smart_prompt 0.4.4 → 0.5.0

data/examples/tts_example.rb

@@ -0,0 +1,244 @@
+# TTS Example for SmartPrompt
+# This example demonstrates how to use the new TTSAdapter
+
+require_relative '../lib/smart_prompt'
+
+# Configuration for TTS capabilities
+config = {
+  "adapters" => {
+    "multimodal" => "MultimodalAdapter",
+    "image_generation" => "ImageGenerationAdapter",
+    "video_generation" => "VideoGenerationAdapter",
+    "tts" => "TTSAdapter"
+  },
+  "llms" => {
+    "qwen_vl" => {
+      "adapter" => "multimodal",
+      "url" => "https://api.siliconflow.cn/v1/",
+      "api_key" => ENV["SILICONFLOW_API_KEY"],
+      "model" => "Qwen/Qwen2.5-VL-7B-Instruct"
+    },
+    "image_gen" => {
+      "adapter" => "image_generation",
+      "url" => "https://api.siliconflow.cn/v1/",
+      "api_key" => ENV["SILICONFLOW_API_KEY"],
+      "model" => "stabilityai/stable-diffusion-xl-base-1.0"
+    },
+    "video_gen" => {
+      "adapter" => "video_generation",
+      "url" => "https://api.siliconflow.cn/v1/",
+      "api_key" => ENV["SILICONFLOW_API_KEY"],
+      "model" => "Wan-AI/Wan2.2-T2V-A14B"
+    },
+    "tts_service" => {
+      "adapter" => "tts",
+      "url" => "https://api.siliconflow.cn/v1/",
+      "api_key" => ENV["SILICONFLOW_API_KEY"],
+      "model" => "FunAudioLLM/CosyVoice2-0.5B"
+    }
+  },
+  "default_llm" => "qwen_vl",
+  "template_path" => "./templates",
+  "worker_path" => "./workers",
+  "logger_file" => "./logs/smart_prompt.log"
+}
+
+# Write config to file
+File.write('tts_config.yml', config.to_yaml)
+
+# Initialize engine
+engine = SmartPrompt::Engine.new('tts_config.yml')
+
+puts "=== SmartPrompt TTS Demo ==="
+
+# Example 1: Basic text-to-speech synthesis
+puts "\n=== Example 1: Basic TTS Synthesis ==="
+begin
+  result = engine.call_worker(:tts_synthesizer, {
+    text: "欢迎使用智能提示系统，这是一个文本转语音功能的演示。",
+    voice: "alloy",
+    speed: 1.0,
+    response_format: "mp3",
+    save_to_file: true,
+    output_dir: "./generated_audio",
+    filename_prefix: "basic_tts"
+  })
+
+  puts "TTS synthesis successful!"
+  puts "Audio file: #{result[:audio_file][:file_path]}"
+  puts "Text length: #{result[:audio_file][:text_length]} characters"
+  puts "Voice: #{result[:audio_file][:voice]}"
+
+rescue => e
+  puts "Error in TTS synthesis: #{e.message}"
+  puts "Note: This example requires a valid SILICONFLOW_API_KEY environment variable"
+end
+
+# Example 2: Multilingual TTS
+puts "\n=== Example 2: Multilingual TTS ==="
+begin
+  # English text
+  result_en = engine.call_worker(:multilingual_tts, {
+    text: "Hello, this is a demonstration of text-to-speech functionality.",
+    voice: "echo",
+    save_to_file: true,
+    output_dir: "./multilingual_audio",
+    filename_prefix: "english_tts"
+  })
+
+  puts "English TTS successful!"
+  puts "Detected language: #{result_en[:detected_language]}"
+  puts "Audio file: #{result_en[:audio_file][:file_path]}"
+
+  # Chinese text
+  result_zh = engine.call_worker(:multilingual_tts, {
+    text: "这是一个中文文本转语音的演示，支持多种语言。",
+    voice: "nova",
+    save_to_file: true,
+    output_dir: "./multilingual_audio",
+    filename_prefix: "chinese_tts"
+  })
+
+  puts "Chinese TTS successful!"
+  puts "Detected language: #{result_zh[:detected_language]}"
+  puts "Audio file: #{result_zh[:audio_file][:file_path]}"
+
+rescue => e
+  puts "Error in multilingual TTS: #{e.message}"
+end
+
+# Example 3: Voice selection demo
+puts "\n=== Example 3: Voice Selection Demo ==="
+begin
+  result = engine.call_worker(:voice_selector, {
+    text: "这是一个不同音色的演示，您可以听到不同声音的朗读效果。",
+    save_to_file: true,
+    output_dir: "./voice_demos"
+  })
+
+  puts "Voice selection demo successful!"
