cld3 3.2.6 → 3.4.3
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/Gemfile +1 -1
- data/LICENSE +2 -2
- data/README.md +3 -3
- data/cld3.gemspec +9 -7
- data/ext/cld3/Makefile +45 -44
- data/ext/cld3/base.o +0 -0
- data/ext/cld3/embedding_feature_extractor.o +0 -0
- data/ext/cld3/embedding_network.cc +1 -0
- data/ext/cld3/embedding_network.o +0 -0
- data/ext/cld3/extconf.rb +3 -2
- data/ext/cld3/feature_extractor.o +0 -0
- data/ext/cld3/feature_extractor.pb.o +0 -0
- data/ext/cld3/feature_types.o +0 -0
- data/ext/cld3/fixunicodevalue.o +0 -0
- data/ext/cld3/fml_parser.o +0 -0
- data/ext/cld3/generated_entities.o +0 -0
- data/ext/cld3/generated_ulscript.o +0 -0
- data/ext/cld3/getonescriptspan.h +2 -2
- data/ext/cld3/getonescriptspan.o +0 -0
- data/ext/cld3/lang_id_nn_params.o +0 -0
- data/ext/cld3/language_identifier_features.o +0 -0
- data/ext/cld3/libcld3.def +8 -0
- data/ext/cld3/libcld3.so +0 -0
- data/ext/cld3/mkmf.log +10 -9
- data/ext/cld3/nnet_language_identifier.cc +3 -5
- data/ext/cld3/nnet_language_identifier.o +0 -0
- data/ext/cld3/nnet_language_identifier_c.cc +71 -23
- data/ext/cld3/nnet_language_identifier_c.o +0 -0
- data/ext/cld3/offsetmap.o +0 -0
- data/ext/cld3/registry.o +0 -0
- data/ext/cld3/relevant_script_feature.o +0 -0
- data/ext/cld3/sentence.pb.o +0 -0
- data/ext/cld3/sentence_features.cc +4 -4
- data/ext/cld3/sentence_features.h +13 -3
- data/ext/cld3/sentence_features.o +0 -0
- data/ext/cld3/task_context.o +0 -0
- data/ext/cld3/task_context_params.o +0 -0
- data/ext/cld3/task_spec.pb.o +0 -0
- data/ext/cld3/text_processing.o +0 -0
- data/ext/cld3/unicodetext.o +0 -0
- data/ext/cld3/utf8statetable.o +0 -0
- data/ext/cld3/utils.o +0 -0
- data/ext/cld3/workspace.o +0 -0
- data/lib/cld3/unstable.rb +58 -0
- data/lib/cld3.rb +88 -40
- data/sig/cld3.rbs +65 -0
- metadata +56 -13
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA256:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: 2c161cbf12d260074efd2e9db3981b6615af20ee04c234d6b2710bd52a283a4e
|
4
|
+
data.tar.gz: c388ae6b529d95e015ecdb7d21cdd7f1ceaca72d167d0f8008b5477d5bce5b3c
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: 8e3c1c07283730e722c450acc308a497756fd501595a02a7fc066d0b3e59b96e1ab1e7941549293b02e41274b176772bdae3779a041eb28f8ae53f5c44308cc0
|
7
|
+
data.tar.gz: 52e95027de7a595b2eabc49745a11f664e305c18f9926bc9d649642a92fea9846efdd23da699529795d80609b8871b00e77f9379449d2e4f6cb79ecbcf2785db
|
data/Gemfile
CHANGED
data/LICENSE
CHANGED
@@ -1,4 +1,4 @@
|
|
1
|
-
Copyright 2017 Akihiko Odaki <akihiko.odaki
|
1
|
+
Copyright 2017 Akihiko Odaki <akihiko.odaki@gmail.com>
|
2
2
|
All rights reserved.
|
3
3
|
|
4
4
|
Apache License
|
@@ -189,7 +189,7 @@ All rights reserved.
|
|
189
189
|
same "printed page" as the copyright notice for easier
|
190
190
|
identification within third-party archives.
|
191
191
|
|
192
|
-
Copyright 2017, Akihiko Odaki <akihiko.odaki
|
192
|
+
Copyright 2017, Akihiko Odaki <akihiko.odaki@gmail.com>
|
193
193
|
|
194
194
|
Licensed under the Apache License, Version 2.0 (the "License");
|
195
195
|
you may not use this file except in compliance with the License.
|
data/README.md
CHANGED
@@ -8,11 +8,11 @@ require 'cld3'
|
|
8
8
|
|
9
9
|
cld3 = CLD3::NNetLanguageIdentifier.new(0, 1000)
|
10
10
|
|
11
|
-
cld3.find_language("こんにちは") # => #<struct Struct::Result language=:ja, probability=1.0, reliable?=true, proportion=1.0>
|
11
|
+
cld3.find_language("こんにちは") # => #<struct Struct::Result language=:ja, probability=1.0, reliable?=true, proportion=1.0, byte_ranges=[]>
|
12
12
|
|
13
|
-
cld3.find_language("This is a pen.") # => #<struct Struct::Result language=:en, probability=0.9999408721923828, reliable?=true, proportion=1.0>
|
13
|
+
cld3.find_language("This is a pen.") # => #<struct Struct::Result language=:en, probability=0.9999408721923828, reliable?=true, proportion=1.0, byte_ranges=[]>
|
14
14
|
|
15
|
-
cld3.find_language("здравствуйте") # => #<struct Struct::Result language=:ru, probability=0.3140212297439575, reliable?=false, proportion=1.0>
|
15
|
+
cld3.find_language("здравствуйте") # => #<struct Struct::Result language=:ru, probability=0.3140212297439575, reliable?=false, proportion=1.0, byte_ranges=[]>
|
16
16
|
```
|
17
17
|
|
18
18
|
## Installation
|
data/cld3.gemspec
CHANGED
@@ -1,4 +1,4 @@
|
|
1
|
-
# Copyright 2017 Akihiko Odaki <akihiko.odaki
|
1
|
+
# Copyright 2017 Akihiko Odaki <akihiko.odaki@gmail.com>
|
2
2
|
# All Rights Reserved.
|
3
3
|
#
|
4
4
|
# Licensed under the Apache License, Version 2.0 (the "License");
|
@@ -16,19 +16,21 @@
|
|
16
16
|
|
17
17
|
Gem::Specification.new do |gem|
|
18
18
|
gem.name = "cld3"
|
19
|
-
gem.version = "3.
|
19
|
+
gem.version = "3.4.3"
|
20
20
|
gem.summary = "Compact Language Detector v3 (CLD3)"
|
21
21
|
gem.description = "Compact Language Detector v3 (CLD3) is a neural network model for language identification."
|
22
22
|
gem.license = "Apache-2.0"
|
23
23
|
gem.homepage = "https://github.com/akihikodaki/cld3-ruby"
|
24
24
|
gem.author = "Akihiko Odaki"
|
25
|
-
gem.email = "akihiko.odaki
|
26
|
-
gem.required_ruby_version = [ ">= 2.
|
27
|
-
gem.add_dependency "ffi", [ ">= 1.1.0", "< 1.
|
28
|
-
gem.add_development_dependency "
|
25
|
+
gem.email = "akihiko.odaki@gmail.com"
|
26
|
+
gem.required_ruby_version = [ ">= 2.6.0", "< 3.2.0" ]
|
27
|
+
gem.add_dependency "ffi", [ ">= 1.1.0", "< 1.16.0" ]
|
28
|
+
gem.add_development_dependency "rbs", [ ">= 1.7.0", "< 1.8.0" ]
|
29
|
+
gem.add_development_dependency "rspec", [ ">=3.0.0", "< 3.11.0" ]
|
30
|
+
gem.add_development_dependency "steep", [ ">= 0.46.0", "< 0.47.0" ]
|
29
31
|
gem.files = Dir[
|
30
32
|
"Gemfile", "LICENSE", "LICENSE_CLD3", "README.md",
|
31
|
-
"cld3.gemspec", "ext/**/*", "lib/**/*"
|
33
|
+
"cld3.gemspec", "ext/**/*", "lib/**/*", "sig/**/*"
|
32
34
|
]
|
33
35
|
gem.require_paths = [ "lib" ]
|
34
36
|
gem.extensions = [ "ext/cld3/extconf.rb" ]
|
data/ext/cld3/Makefile
CHANGED
@@ -2,7 +2,7 @@
|
|
2
2
|
SHELL = /bin/sh
|
3
3
|
|
4
4
|
# V=0 quiet, V=1 verbose. other values don't work.
|
5
|
-
V =
|
5
|
+
V = 1
|
6
6
|
Q1 = $(V:1=)
|
7
7
|
Q = $(Q1:0=@)
|
8
8
|
ECHO1 = $(V:1=@ :)
|
@@ -12,54 +12,55 @@ NULLCMD = :
|
|
12
12
|
#### Start of system configuration section. ####
|
13
13
|
|
14
14
|
srcdir = .
