ElevenlabsClient

gem 'elevenlabs_client'

ElevenlabsClient::Settings.configure do |config|
  config.properties = {
    elevenlabs_base_uri: ENV["ELEVENLABS_BASE_URL"],
    elevenlabs_api_key: ENV["ELEVENLABS_API_KEY"]
  }
end

# Module-level configuration
ElevenlabsClient.configure do |config|
  config.properties = {
    elevenlabs_base_uri: "https://api.elevenlabs.io",
    elevenlabs_api_key: "your_api_key_here"
  }
end

# Or pass directly to client
client = ElevenlabsClient.new(
  api_key: "your_api_key_here",
  base_url: "https://api.elevenlabs.io"
)

# Initialize client (uses configured settings)
client = ElevenlabsClient.new

# Text-to-Speech
audio_data = client.text_to_speech.convert("21m00Tcm4TlvDq8ikWAM", "Hello, world!")
File.open("hello.mp3", "wb") { |f| f.write(audio_data) }

# Dubbing
File.open("video.mp4", "rb") do |file|
  result = client.dubs.create(
    file_io: file,
    filename: "video.mp4",
    target_languages: ["es", "fr", "de"]
  )
end

# Dialogue Generation
dialogue = [
  { text: "Hello, how are you?", voice_id: "voice_1" },
  { text: "I'm doing great, thanks!", voice_id: "voice_2" }
]
audio_data = client.text_to_dialogue.convert(dialogue)

# Sound Generation
audio_data = client.sound_generation.generate("Ocean waves crashing on rocks")

# Voice Design
design_result = client.text_to_voice.design("Warm, professional female voice")
generated_voice_id = design_result["previews"].first["generated_voice_id"]

voice_result = client.text_to_voice.create(
  "Professional Voice",
  "Warm, professional female voice",
  generated_voice_id
)

# List Available Models
models = client.models.list
fastest_model = models["models"].min_by { |m| m["token_cost_factor"] }
puts "Fastest model: #{fastest_model['name']}"

# Voice Management
voices = client.voices.list
puts "Total voices: #{voices['voices'].length}"

# Create custom voice from audio samples
File.open("sample1.mp3", "rb") do |sample|
  voice = client.voices.create("My Voice", [sample], description: "Custom narrator voice")
  puts "Created voice: #{voice['voice_id']}"
end

# Music Generation
music_data = client.music.compose(
  prompt: "Upbeat electronic dance track with synthesizers",
  music_length_ms: 30000
)
File.open("generated_music.mp3", "wb") { |f| f.write(music_data) }

# Speech-to-Speech (Voice Changer)
File.open("input_audio.mp3", "rb") do |audio_file|
  converted_audio = client.speech_to_speech.convert(
    "target_voice_id", 
    audio_file, 
    "input_audio.mp3",
    remove_background_noise: true
  )
  File.open("converted_audio.mp3", "wb") { |f| f.write(converted_audio) }
end

# Speech-to-Text Transcription
File.open("audio.mp3", "rb") do |audio_file|
  transcription = client.speech_to_text.create(
    "scribe_v1",
    file: audio_file,
    filename: "audio.mp3",
    diarize: true,
    timestamps_granularity: "word"
  )
  puts "Transcribed: #{transcription['text']}"
end

# Audio Isolation (Background Noise Removal)
File.open("noisy_audio.mp3", "rb") do |audio_file|
  clean_audio = client.audio_isolation.isolate(audio_file, "noisy_audio.mp3")
  File.open("clean_audio.mp3", "wb") { |f| f.write(clean_audio) }
end

# Audio Native (Embeddable Player)
File.open("article.html", "rb") do |html_file|
  project = client.audio_native.create(
    "My Article",
    file: html_file,
    filename: "article.html",
    voice_id: "voice_id",
    auto_convert: true
  )
  puts "Player HTML: #{project['html_snippet']}"
end

# Forced Alignment
File.open("speech.wav", "rb") do |audio_file|
  alignment = client.forced_alignment.create(
    audio_file,
    "speech.wav",
    "Hello world, this is a test transcript"
  )
  
  alignment['words'].each do |word|
    puts "#{word['text']}: #{word['start']}s - #{word['end']}s"
  end
end

# Streaming Text-to-Speech
client.text_to_speech_stream.stream("voice_id", "Streaming text") do |chunk|
  # Process audio chunk in real-time
  puts "Received #{chunk.bytesize} bytes"
end

client = ElevenlabsClient.new(
  api_key_env: "CUSTOM_API_KEY_VAR",
  base_url_env: "CUSTOM_BASE_URL_VAR"
)

begin
  result = client.text_to_speech.convert(voice_id, text)
rescue ElevenlabsClient::AuthenticationError
  puts "Invalid API key"
rescue ElevenlabsClient::RateLimitError
  puts "Rate limit exceeded"
rescue ElevenlabsClient::ValidationError => e
  puts "Invalid parameters: #{e.message}"
rescue ElevenlabsClient::APIError => e
  puts "API error: #{e.message}"
end