ruby_llm_community 0.0.6 → 1.1.0

data/lib/tasks/models_docs.rake

@@ -1,224 +0,0 @@
-# frozen_string_literal: true
-
-require 'dotenv/load'
-require 'fileutils'
-
-namespace :models do
-  desc 'Generate available models documentation'
-  task :docs do
-    FileUtils.mkdir_p('docs') # ensure output directory exists
-
-    # Generate markdown content
-    output = generate_models_markdown
-
-    # Write the output
-    File.write('docs/_reference/available-models.md', output)
-    puts 'Generated docs/_reference/available-models.md'
-  end
-end
-
-def generate_models_markdown
-  <<~MARKDOWN
-    ---
-    layout: default
-    title: Available Models
-    nav_order: 1
-    description: Browse hundreds of AI models from every major provider. Always up-to-date, automatically generated.
-    redirect_from:
-      - /guides/available-models
-    ---
-
-    # {{ page.title }}
-    {: .no_toc }
-
-    {{ page.description }}
-    {: .fs-6 .fw-300 }
-
-    ## Table of contents
-    {: .no_toc .text-delta }
-
-    1. TOC
-    {:toc}
-
-    ---
-
-    After reading this guide, you will know:
-
-    * How RubyLLM's model registry works and where data comes from
-    * How to find models by provider, capability, or purpose
-    * What information is available for each model
-    * How to use model aliases for simpler configuration
-
-    ## How Model Data Works
-
-    RubyLLM's model registry combines data from multiple sources:
-
-    - **OpenAI, Anthropic, DeepSeek, Gemini**: Data from [Parsera](https://api.parsera.org/v1/llm-specs)
-    - **OpenRouter**: Direct from OpenRouter's API
-    - **Other providers**: Defined in `capabilities.rb` files
-
-    ## Contributing Model Updates
-
-    **For major providers** (OpenAI, Anthropic, DeepSeek, Gemini): File issues with [Parsera](https://github.com/parsera-labs/api-llm-specs/issues) for public model data corrections.
-
-    **For other providers**: Edit `lib/ruby_llm/providers/<provider>/capabilities.rb` then run `rake models:update`.
-
-    See the [Contributing Guide](https://github.com/crmne/ruby_llm/blob/main/CONTRIBUTING.md) for details.
-
-    ## Last Updated
-    {: .d-inline-block }
-
-    #{Time.now.utc.strftime('%Y-%m-%d')}
-    {: .label .label-green }
-
-    ## Models by Provider
-
-    #{generate_provider_sections}
-
-    ## Models by Capability
-
-    #{generate_capability_sections}
-
-    ## Models by Modality
-
-    #{generate_modality_sections}
-  MARKDOWN
-end
-
-def generate_provider_sections
-  RubyLLM::Provider.providers.filter_map do |provider, provider_class|
-    models = RubyLLM.models.by_provider(provider)
-    next if models.none?
-
-    <<~PROVIDER
-      ### #{provider_class.name} (#{models.count})
-
-      #{models_table(models)}
-    PROVIDER
-  end.join("\n\n")
-end
-
-def generate_capability_sections
-  capabilities = {
-    'Function Calling' => RubyLLM.models.select(&:function_calling?),
-    'Structured Output' => RubyLLM.models.select(&:structured_output?),
-    'Streaming' => RubyLLM.models.select { |m| m.capabilities.include?('streaming') },
-    # 'Reasoning' => RubyLLM.models.select { |m| m.capabilities.include?('reasoning') },
-    'Batch Processing' => RubyLLM.models.select { |m| m.capabilities.include?('batch') }
-  }
-
-  capabilities.filter_map do |capability, models|
-    next if models.none?
-
-    <<~CAPABILITY
-      ### #{capability} (#{models.count})
-
-      #{models_table(models)}
-    CAPABILITY
-  end.join("\n\n")
-end
-
-def generate_modality_sections # rubocop:disable Metrics/PerceivedComplexity
-  sections = []
-
-  # Models that support vision/images
-  vision_models = RubyLLM.models.select { |m| (m.modalities.input || []).include?('image') }
-  if vision_models.any?
-    sections << <<~SECTION
-      ### Vision Models (#{vision_models.count})
-
-      Models that can process images:
-
-      #{models_table(vision_models)}
-    SECTION
-  end
-
-  # Models that support audio
-  audio_models = RubyLLM.models.select { |m| (m.modalities.input || []).include?('audio') }
-  if audio_models.any?
-    sections << <<~SECTION
-      ### Audio Input Models (#{audio_models.count})
-
-      Models that can process audio:
-
-      #{models_table(audio_models)}
-    SECTION
-  end
-
-  # Models that support PDFs
-  pdf_models = RubyLLM.models.select { |m| (m.modalities.input || []).include?('pdf') }
-  if pdf_models.any?
-    sections << <<~SECTION
-      ### PDF Models (#{pdf_models.count})
-
-      Models that can process PDF documents:
-
-      #{models_table(pdf_models)}
-    SECTION
-  end
-
-  # Models for embeddings
-  embedding_models = RubyLLM.models.select { |m| (m.modalities.output || []).include?('embeddings') }
-  if embedding_models.any?
-    sections << <<~SECTION
-      ### Embedding Models (#{embedding_models.count})
-
-      Models that generate embeddings:
-
-      #{models_table(embedding_models)}
-    SECTION
-  end
-
-  sections.join("\n\n")
-end
-
-def models_table(models)
-  return '*No models found*' if models.none?
-
-  headers = ['Model', 'Provider', 'Context', 'Max Output', 'Standard Pricing (per 1M tokens)']
-  alignment = [':--', ':--', '--:', '--:', ':--']
-
-  rows = models.sort_by { |m| [m.provider, m.name] }.map do |model|
-    # Format pricing information
-    pricing = standard_pricing_display(model)
-
-    [
-      model.id,
-      model.provider,
-      model.context_window || '-',
-      model.max_output_tokens || '-',
-      pricing
-    ]
-  end
-
-  table = []
-  table << "| #{headers.join(' | ')} |"
-  table << "| #{alignment.join(' | ')} |"
-
-  rows.each do |row|
-    table << "| #{row.join(' | ')} |"
-  end
-
-  table.join("\n")
-end
-
-def standard_pricing_display(model)
-  # Access pricing data using to_h to get the raw hash
-  pricing_data = model.pricing.to_h[:text_tokens]&.dig(:standard) || {}
-
-  if pricing_data.any?
-    parts = []
-
-    parts << "In: $#{format('%.2f', pricing_data[:input_per_million])}" if pricing_data[:input_per_million]
-
-    parts << "Out: $#{format('%.2f', pricing_data[:output_per_million])}" if pricing_data[:output_per_million]
-
-    if pricing_data[:cached_input_per_million]
-      parts << "Cache: $#{format('%.2f', pricing_data[:cached_input_per_million])}"
-    end
-
-    return parts.join(', ') if parts.any?
-  end
-
-  '-'
-end

