tomoto 0.3.1 → 0.3.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: a7ebf68dfdd1842174f9f8c3b8459de5cb9943f194c0df0e75bad6b2747d7546
-  data.tar.gz: 97ed983d26beb5e172dbb42316b67256930032f017e9efabc27d0dbcf419959b
+  metadata.gz: daa9c454c4cf09e120dbbe17305d225be58ac5937c463886e87ea1b3c3b5d466
+  data.tar.gz: f6c0c353a0efcc6026964e9125f1156b50e0d119506ecab2812522f7b716042d
 SHA512:
-  metadata.gz: 5173e759d7091f2621f70b8e45906585a33bdaf3093b4ed9e6ce637e172773e1d0cc7d8233156ec3715707b33de221b42f643582a75aa2e2506bb7de49bb1e11
-  data.tar.gz: 0bd48bfe76e6ce72193c9b1b7e2f50967dd76c3125aeb0cbade7dc91979017e5a01d5879c205998ac8af87ef547b2604743d2571190b24730646d847910c1200
+  metadata.gz: 874f531a75a62d2291793ded080f380f8103682c2ae2b087dd31a014533443d5f35a7ea4e634aabd246fab1564ece35679c60c0b6ffbb6a627d57048e32bf790
+  data.tar.gz: e91bf3c618394f34f208fe4945729db6719a2cab1a8e7192e646b232d7e38274f2087808d145dccddb94b3cc632b9dddce82f942f051a4bf522ec5ec9d2c43b3

data/CHANGELOG.md

@@ -1,3 +1,13 @@
+## 0.3.3 (2023-02-01)
+
+- Added `topic_label_dict` method to `LLDA`
+- Fixed error with `infer` with loaded model
+
+## 0.3.2 (2023-01-22)
+
+- Added precompiled gem for Mac ARM
+- Updated tomoto to 0.12.4
+
 ## 0.3.1 (2023-01-12)
 
 - Added support for Ruby 3.2

data/LICENSE.txt

@@ -1,7 +1,7 @@
 MIT License
 
 Copyright (c) 2019, bab2min
-Copyright (c) 2020-2021 Andrew Kane
+Copyright (c) 2020-2023 Andrew Kane
 
 Permission is hereby granted, free of charge, to any person obtaining a copy
 of this software and associated documentation files (the "Software"), to deal

data/README.md

@@ -12,17 +12,15 @@ Add this line to your application’s Gemfile:
 gem "tomoto"
 ```
 
-ARM is not currently supported
-
 ## Getting Started
 
 Train a model
 
 ```ruby
 model = Tomoto::LDA.new(k: 2)
-model.add_doc("text from document one")
-model.add_doc("text from document two")
-model.add_doc("text from document three")
+model.add_doc(["tokens", "from", "document", "one"])
+model.add_doc(["tokens", "from", "document", "two"])
+model.add_doc(["tokens", "from", "document", "three"])
 model.train(100) # iterations
 ```
 
@@ -78,7 +76,7 @@ model.ll_per_word
 Perform inference for unseen documents
 
 ```ruby
-doc = model.make_doc("unseen doc")
+doc = model.make_doc(["unseen", "doc"])
 topic_dist, ll = model.infer(doc)
 ```
 
@@ -114,14 +112,6 @@ If a method or option you need isn’t supported, feel free to open an issue.
 - [LDA](examples/lda_basic.rb)
 - [HDP](examples/hdp_basic.rb)
 
-## Tokenization
-
-Documents are tokenized by whitespace by default, or you can perform your own tokenization.
