ruby_llm 0.1.0.pre35 → 0.1.0.pre37

data/lib/ruby_llm/providers/openai/capabilities.rb

@@ -7,76 +7,108 @@ module RubyLLM
       module Capabilities # rubocop:disable Metrics/ModuleLength
         module_function
 
+        # Returns the context window size for the given model ID
+        # @param model_id [String] the model identifier
+        # @return [Integer] the context window size in tokens
         def context_window_for(model_id)
           case model_id
-          when /o[13]-mini/, /o3-mini-2025/         then 200_000
-          when /o1-2024/                            then 200_000
-          when /gpt-4o/, /gpt-4-turbo/             then 128_000
-          when /gpt-4-0[0-9]{3}/                   then 8_192
-          when /gpt-3.5-turbo-instruct/            then 4_096
-          when /gpt-3.5/                           then 16_385
+          when /o1-2024/, /o3-mini/, /o3-mini-2025/ then 200_000
+          when /gpt-4o/, /gpt-4o-mini/, /gpt-4-turbo/, /o1-mini/ then 128_000
+          when /gpt-4-0[0-9]{3}/ then 8_192
+          when /gpt-3.5/ then 16_385
+          when /babbage-002/, /davinci-002/ then 16_384
           else 4_096
           end
         end
 
-        def max_tokens_for(model_id) # rubocop:disable Metrics/CyclomaticComplexity
+        # Returns the maximum output tokens for the given model ID
+        # @param model_id [String] the model identifier
+        # @return [Integer] the maximum output tokens
+        def max_tokens_for(model_id)
           case model_id
-          when /o1-2024/, /o3-mini/                then 100_000
-          when /o1-mini-2024/                      then 65_536
-          when /gpt-4o-2024-05-13/                then 4_096
-          when /gpt-4o/, /gpt-4o-mini/            then 16_384
-          when /gpt-4o-realtime/                  then 4_096
-          when /gpt-4-0[0-9]{3}/                  then 8_192
-          when /gpt-3.5-turbo/                    then 4_096
+          when /o1-2024/, /o3-mini/, /o3-mini-2025/ then 100_000
+          when /o1-mini-2024/ then 65_536
+          when /gpt-4o/, /gpt-4o-mini/, /gpt-4o-audio/, /gpt-4o-mini-audio/, /babbage-002/, /davinci-002/ then 16_384
+          when /gpt-4-0[0-9]{3}/ then 8_192
           else 4_096
           end
         end
 
+        # Returns the input price per million tokens for the given model ID
+        # @param model_id [String] the model identifier
+        # @return [Float] the price per million tokens for input
         def input_price_for(model_id)
           PRICES.dig(model_family(model_id), :input) || default_input_price
         end
 
+        # Returns the output price per million tokens for the given model ID
+        # @param model_id [String] the model identifier
+        # @return [Float] the price per million tokens for output
         def output_price_for(model_id)
           PRICES.dig(model_family(model_id), :output) || default_output_price
         end
 
+        # Determines if the model supports vision capabilities
+        # @param model_id [String] the model identifier
+        # @return [Boolean] true if the model supports vision
         def supports_vision?(model_id)
           model_id.match?(/gpt-4o|o1/) || model_id.match?(/gpt-4-(?!0314|0613)/)
         end
 
+        # Determines if the model supports function calling
+        # @param model_id [String] the model identifier
+        # @return [Boolean] true if the model supports functions
         def supports_functions?(model_id)
           !model_id.include?('instruct')
         end
 
+        # Determines if the model supports audio input/output
+        # @param model_id [String] the model identifier
+        # @return [Boolean] true if the model supports audio
         def supports_audio?(model_id)
           model_id.match?(/audio-preview|realtime-preview|whisper|tts/)
         end
 
+        # Determines if the model supports JSON mode
+        # @param model_id [String] the model identifier
+        # @return [Boolean] true if the model supports JSON mode
         def supports_json_mode?(model_id)
           model_id.match?(/gpt-4-\d{4}-preview/) ||
             model_id.include?('turbo') ||
             model_id.match?(/gpt-3.5-turbo-(?!0301|0613)/)
         end
 
