ruby_llm 1.3.0rc1 → 1.3.0

data/lib/tasks/aliases.rake

@@ -0,0 +1,235 @@
+# frozen_string_literal: true
+
+require 'json'
+
+namespace :aliases do # rubocop:disable Metrics/BlockLength
+  desc 'Generate aliases.json from models in the registry'
+  task :generate do # rubocop:disable Metrics/BlockLength
+    require 'ruby_llm'
+
+    # Group models by provider
+    models = Hash.new { |h, k| h[k] = [] }
+
+    RubyLLM.models.all.each do |model|
+      models[model.provider] << model.id
+    end
+
+    aliases = {}
+
+    # OpenAI models
+    models['openai'].each do |model|
+      openrouter_model = "openai/#{model}"
+      next unless models['openrouter'].include?(openrouter_model)
+
+      alias_key = model.gsub('-latest', '')
+      aliases[alias_key] = {
+        'openai' => model,
+        'openrouter' => openrouter_model
+      }
+    end
+
+    # Anthropic models - group by base name and find latest
+    anthropic_latest = group_anthropic_models_by_base_name(models['anthropic'])
+
+    anthropic_latest.each do |base_name, latest_model|
+      # Check OpenRouter naming patterns for the BASE NAME (not the full dated model)
+      openrouter_variants = [
+        "anthropic/#{base_name}", # anthropic/claude-3-5-sonnet
+        "anthropic/#{base_name.gsub(/-(\d)/, '.\1')}", # anthropic/claude-3.5-sonnet
+        "anthropic/#{base_name.gsub(/claude-(\d+)-(\d+)/, 'claude-\1.\2')}", # claude-3-5 -> claude-3.5
+        "anthropic/#{base_name.gsub(/(\d+)-(\d+)/, '\1.\2')}" # any X-Y -> X.Y pattern
+      ]
+
+      openrouter_model = openrouter_variants.find { |v| models['openrouter'].include?(v) }
+
+      # Find corresponding Bedrock model
+      bedrock_model = find_best_bedrock_model(latest_model, models['bedrock'])
+
+      # Create alias if we have any match (OpenRouter OR Bedrock) OR if it's Anthropic-only
+      next unless openrouter_model || bedrock_model || models['anthropic'].include?(latest_model)
+
+      aliases[base_name] = {
+        'anthropic' => latest_model
+      }
+
+      aliases[base_name]['openrouter'] = openrouter_model if openrouter_model
+      aliases[base_name]['bedrock'] = bedrock_model if bedrock_model
+    end
+
+    # Also check if Bedrock has models that Anthropic doesn't
+    models['bedrock'].each do |bedrock_model|
+      next unless bedrock_model.start_with?('anthropic.')
+
+      # Extract the Claude model name
+      next unless bedrock_model =~ /anthropic\.(claude-[\d\.]+-[a-z]+)/
+
+      base_name = Regexp.last_match(1)
+      # Normalize to Anthropic naming convention
+      anthropic_name = base_name.gsub('.', '-')
+
+      # Skip if we already have an alias for this
+      next if aliases[anthropic_name]
+
+      # Check if this model exists in OpenRouter
+      openrouter_variants = [
+        "anthropic/#{anthropic_name}",
+        "anthropic/#{base_name}" # Keep the dots
+      ]
+
+      openrouter_model = openrouter_variants.find { |v| models['openrouter'].include?(v) }
+
+      aliases[anthropic_name] = {
+        'bedrock' => bedrock_model
+      }
+
+      aliases[anthropic_name]['anthropic'] = anthropic_name if models['anthropic'].include?(anthropic_name)
+      aliases[anthropic_name]['openrouter'] = openrouter_model if openrouter_model
+    end
+
+    # Gemini models
+    models['gemini'].each do |model|
+      # OpenRouter uses "google/" prefix and sometimes different naming
+      openrouter_variants = [
+        "google/#{model}",
+        "google/#{model.gsub('gemini-', 'gemini-').gsub('.', '-')}",
+        "google/#{model.gsub('gemini-', 'gemini-')}"
+      ]
+
+      openrouter_model = openrouter_variants.find { |v| models['openrouter'].include?(v) }
+      next unless openrouter_model
+
+      alias_key = model.gsub('-latest', '')
+      aliases[alias_key] = {
+        'gemini' => model,
+        'openrouter' => openrouter_model
+      }
+    end
+
+    # DeepSeek models
+    models['deepseek'].each do |model|
+      openrouter_model = "deepseek/#{model}"
+      next unless models['openrouter'].include?(openrouter_model)
+
+      alias_key = model.gsub('-latest', '')
+      aliases[alias_key] = {
+        'deepseek' => model,
+        'openrouter' => openrouter_model
