speaky

Speaky

UNDER DEVELOPMENT

Store activerecord models in vector stores and query them with LLMs!

Installation

Add the gem to your Gemfile:

gem 'speaky'

Install the gem:

bundle install

Create a new configuration initializer:

# config/initializers/speaky.rb

Speaky.configure do |config|
  # Set the LLM type to use for querying the vector store.
  # - 'openai' -> require 'ruby-openai' gem
  config.llm_type = 'openai'

  # Set the LLM configuration options.
  # - for 'openai' LLMs, view https://github.com/alexrudall/ruby-openai for configuration options
  config.llm_config = {
    access_token: 'YOUR_ACCESS_TOKEN',
    organization_id: 'YOUR_ORGANIZATION_ID',
  }

  # Set the vector store type to use for storing model instances.
  # - 'qdrant' -> require 'qdrant-ruby' gem
  config.vectorstore_type = 'qdrant'

  # Set the vector store configuration options.
  # - for 'qdrant' vector stores, :url, :api_key, :collection_name are required
  config.vectorstore_config = {
    url: 'YOUR_URL',
    api_key: 'YOUR_API_KEY',
    collection_name: 'YOUR_COLLECTION_NAME',
  }
end

Usage

To use the gem, include the Speaky::Concern module in your ActiveRecord models that should be stored in the vector store:

class MyModel < ApplicationRecord
  include Speaky::Concern

  def as_speaky
    # This method should return a string representation of the model instance data that should be stored in the vector store. The default implementation is to call `to_json` on the model instance data.
  end

  # Add any callbacks that should update the model instance data in the vector store.
  after_save :save_for_speaky
  after_destroy :destroy_for_speaky
end

To chat with the LLM, use the Speaky.chat method:

# Using default prompt template
Speaky.ask('What is the capital of France?')

# Using custom prompt template
Speaky.ask(
  'What is the capital of France?',
  template: 'You are a chatbot. Please answer the following question: {{question}} using this context: {{context}}.'
)

Development

1. Clone the repo

2. Install dependencies with bundle install

3. Create your local .env file with cp .env.example .env

4. Run the tests with bundle exec rspec

Publish a new version

1. Update the version in lib/speaky/version.rb

2. Run ruby bin/publish

License

The gem is available as open source under the terms of the MIT License.