Module: GemHadar::OllamaSupport

Included in:
GemHadar
Defined in:
lib/gem_hadar/ollama_support.rb

Overview

A module that provides Ollama AI integration support for GemHadar.

This module includes methods for initializing an Ollama client, retrieving the configured AI model name, and fetching model options for AI generation. It also provides a method to generate responses from Ollama AI models using the configured settings.

Instance Method Summary collapse

Instance Method Details

#ollama ⇒ Ollama::Client

The ollama method initializes and returns an Ollama client instance.

This method constructs an Ollama client by first determining the base URL from the OLLAMA_URL environment variable, falling back to the OLLAMA_HOST environment variable with a default value of ‘localhost:11434’. It then creates a client configuration hash including read and connect timeouts before instantiating and returning a new Ollama::Client object.



19
20
21
22
23
24
25
26
27
 
# File 'lib/gem_hadar/ollama_support.rb', line 19

def ollama
  base_url      = ENV['OLLAMA_URL'] || "http://%s" % ENV.fetch('OLLAMA_HOST', 'localhost:11434')
  client_config = {
    base_url:        base_url,
    read_timeout:    600,
    connect_timeout: 60
  }
  Ollama::Client.new(**client_config)
end

#ollama_generate(system:, prompt:) ⇒ String?

Generates a response from an AI model using the Ollama::Client.



67
68
69
70
71
 
# File 'lib/gem_hadar/ollama_support.rb', line 67

def ollama_generate(system:, prompt:)
  ollama.generate(
    model:, system:, prompt:, options:, stream: false, think: false
  ).response
end

#ollama_model ⇒ String Also known as: model

The ollama_model method retrieves the name of the Ollama AI model to be used for generating responses.

It first checks the OLLAMA_MODEL environment variable for a custom model specification. If the environment variable is not set, it falls back to using the default model name, which is determined by the ollama_model_default dsl method.



39
40
41
 
# File 'lib/gem_hadar/ollama_support.rb', line 39

def ollama_model
  ENV.fetch('OLLAMA_MODEL', ollama_model_default)
end

#options ⇒ Hash

The options method retrieves and configures the Ollama model options for AI generation.

This method fetches the JSON configuration for the Ollama model from the OLLAMA_MODEL_OPTIONS environment variable, or uses an empty JSON object if the variable is not set. It then merges default values for temperature, top_p, and min_p parameters to ensure consistent AI response characteristics.



56
57
58
59
 
# File 'lib/gem_hadar/ollama_support.rb', line 56

def options
  options = JSON(ENV.fetch('OLLAMA_MODEL_OPTIONS', '{}'))
  options.merge!("temperature" => 0, "top_p" => 1, "min_p" => 0.1)
end