|
15
|
-
topdir = /usr/include
|
15
|
+
topdir = /usr/include
|
16
16
|
hdrdir = $(topdir)
|
17
|
-
arch_hdrdir = /usr/include
|
17
|
+
arch_hdrdir = /usr/include
|
18
18
|
PATH_SEPARATOR = :
|
19
19
|
VPATH = $(srcdir):$(arch_hdrdir)/ruby:$(hdrdir)/ruby
|
20
20
|
prefix = $(DESTDIR)/usr
|
21
|
-
rubysitearchprefix = $(
|
22
|
-
rubyarchprefix = $(
|
23
|
-
rubylibprefix = $(
|
24
|
-
exec_prefix = $(
|
25
|
-
vendorarchhdrdir = $(vendorhdrdir)/$(
|
26
|
-
sitearchhdrdir = $(sitehdrdir)/$(
|
27
|
-
rubyarchhdrdir = $(
|
21
|
+
rubysitearchprefix = $(sitearchlibdir)/$(RUBY_BASE_NAME)
|
22
|
+
rubyarchprefix = $(DESTDIR)/usr/lib64/ruby
|
23
|
+
rubylibprefix = $(exec_prefix)/share/ruby
|
24
|
+
exec_prefix = $(DESTDIR)/usr
|
25
|
+
vendorarchhdrdir = $(vendorhdrdir)/$(arch)
|
26
|
+
sitearchhdrdir = $(sitehdrdir)/$(arch)
|
27
|
+
rubyarchhdrdir = $(DESTDIR)/usr/include
|
28
28
|
vendorhdrdir = $(rubyhdrdir)/vendor_ruby
|
29
29
|
sitehdrdir = $(rubyhdrdir)/site_ruby
|
30
|
-
rubyhdrdir = $(
|
31
|
-
|
32
|
-
|
33
|
-
|
34
|
-
|
35
|
-
|
36
|
-
|
37
|
-
|
38
|
-
|
30
|
+
rubyhdrdir = $(DESTDIR)/usr/include
|
31
|
+
rubygemsdir = $(DESTDIR)/usr/share/rubygems
|
32
|
+
vendorarchdir = $(DESTDIR)/usr/lib64/ruby/vendor_ruby
|
33
|
+
vendorlibdir = $(vendordir)
|
34
|
+
vendordir = $(DESTDIR)/usr/share/ruby/vendor_ruby
|
35
|
+
sitearchdir = $(DESTDIR)/usr/local/lib64/ruby/site_ruby
|
36
|
+
sitelibdir = $(sitedir)
|
37
|
+
sitedir = $(DESTDIR)/usr/local/share/ruby/site_ruby
|
38
|
+
rubyarchdir = $(rubyarchprefix)
|
39
|
+
rubylibdir = $(rubylibprefix)
|
39
40
|
sitearchincludedir = $(includedir)/$(sitearch)
|
40
41
|
archincludedir = $(includedir)/$(arch)
|
41
42
|
sitearchlibdir = $(libdir)/$(sitearch)
|
42
|
-
archlibdir = $(
|
43
|
+
archlibdir = $(DESTDIR)/usr/lib64
|
43
44
|
ridir = $(datarootdir)/$(RI_BASE_NAME)
|
44
|
-
mandir = $(
|
45
|
+
mandir = $(DESTDIR)/usr/share/man
|
45
46
|
localedir = $(datarootdir)/locale
|
46
|
-
libdir = $(exec_prefix)/
|
47
|
+
libdir = $(exec_prefix)/lib64
|
47
48
|
psdir = $(docdir)
|
48
49
|
pdfdir = $(docdir)
|
49
50
|
dvidir = $(docdir)
|
50
51
|
htmldir = $(docdir)
|
51
|
-
infodir = $(
|
52
|
+
infodir = $(DESTDIR)/usr/share/info
|
52
53
|
docdir = $(datarootdir)/doc/$(PACKAGE)
|
53
54
|
oldincludedir = $(DESTDIR)/usr/include
|
54
|
-
includedir = $(
|
55
|
+
includedir = $(DESTDIR)/usr/include
|
55
56
|
runstatedir = $(localstatedir)/run
|
56
57
|
localstatedir = $(DESTDIR)/var
|
57
58
|
sharedstatedir = $(DESTDIR)/var/lib
|
58
59
|
sysconfdir = $(DESTDIR)/etc
|
59
|
-
datadir = $(
|
60
|
+
datadir = $(DESTDIR)/usr/share
|
60
61
|
datarootdir = $(prefix)/share
|
61
|
-
libexecdir = $(DESTDIR)/usr/
|
62
|
-
sbindir = $(
|
62
|
+
libexecdir = $(DESTDIR)/usr/libexec
|
63
|
+
sbindir = $(DESTDIR)/usr/sbin
|
63
64
|
bindir = $(exec_prefix)/bin
|
64
65
|
archdir = $(rubyarchdir)
|
65
66
|
|
@@ -78,36 +79,36 @@ CSRCFLAG = $(empty)
|
|
78
79
|
|
79
80
|
RUBY_EXTCONF_H =
|
80
81
|
cflags = $(optflags) $(debugflags) $(warnflags)
|
81
|
-
cxxflags =
|
82
|
+
cxxflags =
|
82
83
|
optflags = -O3
|
83
84
|
debugflags = -ggdb3
|
84
|
-
warnflags = -Wall -Wextra -
|
85
|
+
warnflags = -Wall -Wextra -Wdeprecated-declarations -Wduplicated-cond -Wimplicit-function-declaration -Wimplicit-int -Wmisleading-indentation -Wpointer-arith -Wwrite-strings -Wimplicit-fallthrough=0 -Wmissing-noreturn -Wno-cast-function-type -Wno-constant-logical-operand -Wno-long-long -Wno-missing-field-initializers -Wno-overlength-strings -Wno-packed-bitfield-compat -Wno-parentheses-equality -Wno-self-assign -Wno-tautological-compare -Wno-unused-parameter -Wno-unused-value -Wsuggest-attribute=format -Wsuggest-attribute=noreturn -Wunused-variable
|
85
86
|
cppflags =
|
86
87
|
CCDLFLAGS = -fPIC
|
87
|
-
CFLAGS = $(CCDLFLAGS) -
|
88
|
+
CFLAGS = $(CCDLFLAGS) -O2 -fexceptions -g -grecord-gcc-switches -pipe -Wall -Werror=format-security -Wp,-D_FORTIFY_SOURCE=2 -Wp,-D_GLIBCXX_ASSERTIONS -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -fstack-protector-strong -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -mbranch-protection=standard -fasynchronous-unwind-tables -fstack-clash-protection -fPIC $(ARCH_FLAG)
|
88
89
|
INCFLAGS = -I. -I$(arch_hdrdir) -I$(hdrdir)/ruby/backward -I$(hdrdir) -I$(srcdir)
|
89
90
|
DEFS =
|
90
|
-
CPPFLAGS =
|
91
|
-
CXXFLAGS = $(CCDLFLAGS) -
|
92
|
-
ldflags = -L. -Wl,-
|
93
|
-
dldflags = -Wl,-
|
91
|
+
CPPFLAGS = $(DEFS) $(cppflags)
|
92
|
+
CXXFLAGS = $(CCDLFLAGS) -O2 -fexceptions -g -grecord-gcc-switches -pipe -Wall -Werror=format-security -Wp,-D_FORTIFY_SOURCE=2 -Wp,-D_GLIBCXX_ASSERTIONS -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -fstack-protector-strong -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -mbranch-protection=standard -fasynchronous-unwind-tables -fstack-clash-protection -fvisibility=hidden -std=c++11 $(ARCH_FLAG)
|
93
|
+
ldflags = -L. -Wl,-z,relro -Wl,--as-needed -Wl,-z,now -specs=/usr/lib/rpm/redhat/redhat-hardened-ld -fstack-protector-strong -rdynamic -Wl,-export-dynamic
|
94
|
+