-
-```ruby
-model.add_doc(["tokens", "from", "document", "one"])
-```
-
 ## Performance
 
 tomoto uses AVX2, AVX, or SSE2 instructions to increase performance on machines that support it. Check which instruction set architecture it’s using with:

data/ext/tomoto/extconf.rb

@@ -3,12 +3,17 @@ require "mkmf-rice"
 $CXXFLAGS += " -std=c++17 $(optflags) -DEIGEN_MPL2_ONLY"
 
 unless ENV["RUBY_CC_VERSION"]
-  # ARM not supported yet
-  # https://github.com/bab2min/tomotopy/issues/170
-
-  # AVX-512F not supported yet
-  # https://github.com/bab2min/tomotopy/issues/188
-  $CXXFLAGS << " " << with_config("optflags", "-march=native -mno-avx512f")
+  default_optflags =
+    if RbConfig::CONFIG["host_os"] =~ /darwin/i && RbConfig::CONFIG["host_cpu"] =~ /arm|aarch64/i
+      # -march=native not supported with Mac ARM
+      ""
+    else
+      # AVX-512F not supported yet
+      # https://github.com/bab2min/tomotopy/issues/188
+      "-march=native -mno-avx512f"
+    end
+
+  $CXXFLAGS << " " << with_config("optflags", default_optflags)
 end
 
 apple_clang = RbConfig::CONFIG["CC_VERSION_MESSAGE"] =~ /apple clang/i

data/ext/tomoto/llda.cpp

@@ -29,5 +29,18 @@ void init_llda(Rice::Module& m) {
       "topics_per_label",
       [](tomoto::ILLDAModel& self) {
         return self.getNumTopicsPerLabel();
+      })
+    .define_method(
+      "topic_label_dict",
+      [](tomoto::ILLDAModel& self) {
+        auto dict = self.getTopicLabelDict();
+        Array res;
+        auto utf8 = Rice::Class(rb_cEncoding).call("const_get", "UTF_8");
+        for (size_t i = 0; i < dict.size(); i++) {
+          VALUE value = Rice::detail::To_Ruby<std::string>().convert(dict.toWord(i));
+          Object obj(value);
+          res.push(obj.call("force_encoding", utf8));
+        }
+        return res;
       });
 }

data/lib/tomoto/lda.rb

@@ -24,7 +24,7 @@ module Tomoto
 
     # TODO support multiple docs
     def infer(doc, iter: 100, tolerance: -1, workers: 0, parallel: :default, together: 0)
-      raise "cannot infer with untrained model" unless defined?(@prepared)
+      raise "cannot infer with untrained model" unless trained?
       _infer(doc, iter, tolerance, workers, to_ps(parallel), together)
     end
 
@@ -86,6 +86,7 @@ module Tomoto
       end
     end
 
+    # TODO raise error if iterations < 1
     def train(iterations = 10, workers: 0, parallel: :default)
       prepare
       _train(iterations, workers, to_ps(parallel))
@@ -97,6 +98,10 @@ module Tomoto
 
     private
 
+    def trained?
+      global_step.positive?
+    end
+
     def prepare
       unless defined?(@prepared)
         _prepare(@min_cf, @min_df, @rm_top)

data/lib/tomoto/version.rb

@@ -1,3 +1,3 @@
 module Tomoto
-  VERSION = "0.3.1"
+  VERSION = "0.3.3"
 end

data/vendor/EigenRand/EigenRand/Core.h

@@ -2,8 +2,8 @@
  * @file Core.h
  * @author bab2min (bab2min@gmail.com)
  * @brief 
- * @version 0.3.3
- * @date 2021-03-31
+ * @version 0.4.1
+ * @date 2022-08-13
  * 
  * @copyright Copyright (c) 2020-2021
  * 

data/vendor/EigenRand/EigenRand/Dists/Basic.h

@@ -2,8 +2,8 @@
  * @file Basic.h
  * @author bab2min (bab2min@gmail.com)
  * @brief 
- * @version 0.3.3
- * @date 2021-03-31
+ * @version 0.4.1
+ * @date 2022-08-13
  * 
  * @copyright Copyright (c) 2020-2021
  * 
@@ -360,6 +360,31 @@ namespace Eigen
 			}
 		};
 
+		namespace detail
+		{
+			template<size_t v>
+			struct BitWidth
+			{
+				static constexpr size_t value = BitWidth<v / 2>::value + 1;
+			};
+
+			template<>
+			struct BitWidth<0>
+			{
+				static constexpr size_t value = 0;
