aia 0.9.3rc1 → 0.9.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 9e3aa7b74ca7b64aca1acd2360043829f9ac14ae63f4c66ced12ea1a70ae34d0
-  data.tar.gz: 5802184befd82ae7ff4eecf16da0e5fc1d9d39cc61cdcc5736ad4d56dbe0b87e
+  metadata.gz: 9a4629811acd04507b4ca4f6902281dae7a3d35128af2a52b9278cf040fb85a4
+  data.tar.gz: 8d80eb38222142a366266091aae2863c2669f676fb9f362210abf12fb270c161
 SHA512:
-  metadata.gz: 31dae0e58c2d7f4b97a23f90b1b588b8735bb6a8e9774a2f40554baf64636ca8c56adc2df9fad804c7e05e91ca5d6a2fde358dd886c56219190f1ba8262fdbf0
-  data.tar.gz: d710b9e9fbe52cf014ed8a10274eb84ad4bd1ed00b2ee03a75e71767a4f45abb37d4b5ff6b258597c77b352cf1eb5c8272f1e59c3611eaba9c730447b17e2f2c
+  metadata.gz: 5997271297a683a7664aab0b74612817fdf53b4d206e9cbc4d50b3cb8a7dc798922b3b60f7f762363723d6b77ab282309f10b6fbc25d110ac5bd4fb9a9435faf
+  data.tar.gz: cea1ffbd60baa36963339972b91557dc2fced33ee17e451513de13df3b0d582f3dceda3f6fb611b6f83cbd6b9d9069480cd92ef78507e75dcfa55c223d060889

data/.version

@@ -1 +1 @@
-0.9.3rc1
+0.9.4

data/CHANGELOG.md

@@ -1,10 +1,15 @@
 # Changelog
 ## [Unreleased]
-### [0.9.3] WIP
+
+## Released
+### [0.9.4] 2025-06-03
+- using RubyLLM v1.3.0
+- setting up a docs infrastructure to behave like the ruby_llm gem's guides side
+- fixed bug in the text-to-image workflow
+- discovered that ruby_llm does not have high level support for audio modes
 - need to pay attention to the test suite
 - also need to ensure the non text2text modes are working
 
-## Released
 ### [0.9.3rc1] 2025-05-24
 - using ruby_llm v1.3.0rc1
 - added a models database refresh based on integer days interval with the --refresh option

data/README.md

@@ -107,7 +107,7 @@ The following table provides a comprehensive list of configuration options, thei
 | presence_penalty     | --presence_penalty | 0.0                   | AIA_PRESENCE_PENALTY      |
 | prompt_extname       |             | .txt                        | AIA_PROMPT_EXTNAME        |
 | prompts_dir          | -p, --prompts_dir | ~/.prompts            | AIA_PROMPTS_DIR           |
-| refresh              | --refresh   | 0 (days)                    | AIA_REFRESH               |
+| refresh              | --refresh   | 7 (days)                    | AIA_REFRESH               |
 | require_libs         | --rq --require | []                       | AIA_REQUIRE_LIBS          |
 | role                 | -r, --role  |                             | AIA_ROLE                  |
 | roles_dir            |             | ~/.prompts/roles            | AIA_ROLES_DIR             |

data/lib/aia/chat_processor_service.rb

@@ -12,7 +12,6 @@ module AIA
     end
 
 
-
     def speak(text)
       return unless AIA.speak?
 
@@ -26,11 +25,10 @@ module AIA
     end
 
 
-
-    def process_prompt(prompt, operation_type)
+    def process_prompt(prompt)
       result = nil
-      @ui_presenter.with_spinner("Processing", operation_type) do
-        result = send_to_client(prompt, operation_type)
+      @ui_presenter.with_spinner("Processing", determine_operation_type) do
+        result = send_to_client(prompt)
       end
 
       unless result.is_a? String
@@ -41,42 +39,20 @@ module AIA
     end
 
 
-
-    def send_to_client(prompt, operation_type)
+    # conversation is an Array of Hashes.  Each entry is an interchange
+    # with the LLM.
+    def send_to_client(conversation)
       maybe_change_model
 