+      }
+    end
+
+    # Write the result
+    sorted_aliases = aliases.sort.to_h
+    File.write('lib/ruby_llm/aliases.json', JSON.pretty_generate(sorted_aliases))
+
+    puts "Generated #{sorted_aliases.size} aliases"
+  end
+
+  def group_anthropic_models_by_base_name(anthropic_models) # rubocop:disable Rake/MethodDefinitionInTask
+    grouped = Hash.new { |h, k| h[k] = [] }
+
+    anthropic_models.each do |model|
+      base_name = extract_base_name(model)
+      grouped[base_name] << model
+    end
+
+    # Find the latest model for each base name
+    latest_models = {}
+    grouped.each do |base_name, model_list|
+      if model_list.size == 1
+        latest_models[base_name] = model_list.first
+      else
+        # Sort by date and take the latest
+        latest_model = model_list.max_by { |model| extract_date_from_model(model) }
+        latest_models[base_name] = latest_model
+      end
+    end
+
+    latest_models
+  end
+
+  def extract_base_name(model) # rubocop:disable Rake/MethodDefinitionInTask
+    # Remove date suffix (YYYYMMDD) from model name
+    if model =~ /^(.+)-(\d{8})$/
+      Regexp.last_match(1)
+    else
+      # Models without date suffix (like claude-2.0, claude-2.1)
+      model
+    end
+  end
+
+  def extract_date_from_model(model) # rubocop:disable Rake/MethodDefinitionInTask
+    # Extract date for comparison, return '00000000' for models without dates
+    if model =~ /-(\d{8})$/
+      Regexp.last_match(1)
+    else
+      '00000000' # Ensures models without dates sort before dated ones
+    end
+  end
+
+  def find_best_bedrock_model(anthropic_model, bedrock_models) # rubocop:disable Metrics/PerceivedComplexity,Rake/MethodDefinitionInTask
+    # Special mapping for Claude 2.x models
+    base_pattern = case anthropic_model
+                   when 'claude-2.0', 'claude-2'
+                     'claude-v2'
+                   when 'claude-2.1'
+                     'claude-v2:1'
+                   when 'claude-instant-v1', 'claude-instant'
+                     'claude-instant'
+                   else
+                     # For Claude 3+ models, extract base name
+                     extract_base_name(anthropic_model)
+                   end
+
+    # Find all matching Bedrock models by stripping provider prefix and comparing base name
+    matching_models = bedrock_models.select do |bedrock_model|
+      # Strip any provider prefix (anthropic. or us.anthropic.)
+      model_without_prefix = bedrock_model.sub(/^(?:us\.)?anthropic\./, '')
+      model_without_prefix.start_with?(base_pattern)
+    end
+
+    return nil if matching_models.empty?
+
+    # Get model info to check context window
+    begin
+      model_info = RubyLLM.models.find(anthropic_model)
+      target_context = model_info.context_window
+    rescue StandardError
+      target_context = nil
+    end
+
+    # If we have context window info, try to match it
+    if target_context
+      # Convert to k format (200000 -> 200k)
+      target_k = target_context / 1000
+
+      # Find models with this specific context window
+      with_context = matching_models.select do |m|
+        m.include?(":#{target_k}k") || m.include?(":0:#{target_k}k")
+      end
+
+      return with_context.first if with_context.any?
+    end
+
+    # Otherwise, pick the one with the highest context window or latest version
+    matching_models.min_by do |model|
+      # Extract context window if specified
+      context_priority = if model =~ /:(?:\d+:)?(\d+)k/
+                           -Regexp.last_match(1).to_i # Negative for descending sort
+                         else
+                           0 # No context specified
+                         end
+
+      # Extract version if present
+      version_priority = if model =~ /-v(\d+):/
+                           -Regexp.last_match(1).to_i # Negative for descending sort (latest version first)
+                         else
+                           0
+                         end
+
+      # Prefer models with explicit context windows
+      has_context_priority = model.include?('k') ? -1 : 0
+
+      [has_context_priority, context_priority, version_priority]
+    end
+  end
+end

