whispercpp 1.3.1 → 1.3.2

data/ext/options.rb

@@ -0,0 +1,221 @@
+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"
+    bool "GGML_CCACHE"
+    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"
+    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"
+    bool "GGML_SSE42"
+    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_SYCL_DNN"
+    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"
+    bool "GGML_XTHEADVECTOR"
+    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_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

data/ext/ruby_whisper.c

@@ -0,0 +1,159 @@
+#include <ruby.h>
+#include <ruby/memory_view.h>
+#include "ruby_whisper.h"
+
+VALUE mWhisper;
+VALUE mVAD;
+VALUE cContext;
+VALUE cParams;
+VALUE cVADParams;
+VALUE eError;
+
+VALUE cSegment;
+VALUE cModel;
+
+ID id_to_s;
+ID id_call;
+ID id___method__;
+ID id_to_enum;
+ID id_length;
+ID id_next;
+ID id_new;
+ID id_to_path;
+ID id_URI;
+ID id_pre_converted_models;
+
+static bool is_log_callback_finalized = false;
+
+// High level API
+extern VALUE ruby_whisper_segment_allocate(VALUE klass);
+
+extern void init_ruby_whisper_context(VALUE *mWhisper);
+extern void init_ruby_whisper_params(VALUE *mWhisper);
+extern void init_ruby_whisper_error(VALUE *mWhisper);
+extern void init_ruby_whisper_segment(VALUE *mWhisper, VALUE *cSegment);
+extern void init_ruby_whisper_model(VALUE *mWhisper);
+extern void init_ruby_whisper_vad_params(VALUE *mVAD);
+extern void register_callbacks(ruby_whisper_params *rwp, VALUE *context);
+
+/*
+ * call-seq:
+ *   lang_max_id -> Integer
+ */
+static VALUE ruby_whisper_s_lang_max_id(VALUE self) {
+  return INT2NUM(whisper_lang_max_id());
+}
+
+/*
+ * call-seq:
+ *   lang_id(lang_name) -> Integer
+ */
+static VALUE ruby_whisper_s_lang_id(VALUE self, VALUE lang) {
+  const char * lang_str = StringValueCStr(lang);
+  const int id = whisper_lang_id(lang_str);
+  if (-1 == id) {
+    rb_raise(rb_eArgError, "language not found: %s", lang_str);
+  }
+  return INT2NUM(id);
+}
+
+/*
+ * call-seq:
+ *   lang_str(lang_id) -> String
+ */
+static VALUE ruby_whisper_s_lang_str(VALUE self, VALUE id) {
+  const int lang_id = NUM2INT(id);
+  const char * str = whisper_lang_str(lang_id);
+  if (NULL == str) {
+    rb_raise(rb_eIndexError, "id %d outside of language id", lang_id);
+  }
+  return rb_str_new2(str);
+}
+
+/*
+ * call-seq:
+ *   lang_str(lang_id) -> String
+ */
+static VALUE ruby_whisper_s_lang_str_full(VALUE self, VALUE id) {
+  const int lang_id = NUM2INT(id);
+  const char * str_full = whisper_lang_str_full(lang_id);
+  if (NULL == str_full) {
+    rb_raise(rb_eIndexError, "id %d outside of language id", lang_id);
+  }
+  return rb_str_new2(str_full);
+}
+
+static VALUE ruby_whisper_s_finalize_log_callback(VALUE self, VALUE id) {
+  is_log_callback_finalized = true;
+  return Qnil;
+}
+
+static void
+ruby_whisper_log_callback(enum ggml_log_level level, const char * buffer, void * user_data) {
+  if (is_log_callback_finalized) {
+    return;
+  }
+  VALUE log_callback = rb_iv_get(mWhisper, "log_callback");
+  VALUE udata = rb_iv_get(mWhisper, "user_data");
+  rb_funcall(log_callback, id_call, 3, INT2NUM(level), rb_str_new2(buffer), udata);
+}
+
+/*