+			};
+
+			template<class Rng>
+			struct RngBitSize
+			{
+				static constexpr size_t _min = Rng::min();
+				static constexpr size_t _max = Rng::max();
+
+				static constexpr bool _fullbit_rng = _min == 0 && (_max & (_max + 1)) == 0;
+				static constexpr size_t value = IsPacketRandomEngine<Rng>::value ? sizeof(typename Rng::result_type) * 8 : (_fullbit_rng ? BitWidth<_max>::value : 0);
+			};
+		}
+
 		/**
 		 * @brief Generator of reals in a range `[0, 1)`
 		 * 
@@ -373,20 +398,57 @@ namespace Eigen
 		public:
 			using Scalar = _Scalar;
 
-			template<typename Rng>
+			template<typename Rng, 
+				typename std::enable_if<sizeof(Scalar) * 8 <= detail::RngBitSize<typename std::remove_const<typename std::remove_reference<Rng>::type>::type>::value, int>::type = 0
+			>
 			EIGEN_STRONG_INLINE const _Scalar operator() (Rng&& rng)
 			{
 				using namespace Eigen::internal;
 				return BitScalar<_Scalar>{}.to_ur(ExtractFirstUint<_Scalar>{}(std::forward<Rng>(rng)()));
 			}
 
-			template<typename Rng>
+			template<typename Rng,
+				typename std::enable_if<detail::RngBitSize<typename std::remove_const<typename std::remove_reference<Rng>::type>::type>::value < sizeof(Scalar) * 8, int>::type = 0
+			>
+			EIGEN_STRONG_INLINE const _Scalar operator() (Rng&& rng)
+			{
+				using RRng = typename std::remove_const<typename std::remove_reference<Rng>::type>::type;
+				static_assert(detail::RngBitSize<RRng>::value > 0,
+					"BaseRng must be a kind of mersenne_twister_engine.");
+				using ResultType = typename std::conditional<detail::RngBitSize<RRng>::value == 32, uint32_t, uint64_t>::type;
+				using namespace Eigen::internal;
+				ResultType arr[sizeof(Scalar) / sizeof(ResultType)];
+				for (size_t i = 0; i < sizeof(Scalar) / sizeof(ResultType); ++i)
+				{
+					arr[i] = rng();
+				}
+				return BitScalar<_Scalar>{}.to_ur(*(uint64_t*)arr);
+			}
+
+			template<typename Rng,
+				typename std::enable_if<sizeof(Scalar) <= sizeof(typename std::remove_const<typename std::remove_reference<Rng>::type>::type::result_type), int>::type = 0
+			>
 			EIGEN_STRONG_INLINE const _Scalar nzur_scalar(Rng&& rng)
 			{
 				using namespace Eigen::internal;
 				return BitScalar<_Scalar>{}.to_nzur(ExtractFirstUint<_Scalar>{}(std::forward<Rng>(rng)()));
 			}
 
+			template<typename Rng,
+				typename std::enable_if<sizeof(typename std::remove_const<typename std::remove_reference<Rng>::type>::type::result_type) < sizeof(Scalar), int > ::type = 0
+			>
+			EIGEN_STRONG_INLINE const _Scalar nzur_scalar(Rng&& rng)
+			{
+				using namespace Eigen::internal;
+				using RngResult = typename std::remove_const<typename std::remove_reference<Rng>::type>::type::result_type;
+				RngResult arr[sizeof(Scalar) / sizeof(RngResult)];
+				for (size_t i = 0; i < sizeof(Scalar) / sizeof(RngResult); ++i)
+				{
+					arr[i] = rng();
+				}
+				return BitScalar<_Scalar>{}.to_nzur(*(Scalar*)arr);
+			}
+
 			template<typename Packet, typename Rng>
 			EIGEN_STRONG_INLINE const Packet packetOp(Rng&& rng)
 			{

data/vendor/EigenRand/EigenRand/Dists/Discrete.h

@@ -2,8 +2,8 @@
  * @file Discrete.h
  * @author bab2min (bab2min@gmail.com)
  * @brief 
- * @version 0.3.3
- * @date 2021-03-31
+ * @version 0.4.1
+ * @date 2022-08-13
  *
  * @copyright Copyright (c) 2020-2021
  * 
@@ -264,7 +264,7 @@ namespace Eigen
 					{
 						EIGENRAND_CHECK_INFINITY_LOOP();
 						_Scalar cands = (_Scalar)(rx & bitmask);
-						if (cands <= pdiff) return cands;
+						if (cands <= pdiff) return cands + pmin;