-      case operation_type
-      when :text_to_text
-        AIA.client.chat(prompt)
-      when :text_to_image
-        AIA.client.chat(prompt)
-      when :image_to_text
-        AIA.client.chat(prompt)
-      when :text_to_audio
-        AIA.client.chat(prompt)
-      when :audio_to_text
-        if prompt.strip.end_with?('.mp3', '.wav', '.m4a', '.flac') && File.exist?(prompt.strip)
-          AIA.client.transcribe(prompt.strip)
-        else
-          AIA.client.chat(prompt) # Fall back to regular chat
-        end
-      else
-        AIA.client.chat(prompt)
-      end
+      AIA.client.chat(conversation)
     end
 
 
     def maybe_change_model
-      if AIA.client.model.is_a?(String)
-        client_model = AIA.client.model     # AiClient instance
-      else
-        client_model = AIA.client.model.id  # RubyLLM::Model instance
-      end
+      client_model = AIA.client.model.id  # RubyLLM::Model instance
 
-      # when adapter is ruby_llm must use model.id as the name
       unless AIA.config.model.downcase.include?(client_model.downcase)
-        # FIXME: assumes that the adapter is AiClient.  It might be RUbyLLM
-        AIA.client = AIClientAdapter.new
+        AIA.client = AIA.client.class.new
       end
     end
 
@@ -107,7 +83,6 @@ module AIA
     end
 
 
-
     def process_next_prompts(response, prompt_handler)
       if @directive_processor.directive?(response)
         directive_result = @directive_processor.process(response, @history_manager.history)
@@ -117,16 +92,24 @@ module AIA
     end
 
 
-    def determine_operation_type(model)
-      model = model.to_s.downcase
-      if model.include?('dall') || model.include?('image-generation')
+    def determine_operation_type
+      mode = AIA.config.client.model.modalities
+      if mode.supports?(:text_to_image)
         :text_to_image
-      elsif model.include?('vision') || model.include?('image')
+      elsif mode.supports?(:image_to_text)
         :image_to_text
-      elsif model.include?('whisper') || model.include?('audio')
+      elsif mode.supports?(:audio_to_text)
         :audio_to_text
-      elsif model.include?('speech') || model.include?('tts')
+      elsif mode.supports?(:text_to_audio)
         :text_to_audio
+      elsif mode.supports?(:audio_to_audio)
+        :audio_to_audio
+      elsif mode.supports?(:image_to_image)
+        :image_to_image
+      elsif mode.supports?(:audio_to_image)
+        :audio_to_image
+      elsif mode.supports?(:image_to_audio)
+        :image_to_audio
       else
         :text_to_text
       end

data/lib/aia/config.rb

@@ -17,17 +17,22 @@ require 'fileutils'
 module AIA
   class Config
     DEFAULT_CONFIG = OpenStruct.new({
+      adapter:      'ruby_llm', # 'ruby_llm' or ???
+      #
       aia_dir:      File.join(ENV['HOME'], '.aia'),
       config_file:  File.join(ENV['HOME'], '.aia', 'config.yml'),
       out_file:     'temp.md',
       log_file:     File.join(ENV['HOME'], '.prompts', '_prompts.log'),
-      prompts_dir:  File.join(ENV['HOME'], '.prompts'),
+      context_files: [],
       #
+      prompts_dir:  File.join(ENV['HOME'], '.prompts'),
       prompt_extname: PromptManager::Storage::FileSystemAdapter::PROMPT_EXTENSION,
       #
       roles_prefix: 'roles',
       roles_dir:    File.join(ENV['HOME'], '.prompts', 'roles'),
       role:         '',
+
+      #
       system_prompt: '',
 
       # Tools
@@ -61,23 +66,26 @@ module AIA
       top_p:                1.0,
       frequency_penalty:    0.0,
       presence_penalty:     0.0,
+
+      # Audio Parameters
+      voice:                'alloy',
+      speak_command:        'afplay', # 'afplay' for audio files on MacOS
+
+      # Image Parameters
       image_size:           '1024x1024',
       image_quality:        'standard',
       image_style:          'vivid',
 
+      # Models
       model:                'gpt-4o-mini',
       speech_model:         'tts-1',
       transcription_model:  'whisper-1',
       embedding_model:      'text-embedding-ada-002',
       image_model:          'dall-e-3',
-      refresh:              0, # days between refreshes of model info; 0 means every startup
-      last_refresh:         Date.today - 1,
-
-      voice:                'alloy',
-      adapter:              'ruby_llm', # 'ruby_llm' or ???
 
