ruby : make Ruby bindings installed with build options (#3056)

* Fix signature of URI.new7s return value

* Use path instead of string | _ToPath

* Add document comment to RBS

* Remove unnecessary build flags

* Remove unnecessary line

* Remove files have become unnecessary

* Make gem install accept build options for whisper.cpp

* Add instraction for build options in README

* Add methods for check to Options

* Test build options

* Rename: configs -> options

* Add assert_installed assertion

* Use assert_installed

* Remove unused attribute

* Extract dependency check logic as Dependencies class

* Update README

* Add WHISPER_FFMPEG option

* Test extra build options only on local test

* Bump version to 1.3.2 [skip ci]

bindings/ruby/README.md

@@ -16,6 +16,18 @@ If bundler is not being used to manage dependencies, install the gem by executin
 
     $ gem install whispercpp
 
+You can pass build options for whisper.cpp, for instance:
+
+    $ bundle config build.whispercpp --enable-ggml-cuda
+
+or,
+
+    $ gem install whispercpp -- --enable-ggml-cuda
+
+See whisper.cpp's [README](https://github.com/ggml-org/whisper.cpp/blob/master/README.md) for available options. You need convert options present the README to Ruby-style options.  
+For boolean options like `GGML_CUDA`, the README says `-DGGML_CUDA=1`. You need strip `-D`, prepend `--enable-` for `1` or `ON` (`--disable-` for `0` or `OFF`) and make it kebab-case: `--enable-ggml-cuda`.  
+For options which require arguments like `CMAKE_CUDA_ARCHITECTURES`, the README says `-DCMAKE_CUDA_ARCHITECTURES="86"`. You need strip `-D`, prepend `--`, make it kebab-case, append `=` and append argument: `--cmake-cuda-architectures="86"`.
+
 Usage
 -----
 

bindings/ruby/ext/cpu.mk

@@ -1,13 +0,0 @@
-ggml/src/ggml-cpu/ggml-cpu-cpp.o: \
-	ggml/src/ggml-cpu/ggml-cpu.cpp \
-	ggml/src/ggml-cpu/unary-ops.cpp \
-	ggml/src/ggml-cpu/binary-ops.cpp \
-	ggml/src/ggml-cpu/vec.cpp \
-	ggml/src/ggml-cpu/ops.cpp \
-	ggml/include/ggml-backend.h \
-	ggml/include/ggml.h \
-	ggml/include/ggml-alloc.h \
-	ggml/src/ggml-backend-impl.h \
-	ggml/include/ggml-cpu.h \
-	ggml/src/ggml-impl.h
-	$(CXX) $(CXXFLAGS)   -c $< -o $@

bindings/ruby/ext/dependencies.rb

@@ -0,0 +1,61 @@
+require "tsort"
+
+class Dependencies
+  def initialize(cmake, options)
+    @cmake = cmake
+    @options = options
+
+    generate_dot
+    @libs = parse_dot
+  end
+
+  def to_s
+    @libs.join(" ")
+  end
+
+  private
+
+  def dot_path
+    File.join(__dir__, "build", "whisper.cpp.dot")
+  end
+
+  def generate_dot
+    system @cmake, "-S", "sources", "-B", "build", "--graphviz", dot_path, "-D", "BUILD_SHARED_LIBS=OFF", @options.to_s, exception: true
+  end
+
+  def parse_dot
+    static_lib_shape = nil
+    nodes = {}
+    depends = Hash.new {|h, k| h[k] = []}
+
+    class << depends
+      include TSort
+      alias tsort_each_node each_key
+      def tsort_each_child(node, &block)
+        fetch(node, []).each(&block)
+      end
+    end
+
+    File.open(dot_path).each_line do |line|
+      case line
+      when /\[\s*label\s*=\s*"Static Library"\s*,\s*shape\s*=\s*(?<shape>\w+)\s*\]/
+        static_lib_shape = $~[:shape]
+      when /\A\s*"(?<node>\w+)"\s*\[\s*label\s*=\s*"(?<label>\S+)"\s*,\s*shape\s*=\s*(?<shape>\w+)\s*\]\s*;\s*\z/
+        node = $~[:node]
+        label = $~[:label]
+        shape = $~[:shape]
+        nodes[node] = [label, shape]
+      when /\A\s*"(?<depender>\w+)"\s*->\s*"(?<dependee>\w+)"/
+        depender = $~[:depender]
+        dependee = $~[:dependee]
+        depends[depender] ||= []
+        depends[depender] << dependee
+      end
+    end
+    depends.tsort.filter_map {|node|
+      label, shape = nodes[node]
+      shape == static_lib_shape ? label : nil
+    }.collect {|lib| "lib#{lib}.a"}
+      .reverse
+  end
+end