 						if (bitcnt + bitsize < 32)
 						{
 							rx >>= bitsize;

data/vendor/EigenRand/EigenRand/Dists/GammaPoisson.h

@@ -2,8 +2,8 @@
  * @file GammaPoisson.h
  * @author bab2min (bab2min@gmail.com)
  * @brief
- * @version 0.3.3
- * @date 2021-03-31
+ * @version 0.4.1
+ * @date 2022-08-13
  *
  * @copyright Copyright (c) 2020-2021
  *

data/vendor/EigenRand/EigenRand/Dists/NormalExp.h

@@ -2,8 +2,8 @@
  * @file NormalExp.h
  * @author bab2min (bab2min@gmail.com)
  * @brief 
- * @version 0.3.3
- * @date 2021-03-31
+ * @version 0.4.1
+ * @date 2022-08-13
  * 
  * @copyright Copyright (c) 2020-2021
  * 

data/vendor/EigenRand/EigenRand/EigenRand

@@ -2,8 +2,8 @@
  * @file EigenRand
  * @author bab2min (bab2min@gmail.com)
  * @brief 
- * @version 0.4.0
- * @date 2021-09-17
+ * @version 0.4.1
+ * @date 2022-08-13
  * 
  * @copyright Copyright (c) 2020-2021
  * 

data/vendor/EigenRand/EigenRand/Macro.h

@@ -2,8 +2,8 @@
  * @file Macro.h
  * @author bab2min (bab2min@gmail.com)
  * @brief 
- * @version 0.4.0
- * @date 2021-09-28
+ * @version 0.4.1
+ * @date 2022-08-13
  * 
  * @copyright Copyright (c) 2020-2021
  * 
@@ -14,7 +14,7 @@
 
 #define EIGENRAND_WORLD_VERSION 0
 #define EIGENRAND_MAJOR_VERSION 4
-#define EIGENRAND_MINOR_VERSION 0
+#define EIGENRAND_MINOR_VERSION 1
 
 #if EIGEN_VERSION_AT_LEAST(3,3,10)
 #define EIGENRAND_EIGEN_34_MODE

data/vendor/EigenRand/EigenRand/MorePacketMath.h

@@ -2,8 +2,8 @@
  * @file MorePacketMath.h
  * @author bab2min (bab2min@gmail.com)
  * @brief 
- * @version 0.3.3
- * @date 2021-03-31
+ * @version 0.4.1
+ * @date 2022-08-13
  * 
  * @copyright Copyright (c) 2020-2021
  * 

data/vendor/EigenRand/EigenRand/MvDists/Multinomial.h

@@ -2,8 +2,8 @@
  * @file Multinomial.h
  * @author bab2min (bab2min@gmail.com)
  * @brief
- * @version 0.3.3
- * @date 2021-03-31
+ * @version 0.4.1
+ * @date 2022-08-13
  *
  * @copyright Copyright (c) 2020-2021
  *

data/vendor/EigenRand/EigenRand/MvDists/MvNormal.h

@@ -2,8 +2,8 @@
  * @file MvNormal.h
  * @author bab2min (bab2min@gmail.com)
  * @brief
- * @version 0.3.3
- * @date 2021-03-31
+ * @version 0.4.1
+ * @date 2022-08-13
  *
  * @copyright Copyright (c) 2020-2021
  *

data/vendor/EigenRand/EigenRand/PacketFilter.h

@@ -2,8 +2,8 @@
  * @file PacketFilter.h
  * @author bab2min (bab2min@gmail.com)
  * @brief 
- * @version 0.3.3
- * @date 2021-03-31
+ * @version 0.4.1
+ * @date 2022-08-13
  * 
  * @copyright Copyright (c) 2020-2021
  * 

data/vendor/EigenRand/EigenRand/PacketRandomEngine.h

@@ -2,8 +2,8 @@
  * @file PacketRandomEngine.h
  * @author bab2min (bab2min@gmail.com)
  * @brief
- * @version 0.3.3
- * @date 2021-03-31
+ * @version 0.4.1
+ * @date 2022-08-13
  *
  * @copyright Copyright (c) 2020-2021
  *
@@ -25,11 +25,17 @@ namespace Eigen
 		namespace detail
 		{
 			template<typename T>
-			auto test_integral_result_type(int)->std::integral_constant<bool, std::is_integral<typename T::result_type>::value>;
+			auto test_integral_result_type(int)->std::integral_constant<bool, std::is_integral<typename T::result_type>::value && !(T::min() == 0 && (T::max() & T::max() + 1) == 0)>;
 
 			template<typename T>
 			auto test_integral_result_type(...)->std::false_type;
 
+			template<typename T>
+			auto test_integral_fullbit_result_type(int)->std::integral_constant<bool, std::is_integral<typename T::result_type>::value&& T::min() == 0 && (T::max() & T::max() + 1) == 0>;
+
+			template<typename T>
+			auto test_integral_fullbit_result_type(...)->std::false_type;
+
 			template<typename T>