-      # Default speak command
-      speak_command: 'afplay', # 'afplay' for audio files
+      # Model Regristery
+      refresh:              7, # days between refreshes of model info; 0 means every startup
+      last_refresh:         Date.today - 1,
 
       # Ruby libraries to require for Ruby binding
       require_libs: [],
@@ -450,6 +458,7 @@ module AIA
             config.system_prompt = prompt_id
           end
 
+          ###################################################
           # AI model parameters
           opts.on("-t", "--temperature TEMP", Float, "Temperature for text generation") do |temp|
             config.temperature = temp
@@ -629,6 +638,9 @@ module AIA
 
       config_hash = config.to_h
 
+      # Remove prompt_id to prevent automatic initial pompting in --chat mode
+      config_hash.delete(:prompt_id)
+
       # Remove dump_file key to prevent automatic exit on next load
       config_hash.delete(:dump_file)
 

data/lib/aia/context_manager.rb

@@ -27,7 +27,12 @@ module AIA
     # @return [Array<Hash>] The conversation context array.
     def get_context(system_prompt: nil)
       # Ensure system prompt is present if provided and not already the first message
-      if system_prompt && !system_prompt.strip.empty? && (@context.empty? || @context.first[:role] != 'system')
+      if  system_prompt &&
+          !system_prompt.strip.empty? &&
+          (
+            @context.empty? ||
+            @context.first[:role] != 'system'
+          )
          add_system_prompt(system_prompt)
       end
       @context

data/lib/aia/directive_processor.rb

@@ -4,6 +4,8 @@ require 'faraday'
 
 module AIA
   class DirectiveProcessor
+    using Refinements
+
     PUREMD_API_KEY   = ENV.fetch('PUREMD_API_KEY', nil)
     EXCLUDED_METHODS = %w[ run initialize private? ]
     @descriptions = {}
@@ -117,7 +119,6 @@ module AIA
       !respond_to?(method_name) && respond_to?(method_name, true)
     end
 
-
     ################
     ## Directives ##
     ################
@@ -295,6 +296,45 @@ module AIA
       ""
     end
 
+    desc "All Available models or query on [partial LLM or provider name] Examples: //llms ; //llms openai ; //llms claude"
+    def available_models( args=nil, context_manager=nil)
+      query     = args
+      header    = "Available LLMs"
+
+      if query
+        header += " for #{query.join(' and ')}"
+      end
+
+      puts header + ':'
+
+      q1 = query.select{|q| !q.start_with?(':')}
+      q2 = query.select{|q|  q.start_with?(':')}
+
+      RubyLLM.models.all.each do |llm|
+        inputs  = llm.modalities.input.join(',')
+        outputs = llm.modalities.output.join(',')
+        entry   = "- #{llm.id} (#{llm.provider}) #{inputs} to #{outputs}"
+
+        if query.nil? || query.empty?
+          puts entry
+          next
+        end
+
+        show_it = true
+        q1.each{|q| show_it &&= entry.include?(q)}
+        q2.each{|q| show_it &&= llm.modalities.supports?(q)}
+
+        puts entry if show_it
+      end
+
+      ""
+    end
+    alias_method :am,         :available_models
+    alias_method :available,  :available_models
+    alias_method :models,     :available_models
+    alias_method :all_models, :available_models
+    alias_method :llms,       :available_models
+
     desc "Generates this help content"
     def help(args=nil, context_manager=nil)
       puts

data/lib/aia/ruby_llm_adapter.rb

@@ -1,25 +1,5 @@
 # lib/aia/ruby_llm_adapter.rb
 
-require 'ruby_llm'
-
-class RubyLLM::Modalities
-  def supports?(query_mode)
-    parts = query_mode
-              .to_s
-              .downcase
-              .split(/2|-to-| to |_to_/)
-              .map(&:strip)
-
-    if 2 == parts.size
-      input.include?(parts[0]) && output.include?(parts[1])
-    elsif 1 == parts.size
-      input.include?(parts[0]) || output.include?(parts[0])
-    else
-      false
-    end
-  end
-end
-
 module AIA
   class RubyLLMAdapter
     attr_reader :tools
@@ -135,7 +115,7 @@ module AIA
     end
 
     def transcribe(audio_file)
-      @chat.ask("Transcribe this audio", with: { audio: audio_file })
+      @chat.ask("Transcribe this audio", with: audio_file)
     end
 
     def speak(text)
@@ -243,21 +223,50 @@ module AIA
       end
     end
 
+
+    #########################################
+    ## text
+
     def text_to_text(prompt)
       text_prompt = extract_text_prompt(prompt)
-      @chat.ask(text_prompt).content
+      response  = if AIA.config.context_files.empty?
+                    @chat.ask(text_prompt)
+                  else
+                    @chat.ask(text_prompt, with: AIA.config.context_files)
+                  end
+
+      response.content
+    rescue => e
+      e.message
+    end
+
+
+    #########################################
+    ## Image
+
+    def extract_image_path(prompt)
+      if prompt.is_a?(String)
+        match = prompt.match(/\b[\w\/\.\-_]+?\.(jpg|jpeg|png|gif|webp)\b/i)
+        match ? match[0] : nil
+      elsif prompt.is_a?(Hash)
+        prompt[:image] || prompt[:image_path]
+      else
+        nil
+      end
     end
 