bindings/ruby/ext/extconf.rb

@@ -1,50 +1,12 @@
 require "mkmf"
-require "tsort"
+require_relative "options"
-
+require_relative "dependencies"
-# TODO: options such as CoreML
 
 cmake = find_executable("cmake") || abort
-
+options = Options.new
 have_library("gomp") rescue nil
+libs = Dependencies.new(cmake, options)
 
-prefix = File.join("build", "whisper.cpp.dot")
-system cmake, "-S", "sources", "-B", "build", "--graphviz", prefix, "-D", "BUILD_SHARED_LIBS=OFF", exception: true
-
-static_lib_shape = nil
-nodes = {}
-depends = {}
-class << depends
-  include TSort
-  alias tsort_each_node each_key
-  def tsort_each_child(node, &block)
-    fetch(node, []).each(&block)
-  end
-end
-File.open(File.join("build", "whisper.cpp.dot")).each_line do |line|
-  case line
-  when /\[\s*label\s*=\s*"Static Library"\s*,\s*shape\s*=\s*(?<shape>\w+)\s*\]/
-    static_lib_shape = $~[:shape]
-  when /\A\s*"(?<node>\w+)"\s*\[\s*label\s*=\s*"(?<label>\S+)"\s*,\s*shape\s*=\s*(?<shape>\w+)\s*\]\s*;\s*\z/
-    node = $~[:node]
-    label = $~[:label]
-    shape = $~[:shape]
-    nodes[node] = [label, shape]
-  when /\A\s*"(?<depender>\w+)"\s*->\s*"(?<dependee>\w+)"/
-    depender = $~[:depender]
-    dependee = $~[:dependee]
-    depends[depender] ||= []
-    depends[depender] << dependee
-  end
-end
-libs = depends.tsort.filter_map {|node|
-  label, shape = nodes[node]
-  shape == static_lib_shape ? label : nil
-}.collect {|lib| "lib#{lib}.a"}
- .reverse
- .join(" ")
-
-$CFLAGS << " -std=c11 -fPIC"
-$CXXFLAGS << " -std=c++17 -O3 -DNDEBUG"
 $INCFLAGS << " -Isources/include -Isources/ggml/include -Isources/examples"
 $LOCAL_LIBS << " #{libs}"
 $cleanfiles << " build #{libs}"
@@ -54,8 +16,7 @@ create_makefile "whisper" do |conf|
     $(TARGET_SO): #{libs}
     #{libs}: cmake-targets
     cmake-targets:
-    #{"\t"}#{cmake} -S sources -B build -D BUILD_SHARED_LIBS=OFF -D CMAKE_ARCHIVE_OUTPUT_DIRECTORY=#{__dir__} -D CMAKE_POSITION_INDEPENDENT_CODE=ON
+    #{"\t"}#{cmake} -S sources -B build -D BUILD_SHARED_LIBS=OFF -D CMAKE_ARCHIVE_OUTPUT_DIRECTORY=#{__dir__} -D CMAKE_POSITION_INDEPENDENT_CODE=ON #{options}
     #{"\t"}#{cmake} --build build --config Release --target common whisper
-    #{"\t"}
   EOF
 end

