Class: Typecast::SpeechComposer

Inherits:
Object
  • Object
show all
Defined in:
lib/typecast/composer.rb

Instance Method Summary collapse

Constructor Details

#initialize(compose) ⇒ SpeechComposer

Returns a new instance of SpeechComposer.



29
30
31
32
33
 
# File 'lib/typecast/composer.rb', line 29

def initialize(compose)
  @compose = compose
  @defaults = {}
  @parts = []
end

Instance Method Details

#defaults(voice_id: nil, model: nil, language: nil, prompt: nil, output: nil, seed: nil) ⇒ Object 



35
36
37
38
39
40
41
42
43
44
45
 
# File 'lib/typecast/composer.rb', line 35

def defaults(voice_id: nil, model: nil, language: nil, prompt: nil, output: nil, seed: nil)
  @defaults = merge_settings(@defaults, settings_hash(
    voice_id: voice_id,
    model: model,
    language: language,
    prompt: prompt,
    output: output,
    seed: seed
  ))
  self
end

#generate ⇒ Object

Raises:

  • (ArgumentError) 


75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
 
# File 'lib/typecast/composer.rb', line 75

def generate
  plan = build_plan
  unless plan.any? { |part| part.is_a?(Hash) && part[:kind] == "speech" }
    raise ArgumentError, "at least one speech segment is required"
  end

  formats = plan.each_with_object([]) do |part, values|
    format = part.is_a?(Hash) && part[:kind] == "speech" ? part.dig(:settings, :output, :audio_format) : nil
    values << format if format
  end.uniq
  raise ArgumentError, "composed speech segments must use one audio format" if formats.length > 1

  output_format = formats.first || Models::AUDIO_WAV
  unless [Models::AUDIO_WAV, Models::AUDIO_MP3].include?(output_format)
    raise ArgumentError, "unsupported composed speech output format: #{output_format}"
  end

  segments = plan.map do |part|
    part.is_a?(PausePart) ? { type: "pause", duration_seconds: part.seconds } :
      { type: "tts", **request_from_settings(part[:text], part[:settings], output_format).to_h }
  end
  @compose.call(segments)
end

#pause(seconds) ⇒ Object

Inserts silence between speech segments.

seconds is a duration in seconds. Use 0.3 for 300 ms, 3 for 3 seconds.



66
67
68
69
70
71
72
73
 
# File 'lib/typecast/composer.rb', line 66

def pause(seconds)
  unless seconds.is_a?(Numeric) && seconds.finite? && seconds.positive?
    raise ArgumentError, "pause seconds must be greater than 0"
  end

  @parts << PausePart.new(kind: "pause", seconds: seconds.to_f)
  self
end

#say(text, voice_id: nil, model: nil, language: nil, prompt: nil, output: nil, seed: nil) ⇒ Object 



47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
 
# File 'lib/typecast/composer.rb', line 47

def say(text, voice_id: nil, model: nil, language: nil, prompt: nil, output: nil, seed: nil)
  @parts << {
    kind: "speech",
    text: text.to_s,
    settings: merge_settings(@defaults, settings_hash(
      voice_id: voice_id,
      model: model,
      language: language,
      prompt: prompt,
      output: output,
      seed: seed
    ))
  }
  self
end