dldflags = -Wl,-z,relro -Wl,--as-needed -Wl,-z,now -specs=/usr/lib/rpm/redhat/redhat-hardened-ld
|
94
95
|
ARCH_FLAG =
|
95
96
|
DLDFLAGS = $(ldflags) $(dldflags) $(ARCH_FLAG)
|
96
97
|
LDSHARED = $(CC) -shared
|
97
98
|
LDSHAREDXX = $(CXX) -shared
|
98
|
-
AR = ar
|
99
|
+
AR = gcc-ar
|
99
100
|
EXEEXT =
|
100
101
|
|
101
102
|
RUBY_INSTALL_NAME = $(RUBY_BASE_NAME)
|
102
103
|
RUBY_SO_NAME = ruby
|
103
104
|
RUBYW_INSTALL_NAME =
|
104
|
-
RUBY_VERSION_NAME = $(RUBY_BASE_NAME)-$(
|
105
|
+
RUBY_VERSION_NAME = $(RUBY_BASE_NAME)-$(ruby_version_dir_name)
|
105
106
|
RUBYW_BASE_NAME = rubyw
|
106
107
|
RUBY_BASE_NAME = ruby
|
107
108
|
|
108
|
-
arch =
|
109
|
+
arch = aarch64-linux
|
109
110
|
sitearch = $(arch)
|
110
|
-
ruby_version =
|
111
|
+
ruby_version = 3.0.0
|
111
112
|
ruby = $(bindir)/$(RUBY_BASE_NAME)
|
112
113
|
RUBY = $(ruby)
|
113
114
|
ruby_headers = $(hdrdir)/ruby.h $(hdrdir)/ruby/backward.h $(hdrdir)/ruby/ruby.h $(hdrdir)/ruby/defines.h $(hdrdir)/ruby/missing.h $(hdrdir)/ruby/intern.h $(hdrdir)/ruby/st.h $(hdrdir)/ruby/subst.h $(arch_hdrdir)/ruby/config.h
|
@@ -125,8 +126,8 @@ TOUCH = exit >
|
|
125
126
|
#### End of system configuration section. ####
|
126
127
|
|
127
128
|
preload =
|
128
|
-
libpath = . $(
|
129
|
-
LIBPATH = -L. -L$(
|
129
|
+
libpath = . $(archlibdir)
|
130
|
+
LIBPATH = -L. -L$(archlibdir)
|
130
131
|
DEFFILE =
|
131
132
|
|
132
133
|
CLEANFILES = mkmf.log
|
@@ -137,11 +138,11 @@ extout =
|
|
137
138
|
extout_prefix =
|
138
139
|
target_prefix =
|
139
140
|
LOCAL_LIBS =
|
140
|
-
LIBS =
|
141
|
+
LIBS = -lprotobuf -lpthread -lm -lc
|
141
142
|
ORIG_SRCS = base.cc embedding_feature_extractor.cc embedding_network.cc feature_extractor.cc feature_extractor.pb.cc feature_types.cc fixunicodevalue.cc fml_parser.cc generated_entities.cc generated_ulscript.cc getonescriptspan.cc lang_id_nn_params.cc language_identifier_features.cc nnet_language_identifier.cc nnet_language_identifier_c.cc offsetmap.cc registry.cc relevant_script_feature.cc sentence.pb.cc sentence_features.cc task_context.cc task_context_params.cc task_spec.pb.cc text_processing.cc unicodetext.cc utf8statetable.cc utils.cc workspace.cc
|
142
143
|
SRCS = $(ORIG_SRCS)
|
143
144
|
OBJS = base.o embedding_feature_extractor.o embedding_network.o feature_extractor.o feature_extractor.pb.o feature_types.o fixunicodevalue.o fml_parser.o generated_entities.o generated_ulscript.o getonescriptspan.o lang_id_nn_params.o language_identifier_features.o nnet_language_identifier.o nnet_language_identifier_c.o offsetmap.o registry.o relevant_script_feature.o sentence.pb.o sentence_features.o task_context.o task_context_params.o task_spec.pb.o text_processing.o unicodetext.o utf8statetable.o utils.o workspace.o
|
144
|
-
HDRS = $(srcdir)/base.h $(srcdir)/casts.h $(srcdir)/embedding_feature_extractor.h $(srcdir)/embedding_network.h $(srcdir)/embedding_network_params.h $(srcdir)/feature_extractor.h $(srcdir)/
|
145
|
+
HDRS = $(srcdir)/base.h $(srcdir)/casts.h $(srcdir)/embedding_feature_extractor.h $(srcdir)/embedding_network.h $(srcdir)/embedding_network_params.h $(srcdir)/feature_extractor.h $(srcdir)/feature_extractor.pb.h $(srcdir)/feature_types.h $(srcdir)/fixunicodevalue.h $(srcdir)/float16.h $(srcdir)/fml_parser.h $(srcdir)/generated_ulscript.h $(srcdir)/getonescriptspan.h $(srcdir)/integral_types.h $(srcdir)/lang_id_nn_params.h $(srcdir)/language_identifier_features.h $(srcdir)/nnet_language_identifier.h $(srcdir)/offsetmap.h $(srcdir)/port.h $(srcdir)/registry.h $(srcdir)/relevant_script_feature.h $(srcdir)/script_detector.h $(srcdir)/sentence.pb.h $(srcdir)/sentence_features.h $(srcdir)/simple_adder.h $(srcdir)/stringpiece.h $(srcdir)/task_context.h $(srcdir)/task_context_params.h $(srcdir)/task_spec.pb.h $(srcdir)/text_processing.h $(srcdir)/unicodetext.h $(srcdir)/utf8acceptinterchange.h $(srcdir)/utf8prop_lettermarkscriptnum.h $(srcdir)/utf8repl_lettermarklower.h $(srcdir)/utf8scannot_lettermarkspecial.h $(srcdir)/utf8statetable.h $(srcdir)/utils.h $(srcdir)/workspace.h
|
145
146
|
LOCAL_HDRS =
|
146
147
|
TARGET = libcld3
|
147
148
|
TARGET_NAME = libcld3
|
@@ -155,8 +156,8 @@ BINDIR = $(bindir)
|
|
155
156
|
RUBYCOMMONDIR = $(sitedir)$(target_prefix)
|
156
157
|
RUBYLIBDIR = $(sitelibdir)$(target_prefix)
|
157
158
|
RUBYARCHDIR = $(sitearchdir)$(target_prefix)
|
158
|
-
HDRDIR = $(
|
159
|
-
ARCHHDRDIR = $(
|
159
|
+
HDRDIR = $(sitehdrdir)$(target_prefix)
|
160
|
+
ARCHHDRDIR = $(sitearchhdrdir)$(target_prefix)
|
160
161
|
TARGET_SO_DIR =
|
161
162
|
TARGET_SO = $(TARGET_SO_DIR)$(DLLIB)
|
162
163
|
CLEANLIBS = $(TARGET_SO)
|
data/ext/cld3/base.o
CHANGED
Binary file
|
Binary file
|
@@ -167,6 +167,7 @@ EmbeddingNetwork::EmbeddingNetwork(const EmbeddingNetworkParams *model)
|
|
167
167
|
for (int i = 0; i < model_->embedding_dim_size(); ++i) {
|
168
168
|
CLD3_DCHECK(offset_sum == model_->concat_offset(i));
|
169
169
|
offset_sum += model_->embedding_dim(i) * model_->embedding_num_features(i);
|
170
|
+
(void)offset_sum; // Avoid compiler warning for "unused" variable.