+  puts "Available voices: #{result[:available_voices].keys.join(', ')}"
+  puts "Selected voice: #{result[:selected_voice]}"
+  puts "Audio file: #{result[:audio_file][:file_path]}"
+
+rescue => e
+  puts "Error in voice selection: #{e.message}"
+end
+
+# Example 4: Speed variation demo
+puts "\n=== Example 4: Speed Variation Demo ==="
+begin
+  result = engine.call_worker(:speed_variation_tts, {
+    text: "这是一个语速变化的演示，您可以听到不同语速的朗读效果。",
+    voice: "alloy",
+    speeds: [0.5, 0.75, 1.0, 1.5, 2.0],
+    save_to_file: true,
+    output_dir: "./speed_variations"
+  })
+
+  puts "Speed variation demo successful!"
+  puts "Generated #{result[:speed_variations].size} audio files at different speeds"
+  result[:speed_variations].each do |variation|
+    puts "  - Speed #{variation[:speed]}: #{variation[:audio_file][:file_path]}"
+  end
+
+rescue => e
+  puts "Error in speed variation: #{e.message}"
+end
+
+# Example 5: Custom voice management
+puts "\n=== Example 5: Custom Voice Management ==="
+begin
+  # List available voices
+  result = engine.call_worker(:custom_voice_manager, {
+    action: "list"
+  })
+
+  puts "Voice management demo successful!"
+  puts "Predefined voices: #{result[:predefined_voices].keys.join(', ')}"
+  puts "Custom voices: #{result[:custom_voices].size}"
+
+  # Note: Creating custom voices requires reference audio files
+  # Uncomment the following lines if you have reference audio files:
+  #
+  # result = engine.call_worker(:custom_voice_manager, {
+  #   action: "create",
+  #   name: "my_custom_voice",
+  #   reference_audio_file: "./reference_audio.wav",
+  #   description: "My custom voice created from reference audio"
+  # })
+  #
+  # puts "Custom voice created: #{result[:voice_data][:voice_id]}"
+
+rescue => e
+  puts "Error in voice management: #{e.message}"
+end
+
+# Example 6: Batch TTS processing
+puts "\n=== Example 6: Batch TTS Processing ==="
+begin
+  result = engine.call_worker(:batch_tts, {
+    texts: [
+      "这是第一条文本内容。",
+      "这是第二条文本内容，用于批量处理演示。",
+      "这是第三条文本内容，展示批量文本转语音功能。"
+    ],
+    voice: "alloy",
+    save_to_file: true,
+    output_dir: "./batch_audio"
+  })
+
+  puts "Batch TTS processing successful!"
+  puts "Generated #{result[:batch_results].size} audio files"
+  result[:batch_results].each do |batch_result|
+    puts "  - Text #{batch_result[:index] + 1}: #{batch_result[:audio_file][:file_path]}"
+  end
+
+rescue => e
+  puts "Error in batch TTS: #{e.message}"
+end
+
+# Example 7: Direct adapter usage
+puts "\n=== Example 7: Direct Adapter Usage ==="
+begin
+  # Get the adapter directly
+  adapter = engine.llms["tts_service"]
+
+  # Synthesize speech directly
+  audio_data = adapter.synthesize_speech(
+    "这是直接使用适配器的演示，不通过Worker。",
+    voice: "echo",
+    speed: 1.2,
+    response_format: "mp3"
+  )
+
+  puts "Direct adapter usage successful!"
+  puts "Generated audio data with format: #{audio_data[:format]}"
+  puts "Text length: #{audio_data[:text_length]} characters"
+
+  # Save to file
+  output_path = "./direct_audio/direct_tts_#{Time.now.to_i}.mp3"
+  result = adapter.synthesize_to_file(
+    "这是直接保存到文件的演示。",
+    output_path,
+    voice: "nova",
+    speed: 1.0
+  )
+
+  puts "Direct file synthesis successful!"
+  puts "Audio file: #{result[:file_path]}"
+
+rescue => e
+  puts "Error in direct adapter usage: #{e.message}"
+end
+
+puts "\n=== All examples completed ==="
+puts "\nImportant Notes:"
+puts "1. TTS requires valid SILICONFLOW_API_KEY environment variable"
+puts "2. Audio files are saved in various formats (mp3, wav, etc.)"
