Qdrant

  +  

Ruby wrapper for the Qdrant 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 qdrant-ruby

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

$ gem install qdrant-ruby

Usage

Instantiating API client

require 'qdrant'

client = Qdrant::Client.new(
  url: ENV["QDRANT_URL"],
  api_key: ENV["QDRANT_API_KEY"]
)

Collections

# Get list name of all existing collections
client.collections.list

# Get detailed information about specified existing collection
client.collections.get(collection_name: "string")

# Create new collection with given parameters
client.collections.create(
    collection_name: "string",     # required
    vectors: {},                   # required
    shard_number: nil,
    replication_factor: nil,
    write_consistency_factor: nil,
    on_disk_payload: nil,
    hnsw_config: nil,
    wal_config: nil,
    optimizers_config: nil,
    init_from: nil,
    quantization_config: nil
)

# Update parameters of the existing collection
client.collections.update(
    collection_name: "string",    # required
    optimizers_config: nil,
    params: nil
)

# Drop collection and all associated data
client.collections.delete(collection_name: "string")

# Get list of all aliases (for a collection)
client.collections.aliases(
    collection_name: "string" # optional
)

# Update aliases of the collections
client.collections.update_aliases(
    actions: [{
        # `create_alias:`, `delete_alias` and/or `rename_alias` is required
        create_alias: {
            collection_name: "string", # required
            alias_name: "string"       # required
        }
    }]
)

# Create index for field in collection
client.collections.create_index(
    collection_name: "string", # required
    field_name: "string",      # required
    field_schema: "string",
    wait: "boolean",
    ordering: "ordering"
)

# Delete field index for collection
client.collections.delete_index(
    collection_name: "string", # required
    field_name: "string",      # required
    wait: "boolean",
    ordering: "ordering"
)

# Get cluster information for a collection
client.collections.cluster_info(
    collection_name: "test_collection" # required
)

# Update collection cluster setup
client.collections.update_cluster(
    collection_name: "string", # required
    move_shard: {              # required
        shard_id: "int",
        to_peer_id: "int",
        from_peer_id: "int"
    },
    timeout: "int"
)

# Create new snapshot for a collection
client.collections.create_snapshot(
    collection_name: "string", # required
)

# Get list of snapshots for a collection
client.collections.list_snapshots(
    collection_name: "string", # required
)

# Delete snapshot for a collection
client.collections.delete_snapshot(
    collection_name: "string", # required
    snapshot_name: "string"    # required
)

# Recover local collection data from a snapshot. This will overwrite any data, stored on this node, for the collection. If collection does not exist - it will be created.
client.collections.restore_snapshot(
    collection_name: "string", # required
    filepath: "string",        # required
    wait: "boolean",
    priority: "string"
)

# Download specified snapshot from a collection as a file
client.collections.download_snapshot(
    collection_name: "string",         # required
    snapshot_name: "string",           # required
    filepath: "/dir/filename.snapshot" #require 
)

Points

# Retrieve full information of single point by id
client.points.get(
    collection_name: "string", # required
    id: "int/string",          # required
    consistency: "int"
)

# Retrieve full information of points by ids
client.points.get_all(
    collection_name: "string", # required
    ids: "[int]",              # required
    with_payload: "boolean"
    with_vector: "boolean"
)

# Lists all data objects in reverse order of creation. The data will be returned as an array of objects.
client.points.list(
    collection_name: "string", # required
    ids: "[int/string]",       # required
    with_payload: nil,
    with_vector: nil,
    consistency: nil

)

# Get a single data object.
client.points.upsert(
    collection_name: "string", # required
    batch: {},                 # required
    wait: "boolean",
    ordering: "string"
)

# Delete points
client.points.delete(
    collection_name: "string", # required
    points: "[int/string]",    # either `points:` or `filter:` required
    filter: {},
    wait: "boolean",
    ordering: "string"
)

# Set payload values for points
client.points.set_payload(
    collection_name: "string", # required
    payload: {                 # required
        "property name" => "value" 
    },  
    points: "[int/string]",    # `points:` or `filter:` are required
    filter: {},
    wait: "boolean",
    ordering: "string"
)

