Class: LLM::Gemini

Inherits:
Provider show all
Defined in:
lib/llm/providers/gemini.rb,
lib/llm/providers/gemini/audio.rb,
lib/llm/providers/gemini/files.rb,
lib/llm/providers/gemini/format.rb,
lib/llm/providers/gemini/images.rb,
lib/llm/providers/gemini/models.rb,
lib/llm/providers/gemini/error_handler.rb,
lib/llm/providers/gemini/stream_parser.rb,
lib/llm/providers/gemini/response_parser.rb

Overview

The Gemini class implements a provider for Gemini.

The Gemini provider can accept multiple inputs (text, images, audio, and video). The inputs can be provided inline via the prompt for files under 20MB or via the Gemini Files API for files that are over 20MB

Examples:

example #1

#!/usr/bin/env ruby
require "llm"

llm = LLM.gemini(ENV["KEY"])
bot = LLM::Bot.new(llm)
bot.chat LLM.File("/images/capybara.png")
bot.chat "Describe the image"
bot.messages.select(&:assistant?).each { print "[#{_1.role}]", _1.content, "\n" }

example #2

#!/usr/bin/env ruby
require "llm"

llm = LLM.gemini(ENV["KEY"])
bot = LLM::Bot.new(llm)
bot.chat ["Describe the image", LLM::File("/images/capybara.png")]
bot.messages.select(&:assistant?).each { print "[#{_1.role}]", _1.content, "\n" }

Defined Under Namespace

Classes: Audio, Files, Images, Models

Constant Summary collapse

HOST =
"generativelanguage.googleapis.com"

Instance Method Summary collapse

Methods inherited from Provider

#chat, #chat!, #inspect, #moderations, #respond, #respond!, #responses, #schema, #with

Constructor Details

#initializeGemini

Returns a new instance of Gemini.

Parameters:

  • key (String, nil)

    The secret key for authentication



47
48
49
# File 'lib/llm/providers/gemini.rb', line 47

def initialize(**)
  super(host: HOST, **)
end

Instance Method Details

#default_modelString

Returns the default model for chat completions

Returns:

  • (String)

See Also:



131
132
133
# File 'lib/llm/providers/gemini.rb', line 131

def default_model
  "gemini-1.5-flash"
end

#embed(input, model: "text-embedding-004", **params) ⇒ LLM::Response::Embedding

Provides an embedding

Parameters:

  • input (String, Array<String>)

    The input to embed

  • model (String) (defaults to: "text-embedding-004")

    The embedding model to use

  • params (Hash)

    Other embedding parameters

Returns:



58
59
60
61
62
63
64
65
# File 'lib/llm/providers/gemini.rb', line 58

def embed(input, model: "text-embedding-004", **params)
  model = model.respond_to?(:id) ? model.id : model
  path = ["/v1beta/models/#{model}", "embedContent?key=#{@key}"].join(":")
  req = Net::HTTP::Post.new(path, headers)
  req.body = JSON.dump({content: {parts: [{text: input}]}})
  res = execute(request: req)
  Response::Embedding.new(res).extend(response_parser)
end

#complete(prompt, params = {}) ⇒ LLM::Response::Completion

Provides an interface to the chat completions API

Examples:

llm = LLM.openai(ENV["KEY"])
messages = [{role: "system", content: "Your task is to answer all of my questions"}]
res = llm.complete("5 + 2 ?", messages:)
print "[#{res.choices[0].role}]", res.choices[0].content, "\n"

Parameters:

  • prompt (String)

    The input prompt to be completed

  • params (Hash) (defaults to: {})

    The parameters to maintain throughout the conversation. Any parameter the provider supports can be included and not only those listed here.

Returns:

Raises:

  • (LLM::Error::PromptError)

    When given an object a provider does not understand

See Also:



77
78
79
80
81
82
83
84
85
86
87
88
89
90
# File 'lib/llm/providers/gemini.rb', line 77

def complete(prompt, params = {})
  params = {role: :user, model: default_model}.merge!(params)
  params = [params, format_schema(params), format_tools(params)].inject({}, &:merge!).compact
  role, model, stream = [:role, :model, :stream].map { params.delete(_1) }
  action = stream ? "streamGenerateContent?key=#{@key}&alt=sse" : "generateContent?key=#{@key}"
  model.respond_to?(:id) ? model.id : model
  path = ["/v1beta/models/#{model}", action].join(":")
  req  = Net::HTTP::Post.new(path, headers)
  messages = [*(params.delete(:messages) || []), LLM::Message.new(role, prompt)]
  body = JSON.dump({contents: format(messages)}.merge!(params))
  set_body_stream(req, StringIO.new(body))
  res = execute(request: req, stream:)
  Response::Completion.new(res).extend(response_parser)
end

#audioObject

Provides an interface to Gemini’s audio API

See Also:



95
96
97
# File 'lib/llm/providers/gemini.rb', line 95

def audio
  LLM::Gemini::Audio.new(self)
end

#imagessee LLM::Gemini::Images

Provides an interface to Gemini’s image generation API

Returns:

See Also:



103
104
105
# File 'lib/llm/providers/gemini.rb', line 103

def images
  LLM::Gemini::Images.new(self)
end

#filesObject

Provides an interface to Gemini’s file management API

See Also:



110
111
112
# File 'lib/llm/providers/gemini.rb', line 110

def files
  LLM::Gemini::Files.new(self)
end

#modelsObject

Provides an interface to Gemini’s models API

See Also:



117
118
119
# File 'lib/llm/providers/gemini.rb', line 117

def models
  LLM::Gemini::Models.new(self)
end

#assistant_roleString

Returns the role of the assistant in the conversation. Usually “assistant” or “model”

Returns:

  • (String)

    Returns the role of the assistant in the conversation. Usually “assistant” or “model”



123
124
125
# File 'lib/llm/providers/gemini.rb', line 123

def assistant_role
  "model"
end