data/lib/tasks/release.rake

@@ -0,0 +1,32 @@
+# frozen_string_literal: true
+
+namespace :release do
+  desc 'Verify cassettes are fresh enough for release'
+  task :verify_cassettes do
+    max_age_days = 1
+    cassette_dir = 'spec/fixtures/vcr_cassettes'
+    stale_cassettes = []
+
+    Dir.glob("#{cassette_dir}/**/*.yml").each do |cassette|
+      age_days = (Time.now - File.mtime(cassette)) / 86_400
+
+      next unless age_days > max_age_days
+
+      stale_cassettes << {
+        file: File.basename(cassette),
+        age: age_days.round(1)
+      }
+    end
+
+    if stale_cassettes.any?
+      puts "\n❌ Found stale cassettes (older than #{max_age_days} days):"
+      stale_cassettes.each do |c|
+        puts "   - #{c[:file]} (#{c[:age]} days old)"
+      end
+      puts "\nRun locally: bundle exec rspec"
+      exit 1
+    else
+      puts "✅ All cassettes are fresh (< #{max_age_days} days old)"
+    end
+  end
+end

metadata

@@ -1,14 +1,13 @@
 --- !ruby/object:Gem::Specification
 name: ruby_llm
 version: !ruby/object:Gem::Version
-  version: 1.3.0rc1
+  version: 1.3.0
 platform: ruby
 authors:
 - Carmine Paolino
-autorequire:
 bindir: bin
 cert_chain: []
-date: 2025-05-12 00:00:00.000000000 Z
+date: 1980-01-02 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: base64
@@ -42,58 +41,72 @@ dependencies:
   name: faraday
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ">="
       - !ruby/object:Gem::Version
-        version: '2'
+        version: 1.10.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ">="
       - !ruby/object:Gem::Version
-        version: '2'
+        version: 1.10.0
 - !ruby/object:Gem::Dependency
   name: faraday-multipart
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ">="
       - !ruby/object:Gem::Version
         version: '1'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ">="
       - !ruby/object:Gem::Version
         version: '1'
 - !ruby/object:Gem::Dependency
   name: faraday-net_http
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ">="
       - !ruby/object:Gem::Version
-        version: '3'
+        version: '1'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ">="
       - !ruby/object:Gem::Version
-        version: '3'
+        version: '1'
 - !ruby/object:Gem::Dependency
   name: faraday-retry
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '1'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '1'
+- !ruby/object:Gem::Dependency
+  name: marcel
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '2'
+        version: '1.0'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '2'
+        version: '1.0'
 - !ruby/object:Gem::Dependency
   name: zeitwerk
   requirement: !ruby/object:Gem::Requirement
@@ -126,10 +139,7 @@ files:
 - lib/ruby_llm/active_record/acts_as.rb
 - lib/ruby_llm/aliases.json
 - lib/ruby_llm/aliases.rb
-- lib/ruby_llm/attachments.rb
-- lib/ruby_llm/attachments/audio.rb
-- lib/ruby_llm/attachments/image.rb
-- lib/ruby_llm/attachments/pdf.rb
+- lib/ruby_llm/attachment.rb
 - lib/ruby_llm/chat.rb
 - lib/ruby_llm/chunk.rb
 - lib/ruby_llm/configuration.rb
@@ -140,8 +150,13 @@ files:
 - lib/ruby_llm/error.rb
 - lib/ruby_llm/image.rb
 - lib/ruby_llm/message.rb
-- lib/ruby_llm/mime_types.rb
-- lib/ruby_llm/model_info.rb
+- lib/ruby_llm/mime_type.rb
+- lib/ruby_llm/model.rb
+- lib/ruby_llm/model/info.rb
+- lib/ruby_llm/model/modalities.rb
+- lib/ruby_llm/model/pricing.rb
+- lib/ruby_llm/model/pricing_category.rb
+- lib/ruby_llm/model/pricing_tier.rb
 - lib/ruby_llm/models.json
 - lib/ruby_llm/models.rb
 - lib/ruby_llm/provider.rb
@@ -198,8 +213,10 @@ files:
 - lib/ruby_llm/tool_call.rb
 - lib/ruby_llm/utils.rb
 - lib/ruby_llm/version.rb
-- lib/tasks/models.rake
+- lib/tasks/aliases.rake
 - lib/tasks/models_docs.rake
+- lib/tasks/models_update.rake
+- lib/tasks/release.rake
 - lib/tasks/vcr.rake
 homepage: https://rubyllm.com
 licenses:
@@ -211,7 +228,6 @@ metadata:
   documentation_uri: https://rubyllm.com
   bug_tracker_uri: https://github.com/crmne/ruby_llm/issues
   rubygems_mfa_required: 'true'
-post_install_message:
 rdoc_options: []
 require_paths:
 - lib
@@ -226,8 +242,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.5.22
-signing_key:
+rubygems_version: 3.6.7
 specification_version: 4
 summary: A single delightful Ruby way to work with AI.
 test_files: []