     def text_to_image(prompt)
       text_prompt = extract_text_prompt(prompt)
-      output_file = "#{Time.now.to_i}.png"
+      image_name  = extract_image_path(text_prompt)
 
       begin
-        RubyLLM.paint(text_prompt, output_path: output_file,
-                      size:     AIA.config.image_size,
-                      quality:  AIA.config.image_quality,
-                      style:    AIA.config.image_style)
-        "Image generated and saved to: #{output_file}"
+        image = RubyLLM.paint(text_prompt, size: AIA.config.image_size)
+        if image_name
+          image_path = image.save(image_name)
+          "Image generated and saved to: #{image_path}"
+        else
+          "Image generated and available at: #{image.url}"
+        end
       rescue => e
         "Error generating image: #{e.message}"
       end
@@ -269,7 +278,7 @@ module AIA
 
       if image_path && File.exist?(image_path)
         begin
-          @chat.ask(text_prompt, with: { image: image_path }).content
+          @chat.ask(text_prompt, with: image_path).content
         rescue => e
           "Error analyzing image: #{e.message}"
         end
@@ -278,13 +287,21 @@ module AIA
       end
     end
 
+
+    #########################################
+    ## audio
+
+    def audio_file?(filepath)
+      filepath.to_s.downcase.end_with?('.mp3', '.wav', '.m4a', '.flac')
+    end
+
     def text_to_audio(prompt)
       text_prompt = extract_text_prompt(prompt)
       output_file = "#{Time.now.to_i}.mp3"
 
       begin
         # Note: RubyLLM doesn't have a direct TTS feature
-        # This is a placeholder for a custom implementation
+        # TODO: This is a placeholder for a custom implementation
         File.write(output_file, text_prompt)
         system("#{AIA.config.speak_command} #{output_file}") if File.exist?(output_file) && system("which #{AIA.config.speak_command} > /dev/null 2>&1")
         "Audio generated and saved to: #{output_file}"
@@ -293,16 +310,23 @@ module AIA
       end
     end
 
-    # TODO: what if its a multi-mode model and a text prompt is provided with
-    #       the audio file?
     def audio_to_text(prompt)
-      text = extract_text_prompt(prompt)
-      text = 'Transcribe this audio' if text.nil? || text.empty?
+      text_prompt = extract_text_prompt(prompt)
+      text_prompt = 'Transcribe this audio' if text_prompt.nil? || text_prompt.empty?
+
+      # TODO: I don't think that "prompt" would ever be an audio filepath.
+      #       Check prompt to see if it is a PromptManager object that has context_files
 
-      if prompt.is_a?(String) && File.exist?(prompt) &&
-         prompt.downcase.end_with?('.mp3', '.wav', '.m4a', '.flac')
+      if  prompt.is_a?(String) &&
+          File.exist?(prompt)  &&
+          audio_file?(prompt)
         begin
-          @chat.ask(text, with: { audio: prompt }).content
+          response  = if AIA.config.context_files.empty?
+                        @chat.ask(text_prompt)
+                      else
+                        @chat.ask(text_prompt, with: AIA.config.context_files)
+                      end
+          response.content
         rescue => e
           "Error transcribing audio: #{e.message}"
         end
@@ -311,15 +335,7 @@ module AIA
         text_to_text(prompt)
       end
     end
-
-    def extract_image_path(prompt)
-      if prompt.is_a?(String)
-        prompt.scan(/\b[\w\/\.\-]+\.(jpg|jpeg|png|gif|webp)\b/i).first&.first
-      elsif prompt.is_a?(Hash)
-        prompt[:image] || prompt[:image_path]
-      else
-        nil
-      end
-    end
   end
 end
+
+__END__

data/lib/aia/session.rb

@@ -123,16 +123,13 @@ module AIA
         prompt_text = "#{prompt_text}\n\nContext:\n#{context}"
       end
 
-      # Determine operation type
-      operation_type = @chat_processor.determine_operation_type(AIA.config.model)
-
       # Add initial user prompt to context *before* sending to AI
       @context_manager.add_to_context(role: 'user', content: prompt_text)
 