+puts "3. Custom voice creation requires reference audio files"
+puts "4. Multiple languages are supported (Chinese, English, Japanese, Korean)"
+puts "5. Speed can be adjusted from 0.25x to 4.0x"
+
+# Clean up
+File.delete('tts_config.yml') if File.exist?('tts_config.yml')

data/examples/video_generation_example.rb

@@ -0,0 +1,189 @@
+# Video Generation Example for SmartPrompt
+# This example demonstrates how to use the new VideoGenerationAdapter
+
+require_relative '../lib/smart_prompt'
+
+# Configuration for video generation capabilities
+config = {
+  "adapters" => {
+    "multimodal" => "MultimodalAdapter",
+    "image_generation" => "ImageGenerationAdapter",
+    "video_generation" => "VideoGenerationAdapter"
+  },
+  "llms" => {
+    "qwen_vl" => {
+      "adapter" => "multimodal",
+      "url" => "https://api.siliconflow.cn/v1/",
+      "api_key" => ENV["SILICONFLOW_API_KEY"],
+      "model" => "Qwen/Qwen2.5-VL-7B-Instruct"
+    },
+    "image_gen" => {
+      "adapter" => "image_generation",
+      "url" => "https://api.siliconflow.cn/v1/",
+      "api_key" => ENV["SILICONFLOW_API_KEY"],
+      "model" => "stabilityai/stable-diffusion-xl-base-1.0"
+    },
+    "video_gen" => {
+      "adapter" => "video_generation",
+      "url" => "https://api.siliconflow.cn/v1/",
+      "api_key" => ENV["SILICONFLOW_API_KEY"],
+      "model" => "Wan-AI/Wan2.2-T2V-A14B"
+    }
+  },
+  "default_llm" => "qwen_vl",
+  "template_path" => "./templates",
+  "worker_path" => "./workers",
+  "logger_file" => "./logs/smart_prompt.log"
+}
+
+# Write config to file
+File.write('video_generation_config.yml', config.to_yaml)
+
+# Initialize engine
+engine = SmartPrompt::Engine.new('video_generation_config.yml')
+
+puts "=== SmartPrompt Video Generation Demo ==="
+
+# Example 1: Simple text-to-video generation
+puts "\n=== Example 1: Text-to-Video Generation ==="
+begin
+  result = engine.call_worker(:video_generator, {
+    prompt: "A beautiful sunset over ocean waves, cinematic quality, slow motion",
+    duration: 4,
+    resolution: "720p",
+    fps: 24,
+    wait_for_completion: false, # Set to true to wait for completion
+    download_to_file: false,     # Set to true to download video
+    output_dir: "./generated_videos",
+    filename_prefix: "sunset_video"
+  })
+
+  puts "Video generation job submitted successfully!"
+  puts "Job ID: #{result[:video_data][:job_id]}"
+  puts "Status: #{result[:video_data][:status]}"
+  puts "Created at: #{result[:video_data][:created_at]}"
+
+rescue => e
+  puts "Error in video generation: #{e.message}"
+  puts "Note: This example requires a valid SILICONFLOW_API_KEY environment variable"
+  puts "Note: Video generation may take several minutes to complete"
+end
+
+# Example 2: Creative video generation with style
+puts "\n=== Example 2: Creative Video Generation ==="
+begin
+  result = engine.call_worker(:creative_video_generator, {
+    prompt: "A magical forest with glowing fairies and sparkling lights",
+    video_style: "fantasy animation, Studio Ghibli style",
+    duration: 4,
+    resolution: "720p",
+    fps: 24,
+    wait_for_completion: false,
+    download_to_file: false
+  })
+
+  puts "Creative video generation job submitted successfully!"
+  puts "Job ID: #{result[:video_data][:job_id]}"
+  puts "Status: #{result[:video_data][:status]}"
+
+rescue => e
+  puts "Error in creative video generation: #{e.message}"
+end
+
+# Example 3: Product video generation
+puts "\n=== Example 3: Product Video Generation ==="
+begin
+  result = engine.call_worker(:product_video_generator, {
+    prompt: "A modern smartphone rotating slowly on a marble surface",
+    duration: 4,
+    resolution: "720p",
+    fps: 24,
+    wait_for_completion: false,
+    download_to_file: false
+  })
+
+  puts "Product video generation job submitted successfully!"