data/lib/ruby_llm/attachments/audio.rb

@@ -1,12 +0,0 @@
-# frozen_string_literal: true
-
-module RubyLLM
-  module Attachments
-    # Represents an audio attachment
-    class Audio < Base
-      def format
-        File.extname(@source).downcase.delete('.') || 'wav'
-      end
-    end
-  end
-end

data/lib/ruby_llm/attachments/image.rb

@@ -1,9 +0,0 @@
-# frozen_string_literal: true
-
-module RubyLLM
-  module Attachments
-    # Represents an audio attachment
-    class Image < Base
-    end
-  end
-end

data/lib/ruby_llm/attachments/pdf.rb

@@ -1,9 +0,0 @@
-# frozen_string_literal: true
-
-module RubyLLM
-  module Attachments
-    # Represents a PDF attachment
-    class PDF < Base
-    end
-  end
-end

data/lib/ruby_llm/attachments.rb

@@ -1,78 +0,0 @@
-# frozen_string_literal: true
-
-module RubyLLM
-  # A module for handling attachments and file operations in RubyLLM.
-  module Attachments
-    # Base class for attachments
-    class Base
-      attr_reader :source, :filename
-
-      def initialize(source, filename: nil)
-        @source = source
-        @filename = filename ||
-                    (@source.respond_to?(:original_filename) && @source.original_filename) ||
-                    (@source.respond_to?(:path) && File.basename(@source.path)) ||
-                    (@source.is_a?(String) && File.basename(@source.split('?').first)) || # Basic URL basename
-                    nil
-      end
-
-      def url?
-        @source.is_a?(String) && @source.match?(%r{^https?://})
-      end
-
-      def file_path?
-        @source.is_a?(String) && !url?
-      end
-
-      def io_like?
-        @source.respond_to?(:read) && !file_path?
-      end
-
-      def content
-        return @content if defined?(@content) && !@content.nil?
-
-        if url?
-          fetch_content
-        elsif file_path?
-          load_content_from_path
-        elsif io_like?
-          load_content_from_io
-        else
-          RubyLLM.logger.warn "Attachment source is neither a String nor an IO-like object: #{@source}"
-          nil
-        end
-
-        @content
-      end
-
-      def type
-        self.class.name.demodulize.downcase
-      end
-
-      def encoded
-        Base64.strict_encode64(content)
-      end
-
-      def mime_type
-        RubyLLM::MimeTypes.detect_from_path(@filename)
-      end
-
-      private
-
-      def fetch_content
-        RubyLLM.logger.debug("Fetching content from URL: #{@source}")
-        response = Faraday.get(@source)
-        @content = response.body if response.success?
-      end
-
-      def load_content_from_path
-        @content = File.read(File.expand_path(@source))
-      end
-
-      def load_content_from_io
-        @source.rewind if source.respond_to? :rewind
-        @content = @source.read
-      end
-    end
-  end
-end