bindings/ruby/ext/metal-embed.mk

@@ -1,17 +0,0 @@
-ggml/src/ggml-metal/ggml-metal-embed.o: \
-	ggml/src/ggml-metal/ggml-metal.metal \
-	ggml/src/ggml-metal/ggml-metal-impl.h \
-	ggml/src/ggml-common.h
-	@echo "Embedding Metal library"
-	@sed -e '/__embed_ggml-common.h__/r      ggml/src/ggml-common.h'                -e '/__embed_ggml-common.h__/d'      < ggml/src/ggml-metal/ggml-metal.metal           > ggml/src/ggml-metal/ggml-metal-embed.metal.tmp
-	@sed -e '/#include "ggml-metal-impl.h"/r ggml/src/ggml-metal/ggml-metal-impl.h' -e '/#include "ggml-metal-impl.h"/d' < ggml/src/ggml-metal/ggml-metal-embed.metal.tmp > ggml/src/ggml-metal/ggml-metal-embed.metal
-	$(eval TEMP_ASSEMBLY=$(shell mktemp -d))
-	@echo ".section __DATA, __ggml_metallib"                       >  $(TEMP_ASSEMBLY)/ggml-metal-embed.s
-	@echo ".globl _ggml_metallib_start"                            >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s
-	@echo "_ggml_metallib_start:"                                  >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s
-	@echo ".incbin \"ggml/src/ggml-metal/ggml-metal-embed.metal\"" >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s
-	@echo ".globl _ggml_metallib_end"                              >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s
-	@echo "_ggml_metallib_end:"                                    >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s
-	$(CC) $(CFLAGS) -c $(TEMP_ASSEMBLY)/ggml-metal-embed.s -o $@
-	@rm -f ${TEMP_ASSEMBLY}/ggml-metal-embed.s
-	@rmdir ${TEMP_ASSEMBLY}

bindings/ruby/ext/metal.mk

@@ -1,6 +0,0 @@
-ggml/src/ggml-metal/ggml-metal.o: \
-	ggml/src/ggml-metal/ggml-metal.m \
-	ggml/src/ggml-metal/ggml-metal-impl.h \
-	ggml/include/ggml-metal.h \
-	ggml/include/ggml.h
-	$(CC) $(CFLAGS) -c $< -o $@

