Batch query response
BatchQueryResult: Results for a single query within a batch
BatchQueryTextRequest: Request to execute multiple text queries with automatic embedding in a single call.
BatchQueryTextResponse: Response from a batch text query operation.
ColumnUpsertRequest: Column-based upsert request (Turbopuffer-inspired)
CompactionRequest: Compaction request
CompactionResponse: Compaction response
ComponentHealth: Component health status
ConfigureNamespaceRequest: Request body for PUT /v1/namespaces/:namespace — upsert semantics (v0.6.0).
ConfigureNamespaceResponse: Response from PUT /v1/namespaces/:namespace.
CostEstimate: Cost estimation
CreateNamespaceRequest: Request to create a new namespace.
DakeraClient: Dakeraclient for interacting with the vector database
DakeraClientBuilder: Builder for DakeraClient
DeleteRequest: Delete vectors request
DeleteResponse: Delete response
Document: A document for full-text indexing
ExecutionStage: A stage in query execution
ExportRequest: Request to export vectors from a namespace with pagination
ExportResponse: Response from export operation
ExportedVector: A single exported vector record
FetchRequest: Request to fetch vectors by their IDs.
FetchResponse: Response from a fetch-by-ID operation.
FullTextMatch: Full-text search result
FullTextSearchRequest: Full-text search request
FullTextSearchResponse: Full-text search response
FullTextStats: Full-text index statistics
HealthResponse: Health check response
HealthStatus: Health status for a component
HybridSearchRequest: Hybrid search request combining vector and full-text search
HybridSearchResponse: Hybrid search response
IndexAlternative: Alternative index that was considered
IndexDocumentsRequest: Index documents request
IndexDocumentsResponse: Index documents response
IndexSelection: Index selection details
IndexStatistics: Index statistics
JobInfo: Background job information
ListNamespacesResponse: List namespaces response
Match: A match result from a query
MultiVectorSearchRequest: Request for multi-vector search with positive and negative vectors
MultiVectorSearchResponse: Response from multi-vector search
MultiVectorSearchResult: Single result from multi-vector search
NamespaceInfo: Namespace information
QueryExplainRequest: Query explain request
QueryExplainResponse: Query explain response - detailed execution plan
QueryParams: Query parameters for reference
QueryRequest: Query request for vector similarity search
QueryResponse: Query response
QueryTextRequest: Request to query using natural language text with automatic embedding.
ReadinessResponse: Readiness check response
Recommendation: Performance recommendation
ResourceUsage: Resource usage metrics
StalenessConfig: Configuration for bounded staleness reads
SystemDiagnostics: System diagnostics
SystemInfo: System information
TextDocument: A text document to upsert with automatic embedding generation.
TextQueryResponse: Response from a text query operation.
TextSearchResult: A single text search result.
TextUpsertResponse: Response from a text upsert operation.
UnifiedQueryRequest: Unified query request with flexible ranking options (Turbopuffer-inspired)
UnifiedQueryResponse: Unified query response
UnifiedSearchResult: Single result from unified query
UpsertRequest: Upsert vectors request
UpsertResponse: Upsert response
UpsertTextRequest: Request to upsert text documents with automatic embedding.
Vector: A vector with ID and optional metadata
WarmCacheRequest: Cache warming request with priority hints
WarmCacheResponse: Cache warming response

Enums§

AccessPatternHint: Access pattern hint for cache optimization
AggregateFunction: Aggregate function for computing values across documents
ClientError: Errors that can occur when using the Dakera client
DistanceMetric: Distance metric for similarity search
EmbeddingModel: Supported embedding models for text-based operations.
ExplainQueryType: Query type for explain
RankBy: Ranking function for unified query API Supports vector search (ANN/kNN), full-text BM25, and attribute ordering
ReadConsistency: Read consistency level for queries (Turbopuffer-inspired)
SortDirection: Sort direction for attribute ordering
VectorSearchMethod: Vector search method for unified query
WarmingPriority: Priority level for cache warming operations
WarmingTargetTier: Target cache tier for warming

Type Aliases§

Result: Result type alias for Dakera client operations

Crate dakera_client

Crate dakera_client Copy item path

§Quick Start (HTTP)

§gRPC Client with Connection Pooling

Re-exports§

Modules§

Structs§

Enums§

Type Aliases§

Crate dakera_client