hunspell-i18n 0.2.0
Sign up to get free protection for your applications and to get access to all the features.
- data/.rspec +2 -0
- data/README.md +16 -0
- data/Rakefile +29 -0
- data/VERSION +1 -0
- data/ext/extconf.rb +7 -0
- data/ext/hunspell_wrap.c +129 -0
- data/hunspell-i18n.gemspec +44 -0
- data/spec/hunspell_spec.rb +33 -0
- data/spec/spec_helper.rb +1 -0
- metadata +62 -0
data/.rspec
ADDED
data/README.md
ADDED
@@ -0,0 +1,16 @@
|
|
1
|
+
# Ruby bindings for Hunspell 1.2.x
|
2
|
+
|
3
|
+
This version expects to receive the word in UTF-8 encoding, and it will
|
4
|
+
then use iconv behind the scenes in order to convert to and from
|
5
|
+
the dictionary encoding.
|
6
|
+
|
7
|
+
## Usage
|
8
|
+
|
9
|
+
prefix = '/usr/share/myspell/' # Fedora
|
10
|
+
speller = Hunspell.new(prefix + 'en_US.aff', prefix + 'en_US.dic')
|
11
|
+
speller.valid?('dog') # -> true
|
12
|
+
speller.valid?('dpg') # -> false
|
13
|
+
speller.suggest('dpg') # -> ["dog", "dg", "pg", "deg", "dig", ...]
|
14
|
+
|
15
|
+
|
16
|
+
Copyright 2011 Roman Shterenzon, released under the MIT license
|
data/Rakefile
ADDED
@@ -0,0 +1,29 @@
|
|
1
|
+
require 'rake'
|
2
|
+
require 'rake/rdoctask'
|
3
|
+
|
4
|
+
desc 'Default: create gemspec'
|
5
|
+
task :default => :gemspec
|
6
|
+
|
7
|
+
desc 'Generate documentation'
|
8
|
+
Rake::RDocTask.new(:rdoc) do |rdoc|
|
9
|
+
rdoc.rdoc_dir = 'rdoc'
|
10
|
+
rdoc.title = 'Hunspell'
|
11
|
+
rdoc.options << '--line-numbers' << '--inline-source'
|
12
|
+
rdoc.rdoc_files.include('README')
|
13
|
+
rdoc.rdoc_files.include('lib/**/*.rb')
|
14
|
+
end
|
15
|
+
|
16
|
+
begin
|
17
|
+
require 'jeweler'
|
18
|
+
Jeweler::Tasks.new do |gem|
|
19
|
+
gem.name = "hunspell-i18n"
|
20
|
+
gem.summary = %Q{Ruby bindings for libhunspell-1.2 with i18n support}
|
21
|
+
gem.description = gem.summary
|
22
|
+
gem.email = "romanbsd@yahoo.com"
|
23
|
+
gem.homepage = "http://github.com/romanbsd/hunspell"
|
24
|
+
gem.authors = ["Roman Shterenzon"]
|
25
|
+
gem.extensions = ['ext/extconf.rb']
|
26
|
+
end
|
27
|
+
Jeweler::GemcutterTasks.new
|
28
|
+
rescue LoadError
|
29
|
+
end
|
data/VERSION
ADDED
@@ -0,0 +1 @@
|
|
1
|
+
0.2.0
|
data/ext/extconf.rb
ADDED
data/ext/hunspell_wrap.c
ADDED
@@ -0,0 +1,129 @@
|
|
1
|
+
#include <ruby.h>
|
2
|
+
#include <hunspell/hunspell.h>
|
3
|
+
|
4
|
+
static VALUE rb_cHunspell;
|
5
|
+
static VALUE rb_cIconv;
|
6
|
+
static VALUE enc_utf8;
|
7
|
+
static VALUE enc_iso8859_1;
|
8
|
+
|
9
|
+
static void dict_destroy(Hunhandle *handle) {
|
10
|
+
Hunspell_destroy(handle);
|
11
|
+
}
|
12
|
+
|
13
|
+
/*
|
14
|
+
* call-seq:
|
15
|
+
* Hunspell.new(affix_path, dic_path)
|
16
|
+
*
|
17
|
+
* Instantiate Hunspell with paths pointing to affix and dictionary files.
|
18
|
+
*
|
19
|
+
* Example:
|
20
|
+
* speller = Hunspell.new("/usr/share/myspell/en_US.aff", "/usr/share/myspell/en_US.dic")
|
21
|
+
*
|
22
|
+
*/
|
23
|
+
static VALUE dict_init(VALUE self, VALUE affix_path, VALUE dic_path) {
|
24
|
+
VALUE affpath = StringValue(affix_path), dpath = StringValue(dic_path);
|
25
|
+
Hunhandle *handle = Hunspell_create(RSTRING_PTR(affpath), RSTRING_PTR(dpath));
|
26
|
+
if (!handle) {
|
27
|
+
rb_raise(rb_eRuntimeError, "Failed to initialize Hunspell handle.");
|
28
|
+
}
|
29
|
+
VALUE dict = Data_Wrap_Struct(rb_cHunspell, NULL, dict_destroy, handle);
|
30
|
+
rb_iv_set(self, "dict", dict);
|
31
|
+
return self;
|
32
|
+
}
|
33
|
+
|
34
|
+
static inline Hunhandle* get_handle(VALUE self) {
|
35
|
+
Hunhandle *handle = NULL;
|
36
|
+
Data_Get_Struct(rb_iv_get(self, "dict"), Hunhandle, handle);
|
37
|
+
if (!handle) {
|
38
|
+
rb_raise(rb_eRuntimeError, "Hunspell handle not found.");
|
39
|
+
}
|
40
|
+
return handle;
|
41
|
+
}
|
42
|
+
|
43
|
+
/*
|
44
|
+
* call-seq:
|
45
|
+
* hunspell.encoding -> string
|
46
|
+
*
|
47
|
+
* Returns the encoding of the dictionary.
|
48
|
+
*/
|
49
|
+
static VALUE wrap_encoding(VALUE self) {
|
50
|
+
Hunhandle *handle = get_handle(self);
|
51
|
+
char *enc = Hunspell_get_dic_encoding(handle);
|
52
|
+
return rb_str_new2(enc);
|
53
|
+
}
|
54
|
+
|
55
|
+
static VALUE recode_if_needed(VALUE self, VALUE str, int dir) {
|
56
|
+
VALUE enc = wrap_encoding(self);
|
57
|
+
VALUE from, to;
|
58
|
+