bindings/ruby/ext/options.rb

@@ -0,0 +1,219 @@
+class Options
+  def initialize
+    @options = {}
+    @pending_options = []
+    @ignored_options = []
+
+    configure
+  end
+
+  def help
+    @options
+      .collect_concat {|name, (type, value)|
+        option = option_name(name)
+        if type == :bool
+          ["--enable-#{option}", "--disable-#{option}"]
+        else
+          "--#{option}=#{type.upcase}"
+        end
+      }
+      .join($/)
+  end
+
+  def to_s
+    @options
+      .reject {|name, (type, value)| value.nil?}
+      .collect {|name, (type, value)| "-D #{name}=#{value == true ? "ON" : value == false ? "OFF" : value.shellescape}"}
+      .join(" ")
+  end
+
+  def cmake_options
+    return @cmake_options if @cmake_options
+
+    output = nil
+    Dir.chdir __dir__ do
+      output = `cmake -S sources -B build -L`
+    end
+    started = false
+    @cmake_options = output.lines.filter_map {|line|
+      if line.chomp == "-- Cache values"
+        started = true
+        next
+      end
+      next unless started
+      option, value = line.chomp.split("=", 2)
+      name, type = option.split(":", 2)
+      [name, type, value]
+    }
+  end
+
+  def missing_options
+    cmake_options.collect {|name, type, value| name} -
+      @options.keys - @pending_options - @ignored_options
+  end
+
+  def extra_options
+    @options.keys + @pending_options - @ignored_options -
+      cmake_options.collect {|name, type, value| name}
+  end
+
+  private
+
+  def configure
+    filepath "ACCELERATE_FRAMEWORK"
+    ignored "BUILD_SHARED_LIBS"
+    ignored "BUILD_TESTING"
+    ignored "CMAKE_BUILD_TYPE"
+    ignored "CMAKE_INSTALL_PREFIX"
+    string "CMAKE_OSX_ARCHITECTURES"
+    ignored "CMAKE_OSX_DEPLOYMENT_TARGET"
+    string "CMAKE_OSX_SYSROOT"
+    filepath "FOUNDATION_LIBRARY"
+    bool "GGML_ACCELERATE"
+    bool "GGML_ALL_WARNINGS_3RD_PARTY"
+    bool "GGML_AMX_BF16"
+    bool "GGML_AMX_INT8"
+    bool "GGML_AMX_TILE"
+    bool "GGML_AVX"
+    bool "GGML_AVX2"
+    bool "GGML_AVX512"
+    bool "GGML_AVX512_BF16"
+    bool "GGML_AVX512_VBMI"
+    bool "GGML_AVX512_VNNI"
+    bool "GGML_AVX_VNNI"
+    ignored "GGML_BACKEND_DL"
+    ignored "GGML_BIN_INSTALL_DIR"
+    bool "GGML_BLAS"
+    string "GGML_BLAS_VENDOR"
+    bool "GGML_BMI2"
+    ignored "GGML_BUILD_EXAMPLES"
+    ignored "GGML_BUILD_TESTS"
+    filepath "GGML_CCACHE_FOUND"
+    bool "GGML_CPU"
+    bool "GGML_CPU_AARCH64"
+    ignored "GGML_CPU_ALL_VARIANTS"
+    string "GGML_CPU_ARM_ARCH"
+    bool "GGML_CPU_HBM"
+    bool "GGML_CPU_KLEIDIAI"
+    string "GGML_CPU_POWERPC_CPUTYPE"
+    bool "GGML_CUDA"
+    string "GGML_CUDA_COMPRESSION_MODE"
+    bool "GGML_CUDA_F16"
+    bool "GGML_CUDA_FA"
+    bool "GGML_CUDA_FA_ALL_QUANTS"
+    bool "GGML_CUDA_FORCE_CUBLAS"
+    bool "GGML_CUDA_FORCE_MMQ"
+    ignored "GGML_CUDA_GRAPHS"
+    bool "GGML_CUDA_NO_PEER_COPY"
+    bool "GGML_CUDA_NO_VMM"
+    string "GGML_CUDA_PEER_MAX_BATCH_SIZE"
+    bool "GGML_F16C"
