Crate phago_vectors 

Phago Vectors

Vector database adapters for Phago biological computing framework.

This crate provides a unified interface for storing and searching embeddings across different vector database backends.

Supported Backends

Backend	Feature Flag	Description
In-Memory	(default)	Simple brute-force search, good for testing
Qdrant	qdrant	High-performance vector database
Pinecone	pinecone	Managed vector database service
Weaviate	weaviate	Open-source vector search engine

Quick Start

use phago_vectors::{VectorStore, InMemoryStore, VectorRecord};

#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    let store = InMemoryStore::new(384); // 384-dimensional vectors

    // Store a vector
    let record = VectorRecord::new("doc1", vec![0.1; 384])
        .with_metadata("title", "Introduction to Cells");
    store.upsert(vec![record]).await?;

    // Search for similar vectors
    let results = store.search(&[0.1; 384], 5).await?;
    for result in results {
        println!("{}: {:.3}", result.id, result.score);
    }

    Ok(())
}

Feature Flags

# Use Qdrant backend
phago-vectors = { version = "0.6", features = ["qdrant"] }

# Use all backends
phago-vectors = { version = "0.6", features = ["all"] }

Re-exports

pub use memory::InMemoryStore;

Modules

memory

In-memory vector store implementation.

util

Utility functions for vector operations.

Structs

SearchResult

A search result from the vector store.

VectorRecord

A vector record to store in the database.

VectorStoreConfig

Configuration for creating a vector store.

Enums

BackendConfig

Backend-specific configuration.

DistanceMetric

Distance metric for similarity search.

VectorError

Errors that can occur when working with vector stores.

Traits

VectorStore

Abstract interface for vector storage and search.

Functions

create_store

Create a vector store from configuration.

Type Aliases

VectorResult

Result type for vector operations.