+  puts "Job ID: #{result[:video_data][:job_id]}"
+  puts "Status: #{result[:video_data][:status]}"
+
+rescue => e
+  puts "Error in product video generation: #{e.message}"
+end
+
+# Example 4: Check video status (if we have a job ID from previous examples)
+puts "\n=== Example 4: Video Status Check ==="
+begin
+  # This example requires a valid job_id from a previous video generation
+  # For demonstration, we'll show the method but skip execution
+  puts "To check video status, use:"
+  puts "result = engine.call_worker(:video_status_checker, {"
+  puts "  job_id: 'YOUR_JOB_ID_HERE',"
+  puts "  download_to_file: true"
+  puts "})"
+
+rescue => e
+  puts "Error in status check: #{e.message}"
+end
+
+# Example 5: Direct adapter usage (without worker)
+puts "\n=== Example 5: Direct Adapter Usage ==="
+begin
+  # Get the adapter directly
+  adapter = engine.llms["video_gen"]
+
+  # Generate video directly
+  video_data = adapter.generate_video(
+    "A butterfly flying through a flower garden, nature documentary style",
+    duration: 4,
+    resolution: "720p",
+    fps: 24
+  )
+
+  puts "Direct adapter usage successful!"
+  puts "Job ID: #{video_data[:job_id]}"
+  puts "Status: #{video_data[:status]}"
+
+  # Example of checking status
+  # status = adapter.check_video_status(video_data[:job_id])
+  # puts "Current status: #{status[:status]}, Progress: #{status[:progress]}"
+
+rescue => e
+  puts "Error in direct adapter usage: #{e.message}"
+end
+
+# Example 6: Batch video generation
+puts "\n=== Example 6: Batch Video Generation ==="
+begin
+  result = engine.call_worker(:batch_video_generator, {
+    prompts: [
+      "A cat playing with a ball of yarn",
+      "A dog running through a field",
+      "A bird flying in the sky"
+    ],
+    duration: 3,
+    resolution: "720p",
+    fps: 24,
+    wait_for_completion: false
+  })
+
+  puts "Batch video generation submitted successfully!"
+  puts "Generated #{result[:batch_results].size} video jobs"
+  result[:batch_results].each do |result|
+    puts "  - Prompt: #{result[:prompt][0..50]}..."
+    puts "    Job ID: #{result[:video_data][:job_id]}"
+  end
+
+rescue => e
+  puts "Error in batch video generation: #{e.message}"
+end
+
+puts "\n=== All examples completed ==="
+puts "\nImportant Notes:"
+puts "1. Video generation is an asynchronous process"
+puts "2. Jobs may take several minutes to complete"
+puts "3. Use wait_for_completion: true to wait for completion"
+puts "4. Use download_to_file: true to automatically download videos"
+puts "5. Check status periodically using video_status_checker worker"
+
+# Clean up
+File.delete('video_generation_config.yml') if File.exist?('video_generation_config.yml')

data/examples/zhipu_example.rb

@@ -0,0 +1,151 @@
+# 智谱 AI (BigModel / GLM) Example for SmartPrompt
+#
+# Demonstrates every Zhipu model category through one ZhipuAIAdapter:
+#   1. 文本对话 (chat) — sync + streaming
+#   2. 图文多模态 (vision)
+#   3. 向量模型 (embeddings)
+#   4. 文生图 (CogView)
+#   5. 文生视频 (CogVideoX, async submit -> poll -> download)
+#   6. 语音合成 (GLM-TTS)
+#   7. 语音识别 (GLM-ASR-2512)
+#
+# Requires a valid Zhipu API key in ZHIPUAI_API_KEY (https://open.bigmodel.cn/).
+# Defaults use free-tier models so it works out-of-box once the key is set.
+
+require_relative "../lib/smart_prompt"
+
+api_key = ENV["ZHIPUAI_API_KEY"]
+base = "https://open.bigmodel.cn/api/paas/v4"
+
+config = {
+  "adapters" => { "zhipu" => "ZhipuAIAdapter" },
+  "llms" => {
+    "glm"        => { "adapter" => "zhipu", "url" => base, "api_key" => api_key, "model" => "glm-4-flash" },
+    "glm_vision" => { "adapter" => "zhipu", "url" => base, "api_key" => api_key, "model" => "glm-4v-flash" },
+    "embedding"  => { "adapter" => "zhipu", "url" => base, "api_key" => api_key, "model" => "embedding-3", "dimensions" => 1024 },
+    "cogview"    => { "adapter" => "zhipu", "url" => base, "api_key" => api_key, "model" => "cogview-3-flash" },
+    "cogvideo"   => { "adapter" => "zhipu", "url" => base, "api_key" => api_key, "model" => "cogvideox-flash" },
+    "glm_tts"    => { "adapter" => "zhipu", "url" => base, "api_key" => api_key, "model" => "glm-tts" },
+    "glm_asr"    => { "adapter" => "zhipu", "url" => base, "api_key" => api_key, "model" => "glm-asr-2512" },
+  },
+  "default_llm" => "glm",
+  "template_path" => "./templates",
+  "worker_path" => "./workers",
+  "logger_file" => "./logs/smart_prompt.log",
+}
+
+File.write("zhipu_config.yml", config.to_yaml)
+engine = SmartPrompt::Engine.new("zhipu_config.yml")
+
+puts "=== SmartPrompt 智谱 GLM Demo ==="
+unless api_key
+  puts "Note: ZHIPUAI_API_KEY is not set — the API calls below will fail at the network layer."