+        # Formats the model ID into a human-readable display name
+        # @param model_id [String] the model identifier
+        # @return [String] the formatted display name
         def format_display_name(model_id)
           model_id.then { |id| humanize(id) }
                   .then { |name| apply_special_formatting(name) }
         end
 
+        # Determines the type of model
+        # @param model_id [String] the model identifier
+        # @return [String] the model type (chat, embedding, image, audio, moderation)
         def model_type(model_id)
           case model_id
           when /text-embedding|embedding/ then 'embedding'
           when /dall-e/ then 'image'
           when /tts|whisper/ then 'audio'
-          when /omni-moderation/ then 'moderation'
+          when /omni-moderation|text-moderation/ then 'moderation'
           else 'chat'
           end
         end
 
+        # Determines if the model supports structured output
+        # @param model_id [String] the model identifier
+        # @return [Boolean] true if the model supports structured output
         def supports_structured_output?(model_id)
-          model_id.match?(/gpt-4o|o[13]-mini|o1/)
+          model_id.match?(/gpt-4o|o[13]-mini|o1|o3-mini/)
         end
 
+        # Determines the model family for pricing and capability lookup
+        # @param model_id [String] the model identifier
+        # @return [Symbol] the model family identifier
         def model_family(model_id) # rubocop:disable Metrics/AbcSize,Metrics/CyclomaticComplexity,Metrics/MethodLength
           case model_id
           when /o3-mini/ then 'o3_mini'
@@ -100,13 +132,14 @@ module RubyLLM
           when /tts-1-hd/ then 'tts1_hd'
           when /tts-1/ then 'tts1'
           when /whisper/ then 'whisper1'
-          when /omni-moderation/ then 'moderation'
+          when /omni-moderation|text-moderation/ then 'moderation'
           when /babbage/ then 'babbage'
           when /davinci/ then 'davinci'
           else 'other'
           end
         end
 
+        # Pricing information for OpenAI models (per million tokens unless otherwise specified)
         PRICES = {
           o1: { input: 15.0, cached_input: 7.5, output: 60.0 },
           o1_mini: { input: 1.10, cached_input: 0.55, output: 4.40 },
@@ -150,38 +183,53 @@ module RubyLLM
           embedding2: { price: 0.10 },
           davinci: { input: 2.0, output: 2.0 },
           babbage: { input: 0.40, output: 0.40 },
-          tts1: { price: 15.0 },
-          tts1_hd: { price: 30.0 },
-          whisper1: { price: 0.006 }
+          tts1: { price: 15.0 }, # per million characters
+          tts1_hd: { price: 30.0 }, # per million characters
+          whisper1: { price: 0.006 }, # per minute
+          moderation: { price: 0.0 } # free
         }.freeze
 
+        # Default input price when model-specific pricing is not available
+        # @return [Float] the default price per million tokens
         def default_input_price
           0.50
         end
 
+        # Default output price when model-specific pricing is not available
+        # @return [Float] the default price per million tokens
         def default_output_price
           1.50
         end
 
+        # Converts a model ID to a human-readable format
+        # @param id [String] the model identifier
+        # @return [String] the humanized model name
         def humanize(id)
           id.tr('-', ' ')
-            .split(' ')
+            .split
             .map(&:capitalize)
             .join(' ')
         end
 
+        # Applies special formatting rules to model names
+        # @param name [String] the humanized model name
+        # @return [String] the specially formatted model name
         def apply_special_formatting(name) # rubocop:disable Metrics/MethodLength
           name
             .gsub(/(\d{4}) (\d{2}) (\d{2})/, '\1\2\3')
             .gsub(/^Gpt /, 'GPT-')
             .gsub(/^O([13]) /, 'O\1-')
+            .gsub(/^O3 Mini/, 'O3-Mini')
+            .gsub(/^O1 Mini/, 'O1-Mini')
             .gsub(/^Chatgpt /, 'ChatGPT-')
             .gsub(/^Tts /, 'TTS-')
             .gsub(/^Dall E /, 'DALL-E-')
-            .gsub(/3\.5 /, '3.5-')
-            .gsub(/4 /, '4-')
-            .gsub(/4o (?=Mini|Preview|Turbo|Audio)/, '4o-')
+            .gsub('3.5 ', '3.5-')
+            .gsub('4 ', '4-')
+            .gsub(/4o (?=Mini|Preview|Turbo|Audio|Realtime)/, '4o-')
             .gsub(/\bHd\b/, 'HD')
+            .gsub('Omni Moderation', 'Omni-Moderation')
+            .gsub('Text Moderation', 'Text-Moderation')
         end
       end
     end