|
170
171
|
embedding_matrices_.emplace_back(model_->GetEmbeddingMatrix(i));
|
171
172
|
}
|
172
173
|
|
Binary file
|
data/ext/cld3/extconf.rb
CHANGED
@@ -1,4 +1,4 @@
|
|
1
|
-
# Copyright 2017 Akihiko Odaki <akihiko.odaki
|
1
|
+
# Copyright 2017 Akihiko Odaki <akihiko.odaki@gmail.com>
|
2
2
|
# All Rights Reserved.
|
3
3
|
#
|
4
4
|
# Licensed under the Apache License, Version 2.0 (the "License");
|
@@ -33,7 +33,7 @@ FileUtils.mkdir_p("cld_3/protos")
|
|
33
33
|
FileUtils.mkdir_p("script_span")
|
34
34
|
|
35
35
|
[ "feature_extractor", "sentence", "task_spec" ].each {|name|
|
36
|
-
|
36
|
+
system "protoc", "#{name}.proto", "--cpp_out=.", exception: true
|
37
37
|
ln_fallback("#{name}.pb.h", "cld_3/protos/#{name}.pb.h")
|
38
38
|
}
|
39
39
|
|
@@ -56,4 +56,5 @@ FileUtils.mkdir_p("script_span")
|
|
56
56
|
}
|
57
57
|
|
58
58
|
$CXXFLAGS += " -fvisibility=hidden -std=c++11"
|
59
|
+
$LIBRUBYARG = ""
|
59
60
|
create_makefile("libcld3")
|
Binary file
|
Binary file
|
data/ext/cld3/feature_types.o
CHANGED
Binary file
|
data/ext/cld3/fixunicodevalue.o
CHANGED
Binary file
|
data/ext/cld3/fml_parser.o
CHANGED
Binary file
|
Binary file
|
Binary file
|
data/ext/cld3/getonescriptspan.h
CHANGED
@@ -33,14 +33,14 @@ static const int kMaxScriptBytes = kMaxScriptBuffer - 32; // Leave some room
|
|
33
33
|
static const int kWithinScriptTail = 32; // Stop at word space in last
|
34
34
|
// N bytes of script buffer
|
35
35
|
|
36
|
-
|
36
|
+
struct LangSpan {
|
37
37
|
char* text = nullptr; // Pointer to the span, somewhere
|
38
38
|
int text_bytes = 0; // Number of bytes of text in the span
|
39
39
|
int offset = 0; // Offset of start of span in original input buffer
|
40
40
|
ULScript ulscript = UNKNOWN_ULSCRIPT; // Unicode Letters Script of this span
|
41
41
|
bool truncated = false; // true if buffer filled up before a
|
42
42
|
// different script or EOF was found
|
43
|
-
}
|
43
|
+
};
|
44
44
|
|
45
45
|
static inline bool IsContinuationByte(char c) {
|
46
46
|
return static_cast<signed char>(c) < -64;
|
data/ext/cld3/getonescriptspan.o
CHANGED
Binary file
|
Binary file
|
Binary file
|
data/ext/cld3/libcld3.so
CHANGED
Binary file
|
data/ext/cld3/mkmf.log
CHANGED
@@ -1,36 +1,37 @@
|
|
1
1
|
"pkg-config --exists protobuf"
|
2
2
|
| pkg-config --libs protobuf
|
3
|
-
=> "-lprotobuf \n"
|
4
|
-
"gcc -o conftest -I/usr/include
|
3
|
+
=> "-lprotobuf -lpthread \n"
|
4
|
+
"gcc -o conftest -I/usr/include -I/usr/include/ruby/backward -I/usr/include -I. -O2 -fexceptions -g -grecord-gcc-switches -pipe -Wall -Werror=format-security -Wp,-D_FORTIFY_SOURCE=2 -Wp,-D_GLIBCXX_ASSERTIONS -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -fstack-protector-strong -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -mbranch-protection=standard -fasynchronous-unwind-tables -fstack-clash-protection -fPIC conftest.c -L. -L/usr/lib64 -L. -Wl,-z,relro -Wl,--as-needed -Wl,-z,now -specs=/usr/lib/rpm/redhat/redhat-hardened-ld -fstack-protector-strong -rdynamic -Wl,-export-dynamic -lruby -lm -lc"
|
5
5
|
checked program was:
|
6
6
|
/* begin */
|
7
7
|
1: #include "ruby.h"
|
8
8
|
2:
|
9
9
|
3: int main(int argc, char **argv)
|
10
10
|
4: {
|
11
|
-
5: return
|
11
|
+
5: return !!argv[argc];
|
12
12
|
6: }
|
13
13
|
/* end */
|
14
14
|
|
15
|
-
"gcc -o conftest -I/usr/include
|
15
|
+
"gcc -o conftest -I/usr/include -I/usr/include/ruby/backward -I/usr/include -I. -O2 -fexceptions -g -grecord-gcc-switches -pipe -Wall -Werror=format-security -Wp,-D_FORTIFY_SOURCE=2 -Wp,-D_GLIBCXX_ASSERTIONS -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -fstack-protector-strong -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -mbranch-protection=standard -fasynchronous-unwind-tables -fstack-clash-protection -fPIC conftest.c -L. -L/usr/lib64 -L. -Wl,-z,relro -Wl,--as-needed -Wl,-z,now -specs=/usr/lib/rpm/redhat/redhat-hardened-ld -fstack-protector-strong -rdynamic -Wl,-export-dynamic -lruby -lprotobuf -lpthread -lm -lc"
|
16
16
|
checked program was:
|
17
17
|
/* begin */
|
18
18
|
1: #include "ruby.h"
|
19
19
|
2:
|
20
20
|
3: int main(int argc, char **argv)
|
21
21
|
4: {
|
22
|
-
5: return
|
22
|
+
5: return !!argv[argc];
|
23
23
|
6: }
|
24
24
|
/* end */
|
25
25
|
|
26
26
|
| pkg-config --cflags-only-I protobuf
|
27
27
|
=> "\n"
|
28
28
|
| pkg-config --cflags-only-other protobuf
|
29
|
-
=> "
|
29
|
+
=> "\n"
|
30
30
|
| pkg-config --libs-only-l protobuf
|
31
|
-
=> "-lprotobuf \n"
|
31
|
+
=> "-lprotobuf -lpthread \n"
|
32
32
|
package configuration for protobuf
|
33
|
-
|
33
|
+
incflags:
|
34
|
+
cflags:
|
34
35
|
ldflags:
|
35
|
-
libs: -lprotobuf
|
36
|
+
libs: -lprotobuf -lpthread
|
36
37
|
|
@@ -284,8 +284,6 @@ NNetLanguageIdentifier::FindTopNMostFreqLangs(const string &text,
|
|
284
284
|
CLD2::LangSpan script_span;
|
285
285
|
std::unordered_map<string, LangChunksStats> lang_stats;
|
286
286
|
int total_num_bytes = 0;
|
287
|
-
Result result;
|
288
|
-
string language;
|
289
287
|
int chunk_size = 0; // Use the default.
|
290
288
|
while (ss.GetOneScriptSpanLower(&script_span)) {
|
291
289
|
const int num_original_span_bytes = script_span.text_bytes;
|
@@ -302,8 +300,8 @@ NNetLanguageIdentifier::FindTopNMostFreqLangs(const string &text,
|
|
302
300
|
|
303
301
|
const string selected_text = SelectTextGivenScriptSpan(script_span);
|
304
302
|
|
305
|
-
result = FindLanguageOfValidUTF8(selected_text);
|
306
|
-
language = result.language;
|
303
|
+
Result result = FindLanguageOfValidUTF8(selected_text);
|
304
|
+
string language = result.language;
|
307
305
|
lang_stats[language].byte_sum += num_original_span_bytes;
|
308
306
|
lang_stats[language].prob_sum +=
|
309
307
|
result.probability * num_original_span_bytes;
|
@@ -356,7 +354,7 @@ string NNetLanguageIdentifier::SelectTextGivenBeginAndSize(
|
|
356
354
|
const char *text_begin, int text_size) {
|
357
355
|
string output_text;
|
358
356
|
|
359
|
-
// If the size of the input is greater than the
|
357
|
+
// If the size of the input is greater than the maximum number of bytes needed
|
360
358
|
// for a prediction, then concatenate snippets that are equally spread out
|
361
359
|
// throughout the input.