+    bool "GGML_FMA"
+    bool "GGML_GPROF"
+    bool "GGML_HIP"
+    bool "GGML_HIP_GRAPHS"
+    bool "GGML_HIP_NO_VMM"
+    bool "GGML_HIP_ROCWMMA_FATTN"
+    bool "GGML_HIP_UMA"
+    ignored "GGML_INCLUDE_INSTALL_DIR"
+    bool "GGML_KOMPUTE"
+    bool "GGML_LASX"
+    ignored "GGML_LIB_INSTALL_DIR"
+    ignored "GGML_LLAMAFILE"
+    bool "GGML_LSX"
+    bool "GGML_LTO"
+    bool "GGML_METAL"
+    bool "GGML_METAL_EMBED_LIBRARY"
+    string "GGML_METAL_MACOSX_VERSION_MIN"
+    bool "GGML_METAL_NDEBUG"
+    bool "GGML_METAL_SHADER_DEBUG"
+    string "GGML_METAL_STD"
+    bool "GGML_METAL_USE_BF16"
+    bool "GGML_MUSA"
+    bool "GGML_NATIVE"
+    bool "GGML_OPENCL"
+    bool "GGML_OPENCL_EMBED_KERNELS"
+    bool "GGML_OPENCL_PROFILING"
+    string "GGML_OPENCL_TARGET_VERSION"
+    bool "GGML_OPENCL_USE_ADRENO_KERNELS"
+    bool "GGML_OPENMP"
+    bool "GGML_RPC"
+    bool "GGML_RVV"
+    bool "GGML_RV_ZFH"
+    pending "GGML_SCCACHE_FOUND"
+    string "GGML_SCHED_MAX_COPIES"
+    ignored "GGML_STATIC"
+    bool "GGML_SYCL"
+    string "GGML_SYCL_DEVICE_ARCH"
+    bool "GGML_SYCL_F16"
+    bool "GGML_SYCL_GRAPH"
+    string "GGML_SYCL_TARGET"
+    bool "GGML_VULKAN"
+    bool "GGML_VULKAN_CHECK_RESULTS"
+    bool "GGML_VULKAN_DEBUG"
+    bool "GGML_VULKAN_MEMORY_DEBUG"
+    bool "GGML_VULKAN_PERF"
+    ignored "GGML_VULKAN_RUN_TESTS"
+    filepath "GGML_VULKAN_SHADERS_GEN_TOOLCHAIN"
+    bool "GGML_VULKAN_SHADER_DEBUG_INFO"
+    pending "GGML_VULKAN_VALIDATE"
+    bool "GGML_VXE"
+    filepath "GIT_EXE"
+    filepath "MATH_LIBRARY"
+    filepath "METALKIT_FRAMEWORK"
+    filepath "METAL_FRAMEWORK"
+    bool "WHISPER_ALL_WARNINGS"
+    bool "WHISPER_ALL_WARNINGS_3RD_PARTY"
+    ignored "WHISPER_BIN_INSTALL_DIR"
+    ignored "WHISPER_BUILD_EXAMPLES"
+    ignored "WHISPER_BUILD_SERVER"
+    ignored"WHISPER_BUILD_TESTS"
+    bool "WHISPER_CCACHE"
+    bool "WHISPER_COREML"
+    bool "WHISPER_COREML_ALLOW_FALLBACK"
+    ignored "WHISPER_CURL"
+    bool "WHISPER_FATAL_WARNINGS"
+    ignored "WHISPER_FFMPEG"
+    ignored "WHISPER_INCLUDE_INSTALL_DIR"
+    ignored "WHISPER_LIB_INSTALL_DIR"
+    bool "WHISPER_OPENVINO"
+    bool "WHISPER_SANITIZE_ADDRESS"
+    bool "WHISPER_SANITIZE_THREAD"
+    bool "WHISPER_SANITIZE_UNDEFINED"
+    ignored "WHISPER_SDL2"
+    pending "WHISPER_USE_SYSTEM_GGML"
+  end
+
+  def option_name(name)
+    name.downcase.gsub("_", "-")
+  end
+
+  def bool(name)
+    option = option_name(name)
+    value = enable_config(option)
+    @options[name] = [:bool, value]
+  end
+
+  def string(name, type=:string)
+    option = "--#{option_name(name)}"
+    value = arg_config(option)
+    raise "String expected for #{option}" if value == true || value&.empty?
+    @options[name] = [type, value]
+  end
+
+  def path(name)
+    string(name, :path)
+  end
+
+  def filepath(name)
+    string(name, :filepath)
+  end
+
+  def pending(name)
+    @pending_options << name
+  end
+
+  def ignored(name)
+    @ignored_options << name
+  end
+end