# Replace full payload of points with new one
client.points.overwrite_payload(
    collection_name: "string", # required
    payload: {},               # required 
    wait: "boolean",
    ordering: "string",
    points: "[int/string]",
    filter: {}
)

# Delete specified key payload for points
client.points.clear_payload_keys(
    collection_name: "string", # required
    keys: "[string]",          # required
    points: "[int/string]",
    filter: {},
    wait: "boolean",
    ordering: "string"
)

# Delete specified key payload for points
client.points.clear_payload(
    collection_name: "string", # required
    points: "[int/string]",    # required
    wait: "boolean",
    ordering: "string"
)

# Scroll request - paginate over all points which matches given filtering condition
client.points.scroll(
    collection_name: "string", # required
    limit: "int",
    filter: {},
    offset: "string",
    with_payload: "boolean",
    with_vector: "boolean",
    consistency: "int/string"
)

# Retrieve closest points based on vector similarity and given filtering conditions
client.points.search(
    collection_name: "string", # required
    limit: "int",              # required
    vector: "[int]",           # required
    filter: {},
    params: {},
    offset: "int",
    with_payload: "boolean",
    with_vector: "boolean",
    score_threshold: "float"
)


# Retrieve by batch the closest points based on vector similarity and given filtering conditions
client.points.batch_search(
    collection_name: "string", # required
    searches: [{}],            # required  
    consistency: "int/string"
)

# Look for the points which are closer to stored positive examples and at the same time further to negative examples.
client.points.recommend(
    collection_name: "string", # required
    positive: "[int/string]",  # required; Arrray of point IDs
    limit: "int",              # required
    negative: "[int/string]",
    filter: {},
    params: {},
    offset: "int",
    with_payload: "boolean",
    with_vector: "boolean",
    score_threshold: "float"
    using: "string",
    lookup_from: {},
)

# Look for the points which are closer to stored positive examples and at the same time further to negative examples.
client.points.batch_recommend(
    collection_name: "string", # required
    searches: [{}],            # required
    consistency: "string"
)

# Count points which matches given filtering condition
client.points.count(
    collection_name: "string", # required
    filter: {},
    exact: "boolean"
)

Snapshots

# Get list of snapshots of the whole storage
client.snapshots.list(
    collection_name: "string" # optional
)

# Create new snapshot of the whole storage
client.snapshots.create(
    collection_name: "string" # required
)

# Delete snapshot of the whole storage
client.snapshots.delete(
    collection_name: "string", # required
    snapshot_name: "string"    # required
)

# Download specified snapshot of the whole storage as a file
client.snapshots.download(
    collection_name: "string", # required
    snapshot_name: "string"    # required
    filepath: "~/Downloads/backup.txt" # required
)


# Get the backup
client.backups.get(
    backend: "filesystem",
    id: "my-first-backup"
)

# Restore backup
client.backups.restore(
    backend: "filesystem",
    id: "my-first-backup"
)

# Check the backup restore status
client.backups.restore_status(
    backend: "filesystem",
    id: "my-first-backup"
)

Cluster

# Get information about the current state and composition of the cluster
client.cluster.info

# Tries to recover current peer Raft state.
client.cluster.recover

# Tries to remove peer from the cluster. Will return an error if peer has shards on it.
client.cluster.remove_peer(
    peer_id: "int",  # required
    force: "boolean"
)

Service

# Collect telemetry data including app info, system info, collections info, cluster info, configs and statistics
client.telemetry(
    anonymize: "boolean" # optional
)

# Collect metrics data including app info, collections info, cluster info and statistics
client.metrics(
    anonymize: "boolean" # optional
)

# Get lock options. If write is locked, all write operations and collection creation are forbidden
client.locks

# Set lock options. If write is locked, all write operations and collection creation are forbidden. Returns previous lock options
client.set_lock(
    write: "boolean" # required
    error_message: "string"
)

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/qdrant.

License

qdrant-ruby is licensed under the Apache License, Version 2.0. View a copy of the License file.