|
362
360
|
if (text_size > max_num_bytes_) {
|
Binary file
|
@@ -1,4 +1,4 @@
|
|
1
|
-
/* Copyright 2017 Akihiko Odaki <akihiko.odaki
|
1
|
+
/* Copyright 2017 Akihiko Odaki <akihiko.odaki@gmail.com>
|
2
2
|
All Rights Reserved.
|
3
3
|
|
4
4
|
Licensed under the Apache License, Version 2.0 (the "License");
|
@@ -26,42 +26,90 @@ limitations under the License.
|
|
26
26
|
#define EXPORT __attribute__ ((visibility ("default")))
|
27
27
|
#endif
|
28
28
|
|
29
|
-
struct NNetLanguageIdentifier {
|
30
|
-
chrome_lang_id::NNetLanguageIdentifier context;
|
31
|
-
std::string language;
|
32
|
-
};
|
33
|
-
|
34
29
|
struct Result {
|
35
30
|
struct {
|
36
31
|
const char *data;
|
37
32
|
std::size_t size;
|
38
33
|
} language;
|
34
|
+
struct {
|
35
|
+
const chrome_lang_id::NNetLanguageIdentifier::SpanInfo *data;
|
36
|
+
std::size_t size;
|
37
|
+
} byte_ranges;
|
39
38
|
float probability;
|
40
39
|
float proportion;
|
41
40
|
bool is_reliable;
|
42
41
|
};
|
43
42
|
|
43
|
+
struct OwningResult {
|
44
|
+
OwningResult(chrome_lang_id::NNetLanguageIdentifier::Result&& result) {
|
45
|
+
references.language = std::move(result.language);
|
46
|
+
references.byte_ranges = std::move(result.byte_ranges);
|
47
|
+
plain.language.data = references.language.data();
|
48
|
+
plain.language.size = references.language.size();
|
49
|
+
plain.byte_ranges.data = references.byte_ranges.data();
|
50
|
+
plain.byte_ranges.size = references.byte_ranges.size();
|
51
|
+
plain.probability = result.probability;
|
52
|
+
plain.proportion = result.proportion;
|
53
|
+
plain.is_reliable = result.is_reliable;
|
54
|
+
}
|
55
|
+
|
56
|
+
Result plain;
|
57
|
+
struct {
|
58
|
+
std::string language;
|
59
|
+
std::vector<chrome_lang_id::NNetLanguageIdentifier::SpanInfo> byte_ranges;
|
60
|
+
} references;
|
61
|
+
};
|
62
|
+
|
44
63
|
extern "C" {
|
45
|
-
EXPORT
|
46
|
-
|
47
|
-
|
48
|
-
|
49
|
-
|
50
|
-
instance->language = std::move(result.language);
|
51
|
-
|
52
|
-
return Result {
|
53
|
-
{ instance->language.data(), instance->language.size() },
|
54
|
-
result.probability,
|
55
|
-
result.proportion,
|
56
|
-
result.is_reliable
|
57
|
-
};
|
64
|
+
EXPORT OwningResult *NNetLanguageIdentifier_find_language(
|
65
|
+
chrome_lang_id::NNetLanguageIdentifier *instance,
|
66
|
+
const char *data,
|
67
|
+
std::size_t size) {
|
68
|
+
return new OwningResult(instance->FindLanguage(std::string(data, size)));
|
58
69
|
}
|
59
70
|
|
60
|
-
EXPORT
|
61
|
-
|
71
|
+
EXPORT std::vector<chrome_lang_id::NNetLanguageIdentifier::Result>*
|
72
|
+
NNetLanguageIdentifier_find_top_n_most_freq_langs(
|
73
|
+
chrome_lang_id::NNetLanguageIdentifier *instance,
|
74
|
+
const char *data, std::size_t size, int num_langs) {
|
75
|
+
std::string text(data, size);
|
76
|
+
return new auto(instance->FindTopNMostFreqLangs(text, num_langs));
|
62
77
|
}
|
63
78
|
|
64
|
-
EXPORT void
|
65
|
-
|
79
|
+
EXPORT void delete_NNetLanguageIdentifier(
|
80
|
+
chrome_lang_id::NNetLanguageIdentifier *pointer) {
|
81
|
+
delete pointer;
|
82
|
+
}
|
83
|
+
|
84
|
+
EXPORT void delete_result(OwningResult *pointer) {
|
85
|
+
delete pointer;
|
86
|
+
}
|
87
|
+
|
88
|
+
EXPORT void delete_results(
|
89
|
+
std::vector<chrome_lang_id::NNetLanguageIdentifier::Result> *pointer) {
|
90
|
+
delete pointer;
|
91
|
+
}
|
92
|
+
|
93
|
+
EXPORT chrome_lang_id::NNetLanguageIdentifier *new_NNetLanguageIdentifier(
|
94
|
+
int min_num_bytes, int max_num_bytes) {
|
95
|
+
return new chrome_lang_id::NNetLanguageIdentifier(
|
96
|
+
min_num_bytes, max_num_bytes);
|
97
|
+
}
|
98
|
+
|
99
|
+
EXPORT Result refer_to_nth_result(
|
100
|
+
std::vector<chrome_lang_id::NNetLanguageIdentifier::Result> *results,
|
101
|
+
std::size_t index) {
|
102
|
+
Result c;
|
103
|
+
auto& cc = (*results)[index];
|
104
|
+
|
105
|
+
c.language.data = cc.language.data();
|
106
|
+
c.language.size = cc.language.size();
|
107
|
+
c.byte_ranges.data = cc.byte_ranges.data();
|
108
|
+
c.byte_ranges.size = cc.byte_ranges.size();
|
109
|
+
c.probability = cc.probability;
|
110
|
+
c.proportion = cc.proportion;
|
111
|
+
c.is_reliable = cc.is_reliable;
|
112
|
+
|
113
|
+
return c;
|
66
114
|
}
|
67
115
|
}
|
Binary file
|
data/ext/cld3/offsetmap.o
CHANGED
Binary file
|
data/ext/cld3/registry.o
CHANGED
Binary file
|
Binary file
|
data/ext/cld3/sentence.pb.o
CHANGED
Binary file
|
@@ -19,11 +19,11 @@ limitations under the License.
|
|
19
19
|
|
20
20
|
namespace chrome_lang_id {
|
21
21
|
|
22
|
-
//
|
22
|
+
// Define registry for the whole Sentence feature functions. NOTE: this is not
|
23
23
|
// yet set to anything meaningful. It will be set so in NNetLanguageIdentifier
|
24
24
|
// constructor, *before* we use any feature.
|
25
25
|
template <>
|
26
|
-
WholeSentenceFeature::Registry
|
27
|
-
|
26
|
+
WholeSentenceFeature::Registry*
|
27
|
+
RegisterableClass<WholeSentenceFeature>::registry_ = nullptr;
|
28
28
|
|
29
|
-
} // namespace chrome_lang_id
|
29
|
+
} // namespace chrome_lang_id
|
@@ -26,9 +26,19 @@ limitations under the License.
|
|
26
26
|
namespace chrome_lang_id {
|
27
27
|
|
28
28
|
// Feature function that extracts features for the full Sentence.
|
29
|
-
|
30
|
-
|
31
|
-
|
29
|
+
using WholeSentenceFeature = FeatureFunction<Sentence>;
|
30
|
+
|
31
|
+
using WholeSentenceExtractor = FeatureExtractor<Sentence>;
|
32
|
+
|
33
|
+
// Declare registry for the whole Sentence feature functions. This is required
|
34
|
+
// for clang's -Wundefined-var-template. However, MSVC has a bug which treats
|
35
|
+
// this declaration as a definition, leading to multiple definition errors, so
|
36
|
+
// omit this on MSVC.