bindings/ruby/sig/whisper.rbs

@@ -23,9 +23,20 @@ module Whisper
   def self.log_set: (log_callback, Object? user_data) -> log_callback
 
   class Context
-    def self.new: (string | _ToPath | ::URI::HTTP) -> instance
+    def self.new: (path | ::URI::HTTP) -> instance
+
+    # transcribe a single file
+    # can emit to a block results
+    #
+    #   params = Whisper::Params.new
+    #   params.duration = 60_000
+    #   whisper.transcribe "path/to/audio.wav", params do |text|
+    #     puts text
+    #   end
+    #
     def transcribe: (string, Params) -> self
                   | (string, Params) { (String) -> void } -> self
+
     def model_n_vocab: () -> Integer
     def model_n_audio_ctx: () -> Integer
     def model_n_audio_state: () -> Integer
@@ -34,19 +45,72 @@ module Whisper
     def model_n_mels: () -> Integer
     def model_ftype: () -> Integer
     def model_type: () -> String
+
+    # Yields each Whisper::Segment:
+    #
+    #   whisper.transcribe("path/to/audio.wav", params)
+    #   whisper.each_segment do |segment|
+    #     puts segment.text
+    #   end
+    #
+    # Returns an Enumerator if no block given:
+    #
+    #   whisper.transcribe("path/to/audio.wav", params)
+    #   enum = whisper.each_segment
+    #   enum.to_a # => [#<Whisper::Segment>, ...]
+    #
     def each_segment: { (Segment) -> void } -> void
                     | () -> Enumerator[Segment]
+
     def model: () -> Model
     def full_get_segment: (Integer nth) -> Segment
     def full_n_segments: () -> Integer
+
+    # Language ID, which can be converted to string by Whisper.lang_str and Whisper.lang_str_full.
+    #
     def full_lang_id: () -> Integer
+
+    # Start time of a segment indexed by +segment_index+ in centiseconds (10 times milliseconds).
+    #
+    #   full_get_segment_t0(3) # => 1668 (16680 ms)
+    #
     def full_get_segment_t0: (Integer) -> Integer
+
+    # End time of a segment indexed by +segment_index+ in centiseconds (10 times milliseconds).
+    #
+    #   full_get_segment_t1(3) # => 1668 (16680 ms)
+    #
     def full_get_segment_t1: (Integer) -> Integer
+
+    # Whether the next segment indexed by +segment_index+ is predicated as a speaker turn.
+    #
+    #   full_get_segment_speacker_turn_next(3) # => true
+    #
     def full_get_segment_speaker_turn_next: (Integer) -> (true | false)
+
+    # Text of a segment indexed by +segment_index+.
+    #
+    #   full_get_segment_text(3) # => "ask not what your country can do for you, ..."
+    #
     def full_get_segment_text: (Integer) -> String
+
     def full_get_segment_no_speech_prob: (Integer) -> Float
+
+    # Run the entire model: PCM -> log mel spectrogram -> encoder -> decoder -> text
+    # Not thread safe for same context
+    # Uses the specified decoding strategy to obtain the text.
+    #
+    # The second argument +samples+ must be an array of samples, respond to :length, or be a MemoryView of an array of float. It must be 32 bit float PCM audio data.
+    #
     def full: (Params, Array[Float] samples, ?Integer n_samples) -> self
             | (Params, _Samples, ?Integer n_samples) -> self
+
+    # Split the input audio in chunks and process each chunk separately using whisper_full_with_state()
+    # Result is stored in the default state of the context
+    # Not thread safe if executed in parallel on the same context.
+    # It seems this approach can offer some speedup in some cases.
+    # However, the transcription accuracy can be worse at the beginning and end of each chunk.
+    #
     def full_parallel: (Params, Array[Float], ?Integer n_samples) -> self
                      | (Params, _Samples, ?Integer n_samples) -> self
                      | (Params, _Samples, ?Integer? n_samples, Integer n_processors) -> self
@@ -85,68 +149,202 @@ module Whisper
       ?abort_callback: abort_callback,
       ?abort_callback_user_data: Object
     ) -> instance
+
+    # params.language = "auto" | "en", etc...
+    #
     def language=: (String) -> String # TODO: Enumerate lang names
+
     def language: () -> String
     def translate=: (boolish) -> boolish
     def translate: () -> (true | false)
     def no_context=: (boolish) -> boolish
+
+    # If true, does not use past transcription (if any) as initial prompt for the decoder.
+    #
     def no_context: () -> (true | false)
+
     def single_segment=: (boolish) -> boolish
+
+    # If true, forces single segment output (useful for streaming).
+    #
     def single_segment: () -> (true | false)
+
     def print_special=: (boolish) -> boolish
+
+    # If true, prints special tokens (e.g. <SOT>, <EOT>, <BEG>, etc.).
+    #
     def print_special: () -> (true | false)
+
     def print_progress=: (boolish) -> boolish
+
+    # If true, prints progress information.
+    #
     def print_progress: () -> (true | false)
+
     def print_realtime=: (boolish) -> boolish
+
