Milvus Vector Store Provider

The Milvus provider implements the vectorstore.VectorStore interface using the Milvus v2 REST API. Milvus is an open-source vector database designed for scalable similarity search, supporting both self-hosted and managed (Zilliz Cloud) deployments.

Choose Milvus when you need a vector database designed for large-scale deployments with billions of vectors. Milvus supports multiple index types (IVF, HNSW, DiskANN), GPU-accelerated search, and horizontal scaling. Zilliz Cloud provides a managed option for teams that want Milvus capabilities without self-hosting.

Installation

go get github.com/lookatitude/beluga-ai/rag/vectorstore/providers/milvus

Start Milvus locally:

docker compose up -d  # Using the official Milvus docker-compose

Quick Start

package main

import (
    "context"
    "log"

    "github.com/lookatitude/beluga-ai/config"
    "github.com/lookatitude/beluga-ai/rag/vectorstore"
    "github.com/lookatitude/beluga-ai/schema"
    _ "github.com/lookatitude/beluga-ai/rag/vectorstore/providers/milvus"
)

func main() {
    store, err := vectorstore.New("milvus", config.ProviderConfig{
        BaseURL: "http://localhost:19530",
        Options: map[string]any{
            "collection": "documents",
            "dimension":  float64(1536),
        },
    })
    if err != nil {
        log.Fatal(err)
    }

    ctx := context.Background()

    docs := []schema.Document{
        {ID: "doc1", Content: "Go is a statically typed language", Metadata: map[string]any{"lang": "en"}},
    }
    embeddings := [][]float32{make([]float32, 1536)}

    err = store.Add(ctx, docs, embeddings)
    if err != nil {
        log.Fatal(err)
    }
}

Configuration

Parameter	Type	Default	Description
BaseURL	string	(required)	Milvus server URL
APIKey	string	(optional)	API key for Zilliz Cloud
Options["collection"]	string	documents	Collection name
Options["dimension"]	float64	1536	Vector dimensionality

Direct Construction

import (
    milvusstore "github.com/lookatitude/beluga-ai/rag/vectorstore/providers/milvus"
)

store := milvusstore.New("http://localhost:19530",
    milvusstore.WithCollection("my_collection"),
    milvusstore.WithDimension(768),
)

Metadata Filtering

Milvus supports expression-based filtering:

results, err := store.Search(ctx, queryVec, 10,
    vectorstore.WithFilter(map[string]any{
        "lang": "en",
    }),
)

Filters are translated to Milvus filter expressions (e.g., lang == "en").

Zilliz Cloud

Connect to Zilliz Cloud (managed Milvus) using the API key:

store, err := vectorstore.New("milvus", config.ProviderConfig{
    BaseURL: "https://in03-abc123.api.gcp-us-west1.zillizcloud.com",
    APIKey:  os.Getenv("ZILLIZ_API_KEY"),
    Options: map[string]any{
        "collection": "production",
    },
})

Search with Threshold

results, err := store.Search(ctx, queryVec, 10,
    vectorstore.WithThreshold(0.7),
)
for _, doc := range results {
    fmt.Printf("ID: %s, Score: %.4f\n", doc.ID, doc.Score)
}

Custom HTTP Client

store := milvusstore.New("http://localhost:19530",
    milvusstore.WithHTTPClient(customClient),
    milvusstore.WithCollection("my_collection"),
)