|
37
|
+
#if !defined(COMPILER_MSVC)
|
38
|
+
template <>
|
39
|
+
WholeSentenceFeature::Registry
|
40
|
+
*RegisterableClass<WholeSentenceFeature>::registry_;
|
41
|
+
#endif
|
32
42
|
|
33
43
|
} // namespace chrome_lang_id
|
34
44
|
|
Binary file
|
data/ext/cld3/task_context.o
CHANGED
Binary file
|
Binary file
|
data/ext/cld3/task_spec.pb.o
CHANGED
Binary file
|
data/ext/cld3/text_processing.o
CHANGED
Binary file
|
data/ext/cld3/unicodetext.o
CHANGED
Binary file
|
data/ext/cld3/utf8statetable.o
CHANGED
Binary file
|
data/ext/cld3/utils.o
CHANGED
Binary file
|
data/ext/cld3/workspace.o
CHANGED
Binary file
|
@@ -0,0 +1,58 @@
|
|
1
|
+
|
2
|
+
# Copyright 2021 Akihiko Odaki <akihiko.odaki@gmail.com>
|
3
|
+
# All Rights Reserved.
|
4
|
+
#
|
5
|
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
6
|
+
# you may not use this file except in compliance with the License.
|
7
|
+
# You may obtain a copy of the License at
|
8
|
+
#
|
9
|
+
# http://www.apache.org/licenses/LICENSE-2.0
|
10
|
+
#
|
11
|
+
# Unless required by applicable law or agreed to in writing, software
|
12
|
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
13
|
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
14
|
+
# See the License for the specific language governing permissions and
|
15
|
+
# limitations under the License.
|
16
|
+
# ==============================================================================
|
17
|
+
|
18
|
+
module CLD3
|
19
|
+
module Unstable
|
20
|
+
extend FFI::Library
|
21
|
+
|
22
|
+
ffi_lib File.join(__dir__, "..", "..", "ext", "cld3", "libcld3." + RbConfig::CONFIG["DLEXT"])
|
23
|
+
|
24
|
+
module NNetLanguageIdentifier
|
25
|
+
class Pointer < FFI::AutoPointer
|
26
|
+
def self.release(pointer)
|
27
|
+
Unstable.delete_NNetLanguageIdentifier(pointer)
|
28
|
+
end
|
29
|
+
end
|
30
|
+
|
31
|
+
class SpanInfo < FFI::Struct
|
32
|
+
layout :start_index, :int, :end_index, :int, :probability, :float
|
33
|
+
end
|
34
|
+
|
35
|
+
class Result < FFI::Struct
|
36
|
+
layout :language_data, :pointer, :language_size, :size_t, :byte_ranges_data, :pointer, :byte_ranges_size, :size_t, :probability, :float, :proportion, :float, :reliable?, :bool
|
37
|
+
end
|
38
|
+
end
|
39
|
+
|
40
|
+
attach_function :delete_NNetLanguageIdentifier, [ :pointer ], :void
|
41
|
+
|
42
|
+
attach_function :delete_result, [ :pointer ], :void
|
43
|
+
|
44
|
+
attach_function :delete_results, [ :pointer ], :void
|
45
|
+
|
46
|
+
attach_function :new_NNetLanguageIdentifier, [ :int, :int ], :pointer
|
47
|
+
|
48
|
+
attach_function :refer_to_nth_result, [ :pointer, :size_t ], NNetLanguageIdentifier::Result.by_value
|
49
|
+
|
50
|
+
attach_function :NNetLanguageIdentifier_find_language,
|
51
|
+
[ :pointer, :buffer_in, :size_t ], :pointer
|
52
|
+
|
53
|
+
attach_function :NNetLanguageIdentifier_find_top_n_most_freq_langs,
|
54
|
+
[ :pointer, :buffer_in, :size_t, :int ], :pointer
|
55
|
+
end
|
56
|
+
|
57
|
+
private_constant :Unstable
|
58
|
+
end
|
data/lib/cld3.rb
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
# File including an implementation of CLD3 module. Some documentations are
|
2
2
|
# extracted from ext/cld3/ext/src/nnet_language_identifier.h.
|
3
3
|
#
|
4
|
-
# Copyright 2017 Akihiko Odaki <akihiko.odaki
|
4
|
+
# Copyright 2017 Akihiko Odaki <akihiko.odaki@gmail.com>
|
5
5
|
# All Rights Reserved.
|
6
6
|
#
|
7
7
|
# Licensed under the Apache License, Version 2.0 (the "License");
|
@@ -19,6 +19,7 @@
|
|
19
19
|
|
20
20
|
require "ffi"
|
21
21
|
require "rbconfig"
|
22
|
+
require "cld3/unstable"
|
22
23
|
|
23
24
|
# Module providing an interface for Compact Language Detector v3 (CLD3)
|
24
25
|
module CLD3
|
@@ -49,10 +50,16 @@ module CLD3
|
|
49
50
|
# This is Numeric object.
|
50
51
|
RELIABILITY_HR_BS_THRESHOLD = 0.5
|
51
52
|
|
53
|
+
# Holds probability that Span, specified by start/end indices, is a given
|
54
|
+
# language. The langauge is not stored here; it can be found in Result, which
|
55
|
+
# holds an Array of SpanInfo.
|
56
|
+
# @type const SpanInfo: untyped
|
57
|
+
SpanInfo = Struct.new(:start_index, :end_index, :probability)
|
58
|
+
|
52
59
|
# Information about a predicted language.
|
53
60
|
# This is an instance of Struct with the following members:
|
54
61
|
#
|
55
|
-
# [language] This is symbol
|
62
|
+
# [language] This is symbol.
|
56
63
|
#
|
57
64
|
# [probability] Language probability. This is Numeric object.
|
58
65
|
#
|
@@ -61,33 +68,100 @@ module CLD3
|
|
61
68
|
# [proportion] Proportion of bytes associated with the language. If
|
62
69
|
# #find_language is called, this variable is set to 1.
|
63
70
|
# This is Numeric object.
|
64
|
-
|
71
|
+
#
|
72
|
+
# [byte_ranges] Specifies the byte ranges in UTF-8 that |language| applies to.
|
73
|
+
# This is an Array of SpanInfo.
|
74
|
+
# @type const Result: untyped
|
75
|
+
Result = Struct.new(:language, :probability, :reliable?, :proportion, :byte_ranges)
|
65
76
|
|
66
77
|
# The arguments are two String objects.
|
67
|
-
def initialize(
|
68
|
-
@cc = Unstable::NNetLanguageIdentifier::Pointer.new(Unstable.new_NNetLanguageIdentifier(
|
78
|
+
def initialize(min_num_bytes = MIN_NUM_BYTES_TO_CONSIDER, max_num_bytes = MAX_NUM_BYTES_TO_CONSIDER)
|
79
|
+
@cc = Unstable::NNetLanguageIdentifier::Pointer.new(Unstable.new_NNetLanguageIdentifier(min_num_bytes, max_num_bytes))
