06. Milvus

Introduction to Milvus

Milvus is an open-source, highly scalable vector database optimized for storing and retrieving high-dimensional embeddings. It is particularly useful for applications such as semantic search, recommendation systems, and retrieval-augmented generation (RAG). Milvus supports large-scale vector searches and integrates well with machine learning frameworks.

Setting Up Milvus

1. Installing Milvus

To use Milvus, install the Milvus client:

Copy

pip install pymilvus

If you want to run a local instance of Milvus, you can use Docker:

Copy

docker run -d --name milvus -p 19530:19530 milvusdb/milvus:latest

This will start a Milvus instance locally, accessible on http://localhost:19530.

2. Creating a Milvus Client

Once installed, initialize a Milvus client in Python:

Copy

from pymilvus import connections

connections.connect(host="localhost", port="19530")

If you're using Milvus Cloud, replace localhost with your cloud endpoint and provide authentication credentials.

Integrating Milvus with LangChain

LangChain provides seamless integration with Milvus for vector-based storage and retrieval. The Milvus wrapper in LangChain simplifies adding and retrieving vector embeddings.

1. Creating a Milvus Collection

Before storing vectors, define a schema in Milvus:

Copy

from pymilvus import CollectionSchema, FieldSchema, DataType, Collection

fields = [
    FieldSchema(name="id", dtype=DataType.INT64, is_primary=True, auto_id=True),
    FieldSchema(name="embedding", dtype=DataType.FLOAT_VECTOR, dim=1536)
]
schema = CollectionSchema(fields, description="LangChainDocs collection")

collection = Collection("LangChainDocs", schema)

This creates a collection named LangChainDocs with a vector field of dimension 1536.

2. Storing Embeddings in Milvus

To store vectors, first generate embeddings using an embedding model (e.g., OpenAI or Hugging Face):

Copy

from langchain.embeddings import OpenAIEmbeddings
from langchain.vectorstores import Milvus

embeddings = OpenAIEmbeddings()
vector_db = Milvus(collection_name="LangChainDocs", embeddings=embeddings)

Now, store some text data in Milvus:

Copy

documents = ["This is a sample document.", "LangChain makes working with LLMs easier."]
vector_db.add_texts(texts=documents)

3. Performing Similarity Search

Retrieve documents similar to a given query:

Copy

query = "How does LangChain help with LLMs?"
results = vector_db.similarity_search(query, k=2)

for result in results:
    print(result.page_content)

This fetches the top 2 documents that are most semantically similar to the query.

Best Practices and Optimization

• Optimize Indexing: Use Milvus’s indexing strategies like IVF, HNSW, or PQ for better performance.

• Scale Efficiently: Deploy Milvus with distributed storage for handling large-scale datasets.

• Hybrid Search: Combine keyword and vector searches for more precise retrieval.

• Cloud Deployment: Consider using Milvus Cloud for better scalability and reliability.

Conclusion

Milvus is a powerful, open-source vector database designed for large-scale AI applications. Its integration with LangChain enables efficient storage and retrieval of embeddings, making it an excellent choice for scalable and efficient AI solutions. With proper setup and optimization, you can leverage Milvus for tasks such as search, recommendations, and retrieval-augmented generation (RAG) applications.