 			auto test_intpacket_result_type(int)->std::integral_constant<bool, internal::IsIntPacket<typename T::result_type>::value>;
 
@@ -42,6 +48,11 @@ namespace Eigen
 		{
 		};
 
+		template<typename Ty>
+		struct IsScalarFullBitRandomEngine : decltype(detail::test_integral_fullbit_result_type<Ty>(0))
+		{
+		};
+
 		template<typename Ty>
 		struct IsPacketRandomEngine : decltype(detail::test_intpacket_result_type<Ty>(0))
 		{
@@ -49,14 +60,15 @@ namespace Eigen
 
 		enum class RandomEngineType
 		{
-			none, scalar, packet
+			none, scalar, scalar_fullbit, packet
 		};
 
 		template<typename Ty>
 		struct GetRandomEngineType : std::integral_constant <
 			RandomEngineType,
 			IsPacketRandomEngine<Ty>::value ? RandomEngineType::packet :
-			(IsScalarRandomEngine<Ty>::value ? RandomEngineType::scalar : RandomEngineType::none)
+			IsScalarFullBitRandomEngine<Ty>::value ? RandomEngineType::scalar_fullbit :
+			IsScalarRandomEngine<Ty>::value ? RandomEngineType::scalar : RandomEngineType::none
 		>
 		{
 		};
@@ -257,7 +269,7 @@ namespace Eigen
 			 *
 			 * @return uint64_t
 			 */
-			uint64_t min() const
+			static constexpr uint64_t min()
 			{
 				return 0;
 			}
@@ -267,7 +279,7 @@ namespace Eigen
 			 *
 			 * @return uint64_t
 			 */
-			uint64_t max() const
+			static constexpr uint64_t max()
 			{
 				return _wMask;
 			}
@@ -418,6 +430,7 @@ namespace Eigen
 		class ParallelRandomEngineAdaptor
 		{
 			static_assert(GetRandomEngineType<BaseRng>::value != RandomEngineType::none, "BaseRng must be a kind of Random Engine.");
+			static_assert(GetRandomEngineType<BaseRng>::value != RandomEngineType::scalar, "BaseRng must be a kind of mersenne_twister_engine.");
 		public:
 			using result_type = UIntType;
 
@@ -547,7 +560,7 @@ namespace Eigen
 			IsPacketRandomEngine<typename std::remove_reference<Rng>::type>::value,
 			PacketRandomEngineAdaptor<UIntType, typename std::remove_reference<Rng>::type>,
 			typename std::conditional<
-			IsScalarRandomEngine<typename std::remove_reference<Rng>::type>::value,
+			IsScalarFullBitRandomEngine<typename std::remove_reference<Rng>::type>::value,
 			RandomEngineWrapper<typename std::remove_reference<Rng>::type>,
 			void
 			>::type
@@ -564,6 +577,8 @@ namespace Eigen
 		template<typename UIntType, typename Rng>
 		UniversalRandomEngine<UIntType, Rng> makeUniversalRng(Rng&& rng)
 		{
+			static_assert(IsPacketRandomEngine<typename std::remove_reference<Rng>::type>::value || IsScalarFullBitRandomEngine<typename std::remove_reference<Rng>::type>::value,
+				"`Rng` must be a kind of RandomPacketEngine like std::mersenne_twister_engine");
 			return { std::forward<Rng>(rng) };
 		}
 

data/vendor/EigenRand/EigenRand/RandUtils.h

@@ -2,8 +2,8 @@
  * @file RandUtils.h
  * @author bab2min (bab2min@gmail.com)
  * @brief 
- * @version 0.3.3
- * @date 2021-03-31
+ * @version 0.4.1
+ * @date 2022-08-13
  * 
  * @copyright Copyright (c) 2020-2021
  * 
@@ -55,7 +55,7 @@ namespace Eigen
 		struct scalar_rng_adaptor
 		{
 			static_assert(
-				Rand::IsScalarRandomEngine<
+				Rand::IsScalarFullBitRandomEngine<
 				typename std::remove_reference<Rng>::type
 				>::value ||
 				Rand::IsPacketRandomEngine<
@@ -95,7 +95,7 @@ namespace Eigen
 		struct scalar_rng_adaptor<Gen, _Scalar, Rng, true>
 		{
 			static_assert(
-				Rand::IsScalarRandomEngine<
+				Rand::IsScalarFullBitRandomEngine<
 				typename std::remove_reference<Rng>::type
 				>::value ||
 				Rand::IsPacketRandomEngine<

data/vendor/EigenRand/EigenRand/arch/AVX/MorePacketMath.h

@@ -2,8 +2,8 @@
  * @file MorePacketMath.h
