Milvus

Ruby wrapper for the Milvus vector search database API.

Part of the Langchain.rb stack.

Installation

Install the gem and add to the application's Gemfile by executing:

$ bundle add milvus

If bundler is not being used to manage dependencies, install the gem by executing:

$ gem install milvus

Usage

Instantiating API client

require 'milvus'

client = Milvus::Client.new(
    url: 'http://localhost:9091'
)

Using the Collections endpoints

# Data types: "boolean", "int8", "int16", "int32", "int64", "float", "double", "string", "varchar", "binary_vector", "float_vector"

# Creating a new collection schema
client.collections.create(
  collection_name: "book",
  description: "Test book search",
  auto_id: false,
  fields: [
    {
      "name": "book_id",
      "description": "book id",
      "is_primary_key": true,
      "autoID": false,
      "data_type": Milvus::DATA_TYPES["int64"]
    },
    {
      "name": "word_count",
      "description": "count of words",
      "is_primary_key": false,
      "data_type": Milvus::DATA_TYPES["int64"]
    },
    {
      "name": "book_intro",
      "description": "embedded vector of book introduction",
      "data_type": Milvus::DATA_TYPES["binary_vector"],
      "is_primary_key": false,
      "type_params": [
        {
          "key": "dim",
          "value": "2"
        }
      ]
    }
  ]
)

# Get the collection info
client.collections.get(collection_name: "book")

# Delete the collection
client.collections.delete(collection_name: "book")

# Load the collection to memory before a search or a query
client.collections.load(collection_name: "book")

# Release a collection from memory after a search or a query to reduce memory usage
client.collections.release(collection_name: "book")

Inserting Data

client.entities.insert(
  collection_name: "book",
  num_rows: 5, # Number of rows to be inserted. The number should be the same as the length of each field array.
  fields_data: [
    {
      "field_name": "book_id",
      "type": Milvus::DATA_TYPES["int64"],
      "field": [1,2,3,4,5]
    },
    {
      "field_name": "word_count",
      "type": Milvus::DATA_TYPES["int64"],
      "field": [1000,2000,3000,4000,5000]
    },
    {
      "field_name": "book_intro",
      "type": 101,
      "field": [ [1,1],[2,1],[3,1],[4,1],[5,1] ]
    }
  ]  
)

# Delete the entities with the boolean expression you created
client.entities.delete(
  collection_name: "book",
  expression: "book_id in [0,1]"
)

# Compact data manually
client.entities.compact!(
  collection_id: "book"
)
# => {"status"=>{}, "compactionID"=>440928616022809499}

# Check compaction status
client.entities.compact_status(
  compaction_id: 440928616022809499
)
# => {"status"=>{}, "state"=>2}

Indices

client.indices.create(
  collection_name: "book",
  field_name: "book_intro",
  extra_params: [
    { key: "metric_type", "value": "L2" },
    { key: "index_type", "value": "IVF_FLAT" },
    { key: "params", "value": "{\"nlist\":1024}" }
  ]
)

collection.indices.create(
  field_name: "book_name", 
  index_name: "scalar_index",
)

client.indices.delete(
  collection_name: "book",
  field_name: "book_intro"
)

Search & Querying

client.search(
  collection_name: "book",
  output_fields: ["book_id"], # optional
  anns_field: "book_intro",
  top_k: "2",
  params: "{\"nprobe\": 10}",
  metric_type: "L2",
  round_decimal: "-1",
  vectors: [ [0.1,0.2] ],
  dsl_type: 1
)

client.query(
  collection_name: "book",
  output_fields: ["book_id", "book_intro"],
  expr: "book_id in [2,4,6,8]"
)

Partitions

client.partitions.create(
  "collection_name": "book",
  "partition_name": "novel"
)

client.partitions.get(
  "collection_name": "book",
  "partition_name": "novel"
)

client.partitions.delete(
  "collection_name": "book",
  "partition_name": "novel"
)

client.partitions.load(
  "collection_name": "book",
  "partition_names": ["novel"],
  "replica_number": 1
)

client.partitions.release(
  "collection_name": "book",
  "partition_names": ["novel"],
  "replica_number": 1
)

Health

# Live determines whether the application is alive. It can be used for Kubernetes liveness probe.
client.health

Development

After checking out the repo, run bin/setup to install dependencies. Then, run rake spec to run the tests. You can also run bin/console for an interactive prompt that will allow you to experiment.

To install this gem onto your local machine, run bundle exec rake install. To release a new version, update the version number in version.rb, and then run bundle exec rake release, which will create a git tag for the version, push git commits and the created tag, and push the .gem file to rubygems.org.

Contributing

Bug reports and pull requests are welcome on GitHub at https://github.com/andreibondarev/milvus.

License

milvus is licensed under the Apache License, Version 2.0. View a copy of the License file.