+end
+
+# 1. Chat (sync)
+puts "\n=== Example 1: 文本对话 (sync) ==="
+begin
+  result = engine.call_worker(:glm_chat, { prompt: "用一句话介绍智谱GLM。" })
+  puts "Reply: #{result}"
+rescue => e
+  puts "Error: #{e.message}"
+end
+
+# 2. Chat (streaming)
+puts "\n=== Example 2: 文本对话 (streaming) ==="
+begin
+  engine.call_worker_by_stream(:glm_chat, { prompt: "写两句关于春天的诗。" }) do |chunk, _|
+    print chunk.dig("choices", 0, "delta", "content").to_s
+  end
+  puts
+rescue => e
+  puts "Error: #{e.message}"
+end
+
+# 3. Multimodal vision
+puts "\n=== Example 3: 图文多模态 ==="
+begin
+  result = engine.call_worker(:glm_vision, {
+    image_url: "https://img1.baidu.com/it/u=1966616150,2146512490&fm=253&fmt=auto&app=138&f=JPEG?w=500&h=282",
+    question: "图片里有什么？",
+  })
+  puts "Vision result: #{result}"
+rescue => e
+  puts "Error: #{e.message}"
+end
+
+# 4. Embeddings (embedding-3)
+puts "\n=== Example 4: 向量模型 ==="
+begin
+  vector = engine.call_worker(:glm_embed, { text: "智谱GLM大模型", length: 1024 })
+  puts "Embedding dim: #{vector.is_a?(Array) ? vector.size : vector} (first 5: #{vector.first(5) rescue vector})"
+rescue => e
+  puts "Error: #{e.message}"
+end
+
+# 5. Text-to-image (CogView)
+puts "\n=== Example 5: 文生图 ==="
+begin
+  result = engine.call_worker(:cogview_image, {
+    prompt: "一只在书房里读书的猫，水墨画风格",
+    size: "1024x1024",
+    save_to_file: true,
+    output_dir: "./generated_images",
+    filename_prefix: "zhipu_cat",
+  })
+  if result.is_a?(Hash) && result[:images]
+    puts "Generated #{result[:images].size} image(s); first URL: #{result[:images].first[:url]}"
+    puts "Saved files: #{result[:saved_files]}"
+  else
+    puts "Result: #{result}"
+  end
+rescue => e
+  puts "Error: #{e.message}"
+end
+
+# 6. Text-to-video (CogVideoX, async) — may take a minute or two.
+puts "\n=== Example 6: 文生视频 (async) ==="
+begin
+  result = engine.call_worker(:cogvideo_video, {
+    prompt: "一只猫在阳光下打盹",
+    wait_for_completion: true,
+    download_to_file: true,
+    output_dir: "./generated_videos",
+    timeout: 600,
+  })
+  if result[:video]
+    puts "Video ready: #{result[:video][:video_url]}"
+    puts "Downloaded: #{result[:downloaded_file]}" if result[:downloaded_file]
+  else
+    puts "Submitted task: #{result[:submitted]}"
+  end
+rescue => e
+  puts "Error: #{e.message}"
+end
+
+# 7. TTS (GLM-TTS)
+puts "\n=== Example 7: 语音合成 (TTS) ==="
+begin
+  info = engine.call_worker(:glm_tts, { text: "你好，这是智谱语音合成的测试。", output_path: "./generated_audio/zhipu_tts.wav" })
+  puts "Audio saved: #{info[:file_path]}"
+rescue => e
+  puts "Error: #{e.message}"
+end
+
+# 8. ASR (GLM-ASR-2512) — needs a real audio file path.
+puts "\n=== Example 8: 语音识别 (ASR) ==="
+audio = ENV["ZHIPU_ASR_SAMPLE"] || "./generated_audio/zhipu_tts.wav"
+if File.exist?(audio)
+  begin
+    result = engine.call_worker(:glm_asr, { audio_file: audio })
+    puts "Transcription: #{result[:text]}"
+  rescue => e
+    puts "Error: #{e.message}"
+  end
+else
+  puts "Skipped: set ZHIPU_ASR_SAMPLE to an audio file path (or run TTS first) to test ASR."
+end
+
+puts "\n=== All examples completed ==="
+
+File.delete("zhipu_config.yml") if File.exist?("zhipu_config.yml")