  * @author bab2min (bab2min@gmail.com)
  * @brief
- * @version 0.3.5
- * @date 2021-07-16
+ * @version 0.4.1
+ * @date 2022-08-13
  *
  * @copyright Copyright (c) 2020-2021
  *

data/vendor/EigenRand/EigenRand/arch/AVX/PacketFilter.h

@@ -2,8 +2,8 @@
  * @file PacketFilter.h
  * @author bab2min (bab2min@gmail.com)
  * @brief
- * @version 0.4.0
- * @date 2021-09-17
+ * @version 0.4.1
+ * @date 2022-08-13
  *
  * @copyright Copyright (c) 2020-2021
  *

data/vendor/EigenRand/EigenRand/arch/AVX/RandUtils.h

@@ -2,8 +2,8 @@
  * @file RandUtils.h
  * @author bab2min (bab2min@gmail.com)
  * @brief
- * @version 0.4.0
- * @date 2021-09-17
+ * @version 0.4.1
+ * @date 2022-08-13
  *
  * @copyright Copyright (c) 2020-2021
  *
@@ -57,7 +57,7 @@ namespace Eigen
 		};
 
 		template<typename Rng, typename RngResult>
-		struct RawbitsMaker<Packet8i, Rng, RngResult, Rand::RandomEngineType::scalar>
+		struct RawbitsMaker<Packet8i, Rng, RngResult, Rand::RandomEngineType::scalar_fullbit>
 		{
 			EIGEN_STRONG_INLINE Packet8i rawbits(Rng& rng)
 			{

data/vendor/EigenRand/EigenRand/arch/NEON/MorePacketMath.h

@@ -2,8 +2,8 @@
  * @file MorePacketMath.h
  * @author bab2min (bab2min@gmail.com)
  * @brief
- * @version 0.4.0
- * @date 2021-04-26
+ * @version 0.4.1
+ * @date 2022-08-13
  *
  * @copyright Copyright (c) 2020-2021
  *

data/vendor/EigenRand/EigenRand/arch/NEON/PacketFilter.h

@@ -2,8 +2,8 @@
  * @file PacketFilter.h
  * @author bab2min (bab2min@gmail.com)
  * @brief
- * @version 0.4.0
- * @date 2021-09-17
+ * @version 0.4.1
+ * @date 2022-08-13
  *
  * @copyright Copyright (c) 2020-2021
  *

data/vendor/EigenRand/EigenRand/arch/NEON/RandUtils.h

@@ -2,8 +2,8 @@
  * @file RandUtils.h
  * @author bab2min (bab2min@gmail.com)
  * @brief
- * @version 0.4.0
- * @date 2021-09-17
+ * @version 0.4.1
+ * @date 2022-08-13
  *
  * @copyright Copyright (c) 2020-2021
  *
@@ -19,7 +19,7 @@ namespace Eigen
 	namespace internal
 	{
 		template<typename Rng, typename RngResult>
-		struct RawbitsMaker<Packet4i, Rng, RngResult, Rand::RandomEngineType::scalar>
+		struct RawbitsMaker<Packet4i, Rng, RngResult, Rand::RandomEngineType::scalar_fullbit>
 		{
 			EIGEN_STRONG_INLINE Packet4i rawbits(Rng& rng)
 			{

data/vendor/EigenRand/EigenRand/arch/SSE/MorePacketMath.h

@@ -2,8 +2,8 @@
  * @file MorePacketMath.h
  * @author bab2min (bab2min@gmail.com)
  * @brief
- * @version 0.3.5
- * @date 2021-07-16
+ * @version 0.4.1
+ * @date 2022-08-13
  *
  * @copyright Copyright (c) 2020-2021
  *

data/vendor/EigenRand/EigenRand/arch/SSE/PacketFilter.h

@@ -2,8 +2,8 @@
  * @file PacketFilter.h
  * @author bab2min (bab2min@gmail.com)
  * @brief
- * @version 0.4.0
- * @date 2021-09-17
+ * @version 0.4.1
+ * @date 2022-08-13
  *
  * @copyright Copyright (c) 2020-2021
  *

data/vendor/EigenRand/EigenRand/arch/SSE/RandUtils.h

@@ -2,8 +2,8 @@
  * @file RandUtils.h
  * @author bab2min (bab2min@gmail.com)
  * @brief
- * @version 0.4.0
- * @date 2021-09-17
+ * @version 0.4.1
+ * @date 2022-08-13
  *
  * @copyright Copyright (c) 2020-2021
  *
@@ -19,7 +19,7 @@ namespace Eigen
 	namespace internal
 	{
 		template<typename Rng, typename RngResult>
-		struct RawbitsMaker<Packet4i, Rng, RngResult, Rand::RandomEngineType::scalar>
+		struct RawbitsMaker<Packet4i, Rng, RngResult, Rand::RandomEngineType::scalar_fullbit>
 		{
 			EIGEN_STRONG_INLINE Packet4i rawbits(Rng& rng)
 			{

data/vendor/EigenRand/README.md

@@ -385,6 +385,10 @@ MIT License
 
 ## History
 
+### 0.4.1 (2022-08-13)
+* Fixed a bug where double-type generation with std::mt19937 fails compilation.
