omniai 1.0.9 → 1.1.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 2f4665d73734c57eaf861adb6a941c7e3102c5de00ec4739f29b69123b829a1d
-  data.tar.gz: '03487dbb919fb7821be77fcbecac4d0adefa5c62d511fffa2ca06d6c3f28c946'
+  metadata.gz: 7a31351fb7cc352c108f55c3df8d57020bf2d520b935dce75eb87e62e4a0737c
+  data.tar.gz: a8ca504f40836a0072dcf4549767632a19bb6497da53390404fc1399ae264522
 SHA512:
-  metadata.gz: 105c7d3af61b4f79b1bdfb15c083ea48cbe92bb6b81bc2ccdccef2f8873ed2b358cbf3a7c3418187a7b2b4d27d0768a5c1a8ce36866d263eb107cee91842011f
-  data.tar.gz: 4ddad063456e0d1a98c6b21c284fb66ba5ce672f6b4df4f03522ebf3f341269ebecf28607132a84120b9de621b9e5ca65b219b0f072bbf36bc5f4970461458a4
+  metadata.gz: 76f22ab7edf30b0f5aedbe071d7a1ff0fd5c1b2a5ab1f5a36273b9041c62a27796a510a4c73fd52f27606ff4a5e8d84fbb46cea03a6949da71ada89cdc62daba
+  data.tar.gz: 73d4718a9a593e5c520f48f8a1fa6c06560a2f3447c609f1a85ecbbda9c7d55dc45a6bec9d57de5054b1936be1d21266bbcb3324096a8337db1cf05d7359f09a

data/README.md

@@ -111,3 +111,11 @@ stream = proc do |chunk|
 end
 client.chat('Tell me a joke.', stream:)
 ```
+
+### Transcribe
+
+Clients that support chat (e.g. OpenAI w/ "Whisper", etc) convert recordings to text via the following calls:
+
+```ruby
+client.transcribe(file.path)
+```

data/lib/omniai/chat.rb

@@ -33,6 +33,10 @@ module OmniAI
       SYSTEM = 'system'
     end
 
+    module Format
+      JSON = :json
+    end
+
     def self.process!(...)
       new(...).process!
     end

data/lib/omniai/client.rb

@@ -45,5 +45,19 @@ module OmniAI
     def chat(messages, model:, temperature: nil, format: nil, stream: nil)
       raise NotImplementedError, "#{self.class.name}#chat undefined"
     end
+
+    # @raise [OmniAI::Error]
+    #
+    # @param file [IO]
+    # @param model [String]
+    # @param language [String, nil] optional
+    # @param prompt [String, nil] optional
+    # @param temperature [Float, nil] optional
+    # @param format [Symbol] :text, :srt, :vtt, or :json (default)
+    #
+    # @return text [OmniAI::Transcribe::Transcription]
+    def transcribe(file, model:, language: nil, prompt: nil, temperature: nil, format: Transcription::Format::JSON)
+      raise NotImplementedError, "#{self.class.name}#speak undefined"
+    end
   end
 end

data/lib/omniai/transcribe/transcription.rb

@@ -0,0 +1,26 @@
+# frozen_string_literal: true
+
+module OmniAI
+  class Transcribe
+    # A transcription returned by the API.
+    class Transcription
+      attr_accessor :data, :format
+
+      # @param data [Hash]
+      def initialize(data:, format:)
+        @data = data
+        @format = format
+      end
+
+      # @return [String]
+      def text
+        @data['text']
+      end
+
+      # @return [String]
+      def inspect
+        "#<#{self.class} text=#{text.inspect} format=#{format.inspect}>"
+      end
+    end
+  end
+end

data/lib/omniai/transcribe.rb

@@ -0,0 +1,152 @@
+# frozen_string_literal: true
+
+module OmniAI
+  # An abstract class that provides a consistent interface for processing transcribe requests.