+    # If true, prints results from within whisper.cpp. (avoid it, use callback instead)
+    #
     def print_realtime: () -> (true | false)
+
+    # If true, prints timestamps for each text segment when printing realtime.
+    #
     def print_timestamps=: (boolish) -> boolish
+
     def print_timestamps: () -> (true | false)
+
     def suppress_blank=: (boolish) -> boolish
+
+    # If true, suppresses blank outputs.
+    #
     def suppress_blank: () -> (true | false)
+
     def suppress_nst=: (boolish) -> boolish
+
+    # If true, suppresses non-speech-tokens.
+    #
     def suppress_nst: () -> (true | false)
+
     def token_timestamps=: (boolish) -> boolish
+
+    # If true, enables token-level timestamps.
+    #
     def token_timestamps: () -> (true | false)
+
     def split_on_word=: (boolish) -> boolish
+
+    # If true, split on word rather than on token (when used with max_len).
+    #
     def split_on_word: () -> (true | false)
+
     def initial_prompt=: (_ToS) -> _ToS
+
+    # Tokens to provide to the whisper decoder as initial prompt
+    # these are prepended to any existing text context from a previous call
+    # use whisper_tokenize() to convert text to tokens.
+    # Maximum of whisper_n_text_ctx()/2 tokens are used (typically 224).
+    #
     def initial_prompt: () -> (String | nil)
+
     def diarize=: (boolish) -> boolish
+
+    # If true, enables diarization.
+    #
     def diarize: () -> (true | false)
+
     def offset=: (Integer) -> Integer
+
+    # Start offset in ms.
+    #
     def offset: () -> Integer
+
     def duration=: (Integer) -> Integer
+
+    # Audio duration to process in ms.
+    #
     def duration: () -> Integer
+
     def max_text_tokens=: (Integer) -> Integer
+
+    # Max tokens to use from past text as prompt for the decoder.
+    #
     def max_text_tokens: () -> Integer
+
     def temperature=: (Float) -> Float
     def temperature: () -> Float
     def max_initial_ts=: (Float) -> Float
+
+    # See https://github.com/openai/whisper/blob/f82bc59f5ea234d4b97fb2860842ed38519f7e65/whisper/decoding.py#L97
+    #
     def max_initial_ts: () -> Float
+
     def length_penalty=: (Float) -> Float
     def length_penalty: () -> Float
     def temperature_inc=: (Float) -> Float
     def temperature_inc: () -> Float
     def entropy_thold=: (Float) -> Float
+
+    # Similar to OpenAI's "compression_ratio_threshold"
+    #
     def entropy_thold: () -> Float
+
     def logprob_thold=: (Float) -> Float
     def logprob_thold: () -> Float
     def no_speech_thold=: (Float) -> Float
     def no_speech_thold: () -> Float
+
+    # Sets new segment callback, called for every newly generated text segment.
+    #
+    #   params.new_segment_callback = ->(context, _, n_new, user_data) {
+    #     # ...
+    #   }
+    #
     def new_segment_callback=: (new_segment_callback) -> new_segment_callback
     def new_segment_callback: () -> (new_segment_callback | nil)
+
+    # Sets user data passed to the last argument of new segment callback.
+    #
     def new_segment_callback_user_data=: (Object) -> Object
+
     def new_segment_callback_user_data: () -> Object
+
+    # Sets progress callback, called on each progress update.
+    #
+    #   params.new_segment_callback = ->(context, _, progress, user_data) {
+    #     # ...
+    #   }
+    #
+    # +progress+ is an Integer between 0 and 100.
+    #
     def progress_callback=: (progress_callback) -> progress_callback
+
     def progress_callback: () -> (progress_callback | nil)
+
+    # Sets user data passed to the last argument of progress callback.
+    #
     def progress_callback_user_data=: (Object) -> Object
+
     def progress_callback_user_data: () -> Object
+
+    # Sets abort callback, called to check if the process should be aborted.
+    #
+    #   params.abort_callback = ->(user_data) {
+    #     # ...
+    #   }
+    #
+    #
     def abort_callback=: (abort_callback) -> abort_callback
+
     def abort_callback: () -> (abort_callback | nil)
+
+    # Sets user data passed to the last argument of abort callback.
+    #
     def abort_callback_user_data=: (Object) -> Object
+
     def abort_callback_user_data: () -> Object
+
+    # Hook called on new segment. Yields each Whisper::Segment.
+    #
+    #   whisper.on_new_segment do |segment|
+    #     # ...
+    #   end
+    #
     def on_new_segment: { (Segment) -> void } -> void
+
+    # Hook called on progress update. Yields each progress Integer between 0 and 100.
+    #
     def on_progress: { (Integer progress) -> void } -> void
+
+    # Call block to determine whether abort or not. Return +true+ when you want to abort.
+    #
+    #   params.abort_on do
+    #     if some_condition
+    #       true # abort
+    #     else
+    #       false # continue
+    #     end
+    #   end
+    #
     def abort_on: { (Object user_data) -> boolish } -> void
   end
 