+* Fixed a bug where `UniformIntGen` in scalar mode generates numbers in the wrong range.
+
 ### 0.4.0 alpha (2021-09-28)
 * Now EigenRand supports ARM & ARM64 NEON architecture experimentally. Please report issues about ARM & ARM64 NEON.
 * Now EigenRand has compatibility to `Eigen 3.4.0`.

data/vendor/tomotopy/README.kr.rst

@@ -305,6 +305,13 @@ tomotopy의 Python3 예제 코드는 https://github.com/bab2min/tomotopy/blob/ma
 
 역사
 -------
+* 0.12.4 (2023-01-22)
+    * New features
+        * macOS ARM64 아키텍처에 대한 지원을 추가했습니다.
+    * Bug fixes
+        * `tomotopy.Document.get_sub_topic_dist()`가 bad argument 예외를 발생시키는 문제를 해결했습니다.
+        * 예외 발생이 종종 크래시를 발생시키는 문제를 해결했습니다.
+
 * 0.12.3 (2022-07-19)
     * 기능 개선
         * 이제 `tomotopy.LDAModel.add_doc()`로 빈 문서를 삽입할 경우 예외를 발생시키는 대신 그냥 무시합니다. 새로 추가된 인자인 `ignore_empty_words`를 False로 설정할 경우 이전처럼 예외를 발생시킵니다.

data/vendor/tomotopy/README.rst

@@ -309,6 +309,13 @@ meaning you can use it for any reasonable purpose and remain in complete ownersh
 
 History
 -------
+* 0.12.4 (2023-01-22)
+    * New features
+        * Added support for macOS ARM64 architecture.
+    * Bug fixes
+        * Fixed an issue where `tomotopy.Document.get_sub_topic_dist()` raises a bad argument exception.
+        * Fixed an issue where exception raising sometimes causes crashes.
+
 * 0.12.3 (2022-07-19)
     * New features
         * Now, inserting an empty document using `tomotopy.LDAModel.add_doc()` just ignores it instead of raising an exception. If the newly added argument `ignore_empty_words` is set to False, an exception is raised as before.
@@ -514,3 +521,18 @@ Bundled Libraries and Their License
 
 * Mapbox Variant: `BSD License
   <licenses_bundled/MapboxVariant>`_
+
+Citation
+---------
+::
+
+    @software{minchul_lee_2022_6868418,
+      author       = {Minchul Lee},
+      title        = {bab2min/tomotopy: 0.12.3},
+      month        = jul,
+      year         = 2022,
+      publisher    = {Zenodo},
+      version      = {v0.12.3},
+      doi          = {10.5281/zenodo.6868418},
+      url          = {https://doi.org/10.5281/zenodo.6868418}
+    }

data/vendor/tomotopy/src/Utils/EigenAddonOps.hpp

@@ -93,7 +93,7 @@ namespace Eigen
 #endif
 #ifdef EIGEN_VECTORIZE_NEON
 #include <arm_neon.h>
-
+#include "neon_gamma.h"
 
 namespace Eigen
 {

data/vendor/tomotopy/src/Utils/neon_gamma.h

@@ -0,0 +1,51 @@
+#pragma once
+
+inline float32x4_t accurate_rcp(float32x4_t x)
+{
+	float32x4_t r = vrecpeq_f32(x);
+	return vmulq_f32(vrecpsq_f32(x, r), r);
+}
+
+// approximation : lgamma(z) ~= (z+2.5)ln(z+3) - z - 3 + 0.5 ln (2pi) + 1/12/(z + 3) - ln (z(z+1)(z+2))
+inline float32x4_t lgamma_ps(float32x4_t x)
+{