+  #
+  # Usage:
+  #
+  #   class OmniAI::OpenAI::Transcribe < OmniAI::Transcribe
+  #     module Model
+  #       WHISPER_1 = "whisper-1"
+  #     end
+  #
+  #     protected
+  #
+  #     # @return [Hash]
+  #     def payload
+  #       raise NotImplementedError, "#{self.class.name}#payload undefined"
+  #     end
+  #
+  #     # @return [String]
+  #     def path
+  #       raise NotImplementedError, "#{self.class.name}#path undefined"
+  #     end
+  #   end
+  #
+  #   client.transcribe(File.open("..."), model: "...", format: :json)
+  class Transcribe
+    module Language
+      AFRIKAANS = 'af'
+      ARABIC = 'ar'
+      ARMENIAN = 'hy'
+      AZERBAIJANI = 'az'
+      BELARUSIAN = 'be'
+      BOSNIAN = 'bs'
+      BULGARIAN = 'bg'
+      CATALAN = 'ca'
+      CHINESE = 'zh'
+      CROATIAN = 'hr'
+      CZECH = 'cs'
+      DANISH = 'da'
+      DUTCH = 'nl'
+      ENGLISH = 'en'
+      ESTONIAN = 'et'
+      FINNISH = 'fi'
+      FRENCH = 'fr'
+      GALICIAN = 'gl'
+      GERMAN = 'de'
+      GREEK = 'el'
+      HEBREW = 'he'
+      HINDI = 'hi'
+      HUNGARIAN = 'hu'
+      ICELANDIC = 'is'
+      INDONESIAN = 'id'
+      ITALIAN = 'it'
+      JAPANESE = 'ja'
+      KANNADA = 'kn'
+      KAZAKH = 'kk'
+      KOREAN = 'ko'
+      LATVIAN = 'lv'
+      LITHUANIAN = 'lt'
+      MACEDONIAN = 'mk'
+      MALAY = 'ms'
+      MARATHI = 'mr'
+      MAORI = 'mi'
+      NEPALI = 'ne'
+      NORWEGIAN = 'no'
+      PERSIAN = 'fa'
+      POLISH = 'pl'
+      PORTUGUESE = 'pt'
+      ROMANIAN = 'ro'
+      RUSSIAN = 'ru'
+      SERBIAN = 'sr'
+      SLOVAK = 'sk'
+      SLOVENIAN = 'sl'
+      SPANISH = 'es'
+      SWAHILI = 'sw'
+      SWEDISH = 'sv'
+      TAGALOG = 'tl'
+      TAMIL = 'ta'
+      THAI = 'th'
+      TURKISH = 'tr'
+      UKRAINIAN = 'uk'
+      URDU = 'ur'
+      VIETNAMESE = 'vi'
+      WELSH = 'cy'
+    end
+
+    module Format
+      JSON = 'json'
+      TEXT = 'text'
+      VTT = 'vtt'
+      SRT = 'srt'
+    end
+
+    def self.process!(...)
+      new(...).process!
+    end
+
+    # @param path [String] required
+    # @param client [OmniAI::Client] the client
+    # @param model [String] required
+    # @param language [String, nil] optional
+    # @param prompt [String, nil] optional
+    # @param temperature [Float, nil] optional
+    # @param format [String, nil] optional
+    def initialize(path, client:, model:, language: nil, prompt: nil, temperature: nil, format: Format::JSON)
+      @path = path
+      @model = model
+      @language = language
+      @prompt = prompt
+      @temperature = temperature
+      @format = format
+      @client = client
+    end
+
+    # @return [OmniAI::Transcribe::Transcription]
+    # @raise [ExecutionError]
+    def process!
+      response = request!
+      raise HTTPError, response.flush unless response.status.ok?
+
+      data = @format.eql?(Format::JSON) ? response.parse : { text: String(response.body) }
+      Transcription.new(format: @format, data:)
+    end
+
+    protected
+
+    # @return [Hash]
+    def payload
+      {
+        file: HTTP::FormData::File.new(@path),
+        model: @model,
+        language: @language,
+        prompt: @prompt,
+        temperature: @temperature,
+      }.compact
+    end
+
+    # @return [String]
+    def path
+      raise NotImplementedError, "#{self.class.name}#path undefined"
+    end
+
+    # @return [HTTP::Response]
+    def request!
+      @client
+        .connection
+        .accept(@format.eql?(Format::JSON) ? :json : :text)
+        .post(path, form: payload)
+    end
+  end
+end

data/lib/omniai/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module OmniAI
-  VERSION = '1.0.9'
+  VERSION = '1.1.0'
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: omniai
 version: !ruby/object:Gem::Version
-  version: 1.0.9
+  version: 1.1.0
 platform: ruby
 authors:
 - Kevin Sylvestre
@@ -76,6 +76,8 @@ files:
 - lib/omniai/chat/usage.rb
 - lib/omniai/client.rb
 - lib/omniai/config.rb
+- lib/omniai/transcribe.rb
+- lib/omniai/transcribe/transcription.rb
 - lib/omniai/version.rb
 homepage: https://github.com/ksylvest/omniai
 licenses: []