data/lib/tasks/models_update.rake

@@ -1,108 +0,0 @@
-# frozen_string_literal: true
-
-require 'dotenv/load'
-require 'ruby_llm'
-require 'json-schema'
-
-task default: ['models:update']
-
-namespace :models do
-  desc 'Update available models from providers (API keys needed)'
-  task :update do
-    puts 'Configuring RubyLLM...'
-    configure_from_env
-
-    refresh_models
-    display_model_stats
-  end
-end
-
-def configure_from_env
-  RubyLLM.configure do |config|
-    config.openai_api_key = ENV.fetch('OPENAI_API_KEY', nil)
-    config.anthropic_api_key = ENV.fetch('ANTHROPIC_API_KEY', nil)
-    config.gemini_api_key = ENV.fetch('GEMINI_API_KEY', nil)
-    config.deepseek_api_key = ENV.fetch('DEEPSEEK_API_KEY', nil)
-    config.perplexity_api_key = ENV.fetch('PERPLEXITY_API_KEY', nil)
-    config.openrouter_api_key = ENV.fetch('OPENROUTER_API_KEY', nil)
-    config.mistral_api_key = ENV.fetch('MISTRAL_API_KEY', nil)
-    configure_bedrock(config)
-    config.request_timeout = 30
-  end
-end
-
-def configure_bedrock(config)
-  config.bedrock_api_key = ENV.fetch('AWS_ACCESS_KEY_ID', nil)
-  config.bedrock_secret_key = ENV.fetch('AWS_SECRET_ACCESS_KEY', nil)
-  config.bedrock_region = ENV.fetch('AWS_REGION', nil)
-  config.bedrock_session_token = ENV.fetch('AWS_SESSION_TOKEN', nil)
-end
-
-def refresh_models
-  initial_count = RubyLLM.models.all.size
-  puts "Refreshing models (#{initial_count} cached)..."
-
-  models = RubyLLM.models.refresh!
-
-  if models.all.empty? && initial_count.zero?
-    puts 'Error: Failed to fetch models.'
-    exit(1)
-  elsif models.all.size == initial_count && initial_count.positive?
-    puts 'Warning: Model list unchanged.'
-  else
-    puts 'Validating models...'
-    validate_models!(models)
-
-    puts "Saving models.json (#{models.all.size} models)"
-    models.save_models
-  end
-
-  @models = models
-end
-
-def validate_models!(models)
-  schema_path = File.expand_path('../ruby_llm/models_schema.json', __dir__)
-  models_data = models.all.map(&:to_h)
-
-  validation_errors = JSON::Validator.fully_validate(schema_path, models_data)
-
-  unless validation_errors.empty?
-    # Save failed models for inspection
-    failed_path = File.expand_path('../ruby_llm/models.failed.json', __dir__)
-    File.write(failed_path, JSON.pretty_generate(models_data))
-
-    puts 'ERROR: Models validation failed:'
-    puts "\nValidation errors:"
-    validation_errors.first(10).each { |error| puts "  - #{error}" }
-    puts "  ... and #{validation_errors.size - 10} more errors" if validation_errors.size > 10
-    puts "-> Failed models saved to: #{failed_path}"
-    exit(1)
-  end
-
-  puts '✓ Models validation passed'
-end
-
-def display_model_stats
-  puts "\nModel count:"
-  provider_counts = @models.all.group_by(&:provider).transform_values(&:count)
-
-  RubyLLM::Provider.providers.each do |sym, provider_class|
-    name = provider_class.name
-    count = provider_counts[sym.to_s] || 0
-    status = status(sym)
-    puts "  #{name}: #{count} models #{status}"
-  end
-
-  puts 'Refresh complete.'
-end
-
-def status(provider_sym)
-  provider_class = RubyLLM::Provider.providers[provider_sym]
-  if provider_class.local?
-    ' (LOCAL - SKIP)'
-  elsif provider_class.configured?(RubyLLM.config)
-    ' (OK)'
-  else
-    ' (NOT CONFIGURED)'
-  end
-end