+ * call-seq:
+ *   log_set ->(level, buffer, user_data) { ... }, user_data -> nil
+ */
+static VALUE ruby_whisper_s_log_set(VALUE self, VALUE log_callback, VALUE user_data) {
+  VALUE old_callback = rb_iv_get(self, "log_callback");
+  if (!NIL_P(old_callback)) {
+    rb_undefine_finalizer(old_callback);
+  }
+
+  rb_iv_set(self, "log_callback", log_callback);
+  rb_iv_set(self, "user_data", user_data);
+
+  VALUE finalize_log_callback = rb_funcall(mWhisper, rb_intern("method"), 1, rb_str_new2("finalize_log_callback"));
+  rb_define_finalizer(log_callback, finalize_log_callback);
+
+  whisper_log_set(ruby_whisper_log_callback, NULL);
+
+  return Qnil;
+}
+
+void Init_whisper() {
+  id_to_s = rb_intern("to_s");
+  id_call = rb_intern("call");
+  id___method__ = rb_intern("__method__");
+  id_to_enum = rb_intern("to_enum");
+  id_length = rb_intern("length");
+  id_next = rb_intern("next");
+  id_new = rb_intern("new");
+  id_to_path = rb_intern("to_path");
+  id_URI = rb_intern("URI");
+  id_pre_converted_models = rb_intern("pre_converted_models");
+
+  mWhisper = rb_define_module("Whisper");
+  mVAD = rb_define_module_under(mWhisper, "VAD");
+
+  rb_define_const(mWhisper, "LOG_LEVEL_NONE", INT2NUM(GGML_LOG_LEVEL_NONE));
+  rb_define_const(mWhisper, "LOG_LEVEL_INFO", INT2NUM(GGML_LOG_LEVEL_INFO));
+  rb_define_const(mWhisper, "LOG_LEVEL_WARN", INT2NUM(GGML_LOG_LEVEL_WARN));
+  rb_define_const(mWhisper, "LOG_LEVEL_ERROR", INT2NUM(GGML_LOG_LEVEL_ERROR));
+  rb_define_const(mWhisper, "LOG_LEVEL_DEBUG", INT2NUM(GGML_LOG_LEVEL_DEBUG));
+  rb_define_const(mWhisper, "LOG_LEVEL_CONT", INT2NUM(GGML_LOG_LEVEL_CONT));
+
+  rb_define_singleton_method(mWhisper, "lang_max_id", ruby_whisper_s_lang_max_id, 0);
+  rb_define_singleton_method(mWhisper, "lang_id", ruby_whisper_s_lang_id, 1);
+  rb_define_singleton_method(mWhisper, "lang_str", ruby_whisper_s_lang_str, 1);
+  rb_define_singleton_method(mWhisper, "lang_str_full", ruby_whisper_s_lang_str_full, 1);
+  rb_define_singleton_method(mWhisper, "log_set", ruby_whisper_s_log_set, 2);
+  rb_define_private_method(rb_singleton_class(mWhisper), "finalize_log_callback", ruby_whisper_s_finalize_log_callback, 1);
+
+  init_ruby_whisper_context(&mWhisper);
+  init_ruby_whisper_params(&mWhisper);
+  init_ruby_whisper_error(&mWhisper);
+  init_ruby_whisper_segment(&mWhisper, &cContext);
+  init_ruby_whisper_model(&mWhisper);
+  init_ruby_whisper_vad_params(&mVAD);
+
+  rb_require("whisper/model/uri");
+}

data/ext/ruby_whisper.h

@@ -1,5 +1,5 @@
-#ifndef __RUBY_WHISPER_H
-#define __RUBY_WHISPER_H
+#ifndef RUBY_WHISPER_H
+#define RUBY_WHISPER_H
 
 #include "whisper.h"
 
@@ -19,7 +19,22 @@ typedef struct {
   bool diarize;
   ruby_whisper_callback_container *new_segment_callback_container;
   ruby_whisper_callback_container *progress_callback_container;
+  ruby_whisper_callback_container *encoder_begin_callback_container;
   ruby_whisper_callback_container *abort_callback_container;
+  VALUE vad_params;
 } ruby_whisper_params;
 
+typedef struct {
+  struct whisper_vad_params params;
+} ruby_whisper_vad_params;
+
+typedef struct {
+  VALUE context;
+  int index;
+} ruby_whisper_segment;
+
+typedef struct {
+  VALUE context;
+} ruby_whisper_model;
+
 #endif