@@ -167,16 +365,24 @@ module Whisper
     def type: () -> String
 
     class URI
-      def self.new: (string | ::URI::HTTP) -> self
+      def self.new: (string | ::URI::HTTP) -> instance
       def to_path: -> String
       def clear_cache: -> void
     end
   end
 
   class Segment
+    # Start time in milliseconds.
+    #
     def start_time: () -> Integer
+
+    # End time in milliseconds.
+    #
     def end_time: () -> Integer
+
+    # Whether the next segment is predicted as a speaker turn.
     def speaker_next_turn?: () -> (true | false)
+
     def text: () -> String
     def no_speech_prob: () -> Float
   end

bindings/ruby/tests/helper.rb

@@ -21,4 +21,15 @@ class TestBase < Test::Unit::TestCase
   def whisper
     self.class.whisper
   end
+
+  module BuildOptions
+    load "ext/options.rb", self
+    Options.include self
+
+    def enable_config(name)
+    end
+
+    def arg_config(name)
+    end
+  end
 end

bindings/ruby/tests/test_package.rb

@@ -21,13 +21,26 @@ class TestPackage < TestBase
       match_data = `rake -Tbuild`.match(/(whispercpp-(.+)\.gem)/)
       filename = match_data[1]
       version = match_data[2]
-      basename = "whisper.#{RbConfig::CONFIG["DLEXT"]}"
       Dir.mktmpdir do |dir|
         system "gem", "install", "--install-dir", dir.shellescape, "--no-document", "pkg/#{filename.shellescape}", exception: true
-        assert_path_exist File.join(dir, "gems/whispercpp-#{version}/lib", basename)
+        assert_installed dir, version
-        assert_path_exist File.join(dir, "gems/whispercpp-#{version}/LICENSE")
-        assert_path_not_exist File.join(dir, "gems/whispercpp-#{version}/ext/build")
       end
     end
+
+    private
+
+    def assert_installed(dir, version)
+      assert_path_exist File.join(dir, "gems/whispercpp-#{version}/lib", "whisper.#{RbConfig::CONFIG["DLEXT"]}")
+      assert_path_exist File.join(dir, "gems/whispercpp-#{version}/LICENSE")
+      assert_path_not_exist File.join(dir, "gems/whispercpp-#{version}/ext/build")
+    end
+  end
+
+  def test_build_options
+    options = BuildOptions::Options.new
+    assert_empty options.missing_options
+    unless ENV["CI"]
+      assert_empty options.extra_options
+    end
   end
 end

bindings/ruby/whispercpp.gemspec

@@ -3,8 +3,8 @@ require_relative "extsources"
 Gem::Specification.new do |s|
   s.name    = "whispercpp"
   s.authors = ["Georgi Gerganov", "Todd A. Fisher"]
-  s.version = '1.3.1'
+  s.version = '1.3.2'
-  s.date    = '2024-12-19'
+  s.date    = '2025-04-17'
   s.description = %q{High-performance inference of OpenAI's Whisper automatic speech recognition (ASR) model via Ruby}
   s.email   = 'todd.fisher@gmail.com'
   s.extra_rdoc_files = ['LICENSE', 'README.md']