|
69
80
|
end
|
70
81
|
|
71
82
|
# Finds the most likely language for the given text, along with additional
|
72
83
|
# information (e.g., probability). The prediction is based on the first N
|
73
84
|
# bytes where N is the minumum between the number of interchange valid UTF8
|
74
85
|
# bytes and +max_num_bytes_+. If N is less than +min_num_bytes_+ long, then
|
75
|
-
# this function returns nil
|
86
|
+
# this function returns nil.
|
76
87
|
# The argument is a String object.
|
77
88
|
# The returned value of this function is an instance of Result.
|
78
89
|
def find_language(text)
|
79
90
|
text_utf8 = text.encode(Encoding::UTF_8)
|
80
91
|
pointer = FFI::MemoryPointer.new(:char, text_utf8.bytesize)
|
81
|
-
pointer.put_bytes(0, text_utf8)
|
82
92
|
|
83
|
-
|
84
|
-
|
93
|
+
begin
|
94
|
+
pointer.put_bytes(0, text_utf8)
|
95
|
+
|
96
|
+
result = Unstable.NNetLanguageIdentifier_find_language(@cc, pointer, text_utf8.bytesize)
|
97
|
+
begin
|
98
|
+
convert_result Unstable::NNetLanguageIdentifier::Result.new(result)
|
99
|
+
ensure
|
100
|
+
Unstable.delete_result result
|
101
|
+
end
|
102
|
+
ensure
|
103
|
+
pointer.free
|
104
|
+
end
|
105
|
+
end
|
106
|
+
|
107
|
+
# Splits the input text (up to the first byte, if any, that is not
|
108
|
+
# interchange valid UTF8) into spans based on the script, predicts a language
|
109
|
+
# for each span, and returns a vector storing the top num_langs most frequent
|
110
|
+
# languages along with additional information (e.g., proportions). The number
|
111
|
+
# of bytes considered for each span is the minimum between the size of the
|
112
|
+
# span and +max_num_bytes_+. If more languages are requested than what is
|
113
|
+
# available in the input, then the number of the returned elements will be
|
114
|
+
# the number of the latter. Also, if the size of the span is less than
|
115
|
+
# +min_num_bytes_+ long, then the span is skipped. If the input text is too
|
116
|
+
# long, only the first +MAX_NUM_INPUT_BYTES_TO_CONSIDER+ bytes are processed.
|
117
|
+
# The first argument is a String object.
|
118
|
+
# The second argument is Numeric object.
|
119
|
+
# The returned value of this functions is an Array of Result instances.
|
120
|
+
def find_top_n_most_freq_langs(text, num_langs)
|
121
|
+
# @type var a: untyped
|
122
|
+
|
123
|
+
text_utf8 = text.encode(Encoding::UTF_8)
|
124
|
+
pointer = FFI::MemoryPointer.new(:char, text_utf8.bytesize)
|
125
|
+
|
126
|
+
begin
|
127
|
+
pointer.put_bytes(0, text_utf8)
|
128
|
+
|
129
|
+
results = Unstable.NNetLanguageIdentifier_find_top_n_most_freq_langs(@cc, pointer, text_utf8.bytesize, num_langs)
|
130
|
+
begin
|
131
|
+
a = num_langs.times
|
132
|
+
.lazy
|
133
|
+
.map { |index| convert_result Unstable.refer_to_nth_result(results, index) }
|
134
|
+
.take_while { |result| !result.nil? }
|
135
|
+
.to_a
|
136
|
+
|
137
|
+
a
|
138
|
+
ensure
|
139
|
+
Unstable.delete_results results
|
140
|
+
end
|
141
|
+
ensure
|
142
|
+
pointer.free
|
143
|
+
end
|
144
|
+
end
|
145
|
+
|
146
|
+
private
|
147
|
+
|
148
|
+
def convert_result(result)
|
149
|
+
language = result[:language_data].read_bytes(result[:language_size])
|
150
|
+
return nil if language == "und"
|
151
|
+
|
152
|
+
cursor = result[:byte_ranges_data]
|
153
|
+
byte_ranges = result[:byte_ranges_size].times.map do
|
154
|
+
info = Unstable::NNetLanguageIdentifier::SpanInfo.new(cursor)
|
155
|
+
cursor += Unstable::NNetLanguageIdentifier::SpanInfo.size
|
156
|
+
SpanInfo.new(info[:start_index], info[:end_index], info[:probability])
|
157
|
+
end
|
85
158
|
|
86
159
|
Result.new(
|
87
|
-
language
|
88
|
-
|
89
|
-
|
90
|
-
|
160
|
+
language.to_sym,
|
161
|
+
result[:probability],
|
162
|
+
result[:reliable?],
|
163
|
+
result[:proportion],
|
164
|
+
byte_ranges)
|
91
165
|
end
|
92
166
|
end
|
93
167
|
|
@@ -95,6 +169,7 @@ module CLD3
|
|
95
169
|
# The model weights are loaded statically.
|
96
170
|
module TaskContextParams
|
97
171
|
# This is an frozen Array object containing symbols.
|
172
|
+
# @type const LANGUAGE_NAMES: untyped
|
98
173
|
LANGUAGE_NAMES = [
|
99
174
|
:eo, :co, :eu, :ta, :de, :mt, :ps, :te, :su, :uz, :'zh-Latn', :ne,
|
100
175
|
:nl, :sw, :sq, :hmn, :ja, :no, :mn, :so, :ko, :kk, :sl, :ig,
|
@@ -108,31 +183,4 @@ module CLD3
|
|
108
183
|
:sn, :yo, :pa, :ku,
|
109
184
|
].freeze
|
110
185
|
end
|
111
|
-
|
112
|
-
module Unstable
|
113
|
-
extend FFI::Library
|
114
|
-
|
115
|
-
ffi_lib File.join(File.expand_path(File.dirname(__FILE__)), "..", "ext", "cld3", "libcld3." + RbConfig::CONFIG["DLEXT"])
|
116
|
-
|
117
|
-
module NNetLanguageIdentifier
|
118
|
-
class Pointer < FFI::AutoPointer
|
119
|
-
def self.release(pointer)
|
120
|
-
Unstable.delete_NNetLanguageIdentifier(pointer)
|
121
|
-
end
|
122
|
-
end
|
123
|
-
|
124
|
-
class Result < FFI::Struct
|
125
|
-
layout :language_data, :pointer, :language_size, :size_t, :probability, :float, :proportion, :float, :reliable?, :bool
|
126
|
-
end
|
127
|
-
end
|
128
|
-
|
129
|
-
attach_function :delete_NNetLanguageIdentifier, [ :pointer ], :void
|
130
|
-
|
131
|
-
attach_function :new_NNetLanguageIdentifier, [ :int, :int ], :pointer
|
132
|
-
|
133
|
-
attach_function :NNetLanguageIdentifier_find_language,
|
134
|
-
[ :pointer, :buffer_in, :size_t ], NNetLanguageIdentifier::Result.by_value
|
135
|
-
end
|
136
|
-
|
137
|
-
private_constant :Unstable
|
138
186
|
end
|
data/sig/cld3.rbs
ADDED
@@ -0,0 +1,65 @@
|
|
1
|
+
# Copyright 2021 Akihiko Odaki <akihiko.odaki@gmail.com>
|
2
|
+
# All Rights Reserved.
|
3
|
+
#
|
4
|
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
5
|
+
# you may not use this file except in compliance with the License.
|
6
|
+
# You may obtain a copy of the License at
|
7
|
+
#
|
8
|
+
# http://www.apache.org/licenses/LICENSE-2.0
|
9
|
+
#
|
10
|
+
# Unless required by applicable law or agreed to in writing, software
|
11
|
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
12
|
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
13
|
+
# See the License for the specific language governing permissions and
|
14
|
+
# limitations under the License.
|
15
|
+
# ==============================================================================
|
16
|
+
|
17
|
+
module CLD3
|
18
|
+
class NNetLanguageIdentifier
|
19
|
+
MIN_NUM_BYTES_TO_CONSIDER: Integer
|
20
|
+
MAX_NUM_BYTES_TO_CONSIDER: Integer
|
21
|
+
MAX_NUM_INPUT_BYTES_TO_CONSIDER: Integer
|
22
|
+
RELIABILITY_THRESHOLD: Float
|
23
|
+
RELIABILITY_HR_BS_THRESHOLD: Float
|
24
|
+
|
25
|
+
class SpanInfo < Struct[Float | Integer]
|
26
|
+
attr_accessor start_index(): Integer
|
27
|
+
attr_accessor end_index(): Integer
|
28
|
+
attr_accessor probability(): Float
|
29
|
+
end
|
30
|
+
|
31
|
+
class Result < Struct[Array[SpanInfo] | Float | TaskContextParams::language_names | bool]
|
32
|
+
attr_accessor language(): TaskContextParams::language_names
|
33
|
+
attr_accessor probability(): Float
|
34
|
+
attr_accessor reliable?(): bool
|
35
|
+
attr_accessor proportion(): Float
|
36
|
+
attr_accessor byte_ranges(): Array[SpanInfo]
|
37
|
+
end
|
38
|
+
|
39
|
+
def initialize: (?Integer, ?Integer) -> void
|
40
|
+
def find_language: (String) -> Result?
|
41
|
+
def find_top_n_most_freq_langs: (String, Integer) -> Array[Result]
|
42
|
+
|
43
|
+
private
|
44
|
+
|
45
|
+
def convert_result: (untyped) -> Result?