+	float32x4_t x_3 = vaddq_f32(x, vmovq_n_f32(3));
+	float32x4_t ret = vmulq_f32(vaddq_f32(x_3, vmovq_n_f32(-0.5f)), Eigen::internal::plog(x_3));
+	ret = vsubq_f32(ret, x_3);
+	ret = vaddq_f32(ret, vmovq_n_f32(0.91893853f));
+	ret = vaddq_f32(ret, vdivq_f32(vmovq_n_f32(1 / 12.f), x_3));
+	ret = vsubq_f32(ret, Eigen::internal::plog(vmulq_f32(
+		vmulq_f32(vsubq_f32(x_3, vmovq_n_f32(1)), vsubq_f32(x_3, vmovq_n_f32(2))), x)));
+	return ret;
+}
+
+// approximation : lgamma(z + a) - lgamma(z) = (z + a + 1.5) * log(z + a + 2) - (z + 1.5) * log(z + 2) - a + (1. / (z + a + 2) - 1. / (z + 2)) / 12. - log(((z + a) * (z + a + 1)) / (z * (z + 1)))
+inline float32x4_t lgamma_subt(float32x4_t z, float32x4_t a)
+{
+	float32x4_t _1p5 = vmovq_n_f32(1.5);
+	float32x4_t _2 = vmovq_n_f32(2);
+	float32x4_t za = vaddq_f32(z, a);
+	float32x4_t ret = vmulq_f32(vaddq_f32(za, _1p5), Eigen::internal::plog(vaddq_f32(za, _2)));
+	ret = vsubq_f32(ret, vmulq_f32(vaddq_f32(z, _1p5), Eigen::internal::plog(vaddq_f32(z, _2))));
+	ret = vsubq_f32(ret, a);
+	float32x4_t _1 = vmovq_n_f32(1);
+	float32x4_t _1_12 = vmovq_n_f32(1 / 12.f);
+	ret = vaddq_f32(ret, vsubq_f32(vdivq_f32(_1_12, vaddq_f32(za, _2)), vdivq_f32(_1_12, vaddq_f32(z, _2))));
+	ret = vsubq_f32(ret, Eigen::internal::plog(vdivq_f32(vdivq_f32(vmulq_f32(za, vaddq_f32(za, _1)), z), vaddq_f32(z, _1))));
+	return ret;
+}
+
+
+// approximation : digamma(z) ~= ln(z+4) - 1/2/(z+4) - 1/12/(z+4)^2 - 1/z - 1/(z+1) - 1/(z+2) - 1/(z+3)
+inline float32x4_t digamma_ps(float32x4_t x)
+{
+	float32x4_t x_4 = vaddq_f32(x, vmovq_n_f32(4));
+	float32x4_t ret = Eigen::internal::plog(x_4);
+	ret = vsubq_f32(ret, vdivq_f32(vmovq_n_f32(1 / 2.f), x_4));
+	ret = vsubq_f32(ret, vdivq_f32(vdivq_f32(vmovq_n_f32(1 / 12.f), x_4), x_4));
+	ret = vsubq_f32(ret, accurate_rcp(vsubq_f32(x_4, vmovq_n_f32(1))));
+	ret = vsubq_f32(ret, accurate_rcp(vsubq_f32(x_4, vmovq_n_f32(2))));
+	ret = vsubq_f32(ret, accurate_rcp(vsubq_f32(x_4, vmovq_n_f32(3))));
+	ret = vsubq_f32(ret, accurate_rcp(vsubq_f32(x_4, vmovq_n_f32(4))));
+	return ret;
+}

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: tomoto
 version: !ruby/object:Gem::Version
-  version: 0.3.1
+  version: 0.3.3
 platform: ruby
 authors:
 - Andrew Kane
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2023-01-13 00:00:00.000000000 Z
+date: 2023-02-02 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rice
@@ -518,6 +518,7 @@ files:
 - vendor/tomotopy/src/Utils/avx_mathfun.h
 - vendor/tomotopy/src/Utils/exception.h
 - vendor/tomotopy/src/Utils/math.h
+- vendor/tomotopy/src/Utils/neon_gamma.h
 - vendor/tomotopy/src/Utils/rtnorm.hpp
 - vendor/tomotopy/src/Utils/sample.hpp
 - vendor/tomotopy/src/Utils/serializer.hpp