data/lib/ruby_llm/providers/openai/embeddings.rb

@@ -25,7 +25,7 @@ module RubyLLM
           vectors = data['data'].map { |d| d['embedding'] }
 
           # If we only got one embedding, return it as a single vector
-          vectors = vectors.size == 1 ? vectors.first : vectors
+          vectors = vectors.first if vectors.size == 1
 
           Embedding.new(
             vectors: vectors,

data/lib/ruby_llm/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module RubyLLM
-  VERSION = '0.1.0.pre35'
+  VERSION = '0.1.0.pre37'
 end

data/lib/tasks/models.rake

@@ -16,6 +16,9 @@ PROVIDER_DOCS = {
   },
   deepseek: {
     models: 'https://api-docs.deepseek.com/quick_start/pricing/'
+  },
+  anthropic: {
+    models: 'https://docs.anthropic.com/en/docs/about-claude/models/all-models'
   }
 }.freeze
 
@@ -67,8 +70,8 @@ namespace :models do # rubocop:disable Metrics/BlockLength
     RubyLLM.configure do |config|
       config.openai_api_key = ENV.fetch('OPENAI_API_KEY')
       config.anthropic_api_key = ENV.fetch('ANTHROPIC_API_KEY')
-      config.gemini_api_key = ENV['GEMINI_API_KEY']
-      config.deepseek_api_key = ENV['DEEPSEEK_API_KEY']
+      config.gemini_api_key = ENV.fetch('GEMINI_API_KEY', nil)
+      config.deepseek_api_key = ENV.fetch('DEEPSEEK_API_KEY', nil)
     end
 
     # Get all models
@@ -85,8 +88,10 @@ namespace :models do # rubocop:disable Metrics/BlockLength
     end
   end
 
-  desc 'Update model capabilities modules by scraping provider documentation'
+  desc 'Update model capabilities modules by scraping provider documentation (use PROVIDER=name to update only one)'
   task :update_capabilities do # rubocop:disable Metrics/BlockLength
+    # Check if a specific provider was requested
+    target_provider = ENV['PROVIDER']&.to_sym
     require 'ruby_llm'
     require 'fileutils'
 
@@ -97,8 +102,15 @@ namespace :models do # rubocop:disable Metrics/BlockLength
       config.gemini_api_key = ENV.fetch('GEMINI_API_KEY')
     end
 
+    # Filter providers if a specific one was requested
+    providers_to_process = if target_provider && PROVIDER_DOCS.key?(target_provider)
+                             { target_provider => PROVIDER_DOCS[target_provider] }
+                           else
+                             PROVIDER_DOCS
+                           end
+
     # Process each provider
-    PROVIDER_DOCS.each do |provider, urls| # rubocop:disable Metrics/BlockLength
+    providers_to_process.each do |provider, urls| # rubocop:disable Metrics/BlockLength
       puts "Processing #{provider}..."
 
       # Initialize our AI assistants
@@ -175,12 +187,22 @@ namespace :models do # rubocop:disable Metrics/BlockLength
 
         response = claude.ask(code_prompt)
 
+        # Extract Ruby code from Claude's response
+        puts "  Extracting Ruby code from Claude's response..."
+        ruby_code = nil
+
+        # Look for Ruby code block
+        ruby_code = Regexp.last_match(1).strip if response.content =~ /```ruby\s*(.*?)```/m
+
+        # Verify we found Ruby code
+        raise "No Ruby code block found in Claude's response" if ruby_code.nil? || ruby_code.empty?
+
         # Save the file
         file_path = "lib/ruby_llm/providers/#{provider}/capabilities.rb"
         puts "  Writing #{file_path}..."
 