       # Process the initial prompt
       @ui_presenter.display_thinking_animation
       # Send the current context (which includes the user prompt)
-      response = @chat_processor.process_prompt(@context_manager.get_context, operation_type)
+      response = @chat_processor.process_prompt(@context_manager.get_context)
 
       # Add AI response to context
       @context_manager.add_to_context(role: 'assistant', content: response)
@@ -202,9 +199,8 @@ module AIA
             @context_manager.add_to_context(role: 'user', content: context)
 
             # Process the context
-            operation_type = @chat_processor.determine_operation_type(AIA.config.model)
             @ui_presenter.display_thinking_animation
-            response = @chat_processor.process_prompt(@context_manager.get_context, operation_type)
+            response = @chat_processor.process_prompt(@context_manager.get_context)
 
             # Add AI response to context
             @context_manager.add_to_context(role: 'assistant', content: response)
@@ -228,9 +224,8 @@ module AIA
 
             @context_manager.add_to_context(role: 'user', content: processed_input)
 
-            operation_type = @chat_processor.determine_operation_type(AIA.config.model)
             @ui_presenter.display_thinking_animation
-            response = @chat_processor.process_prompt(@context_manager.get_context, operation_type)
+            response = @chat_processor.process_prompt(@context_manager.get_context)
 
             @context_manager.add_to_context(role: 'assistant', content: response)
             @chat_processor.output_response(response)
@@ -292,9 +287,8 @@ module AIA
           @context_manager.add_to_context(role: 'user', content: processed_prompt)
           conversation = @context_manager.get_context
 
-          operation_type = @chat_processor.determine_operation_type(AIA.config.model)
           @ui_presenter.display_thinking_animation
-          response = @chat_processor.process_prompt(conversation, operation_type)
+          response = @chat_processor.process_prompt(conversation)
 
           @ui_presenter.display_ai_response(response)
           @context_manager.add_to_context(role: 'assistant', content: response)

data/lib/aia.rb

@@ -12,7 +12,14 @@ include DebugMe
 $DEBUG_ME = false
 DebugMeDefaultOptions[:skip1] = true
 
-require_relative 'extensions/openstruct_merge'
+require_relative 'extensions/openstruct_merge'    # adds self.merge self.get_value
+require_relative 'extensions/ruby_llm/modalities' # adds model.modalities.supports? :text-to-text etc.
+
+require_relative 'refinements/string.rb'        # adds #include_any? #include_all?
+
+
+
+
 require_relative 'aia/utility'
 require_relative 'aia/version'
 require_relative 'aia/config'

data/lib/extensions/ruby_llm/modalities.rb

@@ -0,0 +1,26 @@
+# lib/extensions/ruby_llm/modalities.rb
+# A models "modes" are often expressed in terms like:
+#   text-to-text
+#   text_to_audio
+#   audio to image
+#   image2image
+# This new supports? method tests the models modalities against
+# these common expressions
+
+class RubyLLM::Model::Modalities
+  def supports?(query_mode)
+    parts = query_mode
+              .to_s
+              .downcase
+              .split(/2|-to-| to |_to_/)
+              .map(&:strip)
+
+    if 2 == parts.size
+      input.include?(parts[0]) && output.include?(parts[1])
+    elsif 1 == parts.size
+      input.include?(parts[0]) || output.include?(parts[0])
+    else
+      false
+    end
+  end
+end

data/lib/refinements/string.rb

@@ -0,0 +1,16 @@
+# lib/aia_refinements/string.rb
+
+
+module Refinements
+  refine String do
+    def include_all?(substrings)
+      Array(substrings).all? { |substring| self.include?(substring) }
+    end
+    alias :all? :include_all?
+
+    def include_any?(substrings)
+      Array(substrings).any? { |substring| self.include?(substring) }
+    end
+    alias :any?  :include_any?
+  end
+end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: aia
 version: !ruby/object:Gem::Version
-  version: 0.9.3rc1
+  version: 0.9.4
 platform: ruby
 authors:
 - Dewayne VanHoozer
@@ -57,14 +57,14 @@ dependencies:
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 1.3.0rc1
+        version: 1.3.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 1.3.0rc1
+        version: 1.3.0
 - !ruby/object:Gem::Dependency
   name: reline
   requirement: !ruby/object:Gem::Requirement
@@ -311,6 +311,8 @@ files:
 - lib/aia/utility.rb
 - lib/aia/version.rb
 - lib/extensions/openstruct_merge.rb
+- lib/extensions/ruby_llm/modalities.rb
+- lib/refinements/string.rb
 - main.just
 - mcp_servers/README.md
 - mcp_servers/filesystem.json