|
46
|
+
end
|
47
|
+
|
48
|
+
module TaskContextParams
|
49
|
+
type language_names =
|
50
|
+
:eo | :co | :eu | :ta | :de | :mt | :ps | :te | :su | :uz | :'zh-Latn' | :ne |
|
51
|
+
:nl | :sw | :sq | :hmn | :ja | :no | :mn | :so | :ko | :kk | :sl | :ig |
|
52
|
+
:mr | :th | :zu | :ml | :hr | :bs | :lo | :sd | :cy | :hy | :uk | :pt |
|
53
|
+
:lv | :iw | :cs | :vi | :jv | :be | :km | :mk | :tr | :fy | :am | :zh |
|
54
|
+
:da | :sv | :fi | :ht | :af | :la | :id | :fil | :sm | :ca | :el | :ka |
|
55
|
+
:sr | :it | :sk | :ru | :'ru-Latn' | :bg | :ny | :fa | :haw | :gl | :et |
|
56
|
+
:ms | :gd | :'bg-Latn' | :ha | :is | :ur | :mi | :hi | :bn | :'hi-Latn' | :fr |
|
57
|
+
:yi | :hu | :xh | :my | :tg | :ro | :ar | :lb | :'el-Latn' | :st | :ceb |
|
58
|
+
:kn | :az | :si | :ky | :mg | :en | :gu | :es | :pl | :'ja-Latn' | :ga | :lt |
|
59
|
+
:sn | :yo | :pa | :ku
|
60
|
+
|
61
|
+
LANGUAGE_NAMES: Array[language_names]
|
62
|
+
end
|
63
|
+
|
64
|
+
Unstable: untyped
|
65
|
+
end
|
metadata
CHANGED
@@ -1,14 +1,14 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: cld3
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 3.
|
4
|
+
version: 3.4.3
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- Akihiko Odaki
|
8
|
-
autorequire:
|
8
|
+
autorequire:
|
9
9
|
bindir: bin
|
10
10
|
cert_chain: []
|
11
|
-
date:
|
11
|
+
date: 2021-11-25 00:00:00.000000000 Z
|
12
12
|
dependencies:
|
13
13
|
- !ruby/object:Gem::Dependency
|
14
14
|
name: ffi
|
@@ -19,7 +19,7 @@ dependencies:
|
|
19
19
|
version: 1.1.0
|
20
20
|
- - "<"
|
21
21
|
- !ruby/object:Gem::Version
|
22
|
-
version: 1.
|
22
|
+
version: 1.16.0
|
23
23
|
type: :runtime
|
24
24
|
prerelease: false
|
25
25
|
version_requirements: !ruby/object:Gem::Requirement
|
@@ -29,7 +29,27 @@ dependencies:
|
|
29
29
|
version: 1.1.0
|
30
30
|
- - "<"
|
31
31
|
- !ruby/object:Gem::Version
|
32
|
-
version: 1.
|
32
|
+
version: 1.16.0
|
33
|
+
- !ruby/object:Gem::Dependency
|
34
|
+
name: rbs
|
35
|
+
requirement: !ruby/object:Gem::Requirement
|
36
|
+
requirements:
|
37
|
+
- - ">="
|
38
|
+
- !ruby/object:Gem::Version
|
39
|
+
version: 1.7.0
|
40
|
+
- - "<"
|
41
|
+
- !ruby/object:Gem::Version
|
42
|
+
version: 1.8.0
|
43
|
+
type: :development
|
44
|
+
prerelease: false
|
45
|
+
version_requirements: !ruby/object:Gem::Requirement
|
46
|
+
requirements:
|
47
|
+
- - ">="
|
48
|
+
- !ruby/object:Gem::Version
|
49
|
+
version: 1.7.0
|
50
|
+
- - "<"
|
51
|
+
- !ruby/object:Gem::Version
|
52
|
+
version: 1.8.0
|
33
53
|
- !ruby/object:Gem::Dependency
|
34
54
|
name: rspec
|
35
55
|
requirement: !ruby/object:Gem::Requirement
|
@@ -39,7 +59,7 @@ dependencies:
|
|
39
59
|
version: 3.0.0
|
40
60
|
- - "<"
|
41
61
|
- !ruby/object:Gem::Version
|
42
|
-
version: 3.
|
62
|
+
version: 3.11.0
|
43
63
|
type: :development
|
44
64
|
prerelease: false
|
45
65
|
version_requirements: !ruby/object:Gem::Requirement
|
@@ -49,10 +69,30 @@ dependencies:
|
|
49
69
|
version: 3.0.0
|
50
70
|
- - "<"
|
51
71
|
- !ruby/object:Gem::Version
|
52
|
-
version: 3.
|
72
|
+
version: 3.11.0
|
73
|
+
- !ruby/object:Gem::Dependency
|
74
|
+
name: steep
|
75
|
+
requirement: !ruby/object:Gem::Requirement
|
76
|
+
requirements:
|
77
|
+
- - ">="
|
78
|
+
- !ruby/object:Gem::Version
|
79
|
+
version: 0.46.0
|
80
|
+
- - "<"
|
81
|
+
- !ruby/object:Gem::Version
|
82
|
+
version: 0.47.0
|
83
|
+
type: :development
|
84
|
+
prerelease: false
|
85
|
+
version_requirements: !ruby/object:Gem::Requirement
|
86
|
+
requirements:
|
87
|
+
- - ">="
|
88
|
+
- !ruby/object:Gem::Version
|
89
|
+
version: 0.46.0
|
90
|
+
- - "<"
|
91
|
+
- !ruby/object:Gem::Version
|
92
|
+
version: 0.47.0
|
53
93
|
description: Compact Language Detector v3 (CLD3) is a neural network model for language
|
54
94
|
identification.
|
55
|
-
email: akihiko.odaki
|
95
|
+
email: akihiko.odaki@gmail.com
|
56
96
|
executables: []
|
57
97
|
extensions:
|
58
98
|
- ext/cld3/extconf.rb
|
@@ -106,6 +146,7 @@ files:
|
|
106
146
|
- ext/cld3/language_identifier_features.cc
|
107
147
|
- ext/cld3/language_identifier_features.h
|
108
148
|
- ext/cld3/language_identifier_features.o
|
149
|
+
- ext/cld3/libcld3.def
|
109
150
|
- ext/cld3/libcld3.so
|
110
151
|
- ext/cld3/mkmf.log
|
111
152
|
- ext/cld3/nnet_language_identifier.cc
|
@@ -159,11 +200,13 @@ files:
|
|
159
200
|
- ext/cld3/workspace.h
|
160
201
|
- ext/cld3/workspace.o
|
161
202
|
- lib/cld3.rb
|
203
|
+
- lib/cld3/unstable.rb
|
204
|
+
- sig/cld3.rbs
|
162
205
|
homepage: https://github.com/akihikodaki/cld3-ruby
|
163
206
|
licenses:
|
164
207
|
- Apache-2.0
|
165
208
|
metadata: {}
|
166
|
-
post_install_message:
|
209
|
+
post_install_message:
|
167
210
|
rdoc_options: []
|
168
211
|
require_paths:
|
169
212
|
- lib
|
@@ -171,18 +214,18 @@ required_ruby_version: !ruby/object:Gem::Requirement
|
|
171
214
|
requirements:
|
172
215
|
- - ">="
|
173
216
|
- !ruby/object:Gem::Version
|
174
|
-
version: 2.
|
217
|
+
version: 2.6.0
|
175
218
|
- - "<"
|
176
219
|
- !ruby/object:Gem::Version
|
177
|
-
version: 2.
|
220
|
+
version: 3.2.0
|
178
221
|
required_rubygems_version: !ruby/object:Gem::Requirement
|
179
222
|
requirements:
|
180
223
|
- - ">="
|
181
224
|
- !ruby/object:Gem::Version
|
182
225
|
version: '0'
|
183
226
|
requirements: []
|
184
|
-
rubygems_version: 3.
|
185
|
-
signing_key:
|
227
|
+
rubygems_version: 3.2.22
|
228
|
+
signing_key:
|
186
229
|
specification_version: 4
|
187
230
|
summary: Compact Language Detector v3 (CLD3)
|
188
231
|
test_files: []
|