         FileUtils.mkdir_p(File.dirname(file_path))
-        File.write(file_path, response.content)
+        File.write(file_path, ruby_code)
       rescue StandardError => e
         raise "Failed to process #{provider}: #{e.message}"
       end

data/ruby_llm.gemspec

@@ -9,19 +9,21 @@ Gem::Specification.new do |spec|
   spec.email         = ['carmine@paolino.me']
 
   spec.summary       = 'Beautiful Ruby interface to modern AI'
-  spec.description   = 'A delightful Ruby way to work with AI. Chat in text, analyze and generate images, understand' \
-                       ' audio, and use tools through a unified interface to OpenAI, Anthropic, Google, and DeepSeek.' \
-                       ' Built for developer happiness with automatic token counting, proper streaming, and Rails' \
-                       ' integration. No wrapping your head around multiple APIs - just clean Ruby code that works.'
-  spec.homepage      = 'https://github.com/crmne/ruby_llm'
+  spec.description   = 'A delightful Ruby way to work with AI. Chat in text, analyze and generate images, understand ' \
+                       'audio, and use tools through a unified interface to OpenAI, Anthropic, Google, and DeepSeek. ' \
+                       'Built for developer happiness with automatic token counting, proper streaming, and Rails ' \
+                       'integration. No wrapping your head around multiple APIs - just clean Ruby code that works.'
+  spec.homepage      = 'https://rubyllm.com'
   spec.license       = 'MIT'
   spec.required_ruby_version = Gem::Requirement.new('>= 3.1.0')
 
   spec.metadata['homepage_uri'] = spec.homepage
-  spec.metadata['source_code_uri'] = spec.homepage
-  spec.metadata['changelog_uri'] = "#{spec.homepage}/commits/main"
+  spec.metadata['source_code_uri'] = 'https://github.com/crmne/ruby_llm'
+  spec.metadata['changelog_uri'] = "#{spec.metadata['source_code_uri']}/commits/main"
   spec.metadata['documentation_uri'] = spec.homepage
-  spec.metadata['bug_tracker_uri'] = "#{spec.homepage}/issues"
+  spec.metadata['bug_tracker_uri'] = "#{spec.metadata['source_code_uri']}/issues"
+
+  spec.metadata['rubygems_mfa_required'] = 'true'
 
   # Specify which files should be added to the gem when it is released.
   # The `git ls-files -z` loads the files in the RubyGem that have been added into git.
@@ -38,28 +40,4 @@ Gem::Specification.new do |spec|
   spec.add_dependency 'faraday-multipart', '>= 1.0'
   spec.add_dependency 'faraday-retry', '>= 2.0'
   spec.add_dependency 'zeitwerk', '>= 2.6'
-
-  # Rails integration dependencies
-  spec.add_development_dependency 'activerecord', '>= 6.0', '< 9.0'
-  spec.add_development_dependency 'activesupport', '>= 6.0', '< 9.0'
-
-  # Development dependencies
-  spec.add_development_dependency 'bundler', '>= 2.0'
-  spec.add_development_dependency 'codecov'
-  spec.add_development_dependency 'dotenv'
-  spec.add_development_dependency 'irb'
-  spec.add_development_dependency 'nokogiri'
-  spec.add_development_dependency 'overcommit', '>= 0.66'
-  spec.add_development_dependency 'pry', '>= 0.14'
-  spec.add_development_dependency 'rake', '>= 13.0'
-  spec.add_development_dependency 'rdoc'
-  spec.add_development_dependency 'reline'
-  spec.add_development_dependency 'rspec', '~> 3.12'
-  spec.add_development_dependency 'rubocop', '>= 1.0'
-  spec.add_development_dependency 'rubocop-rake', '>= 0.6'
-  spec.add_development_dependency 'simplecov', '>= 0.21'
-  spec.add_development_dependency 'simplecov-cobertura'
-  spec.add_development_dependency 'sqlite3'
-  spec.add_development_dependency 'webmock', '~> 3.18'
-  spec.add_development_dependency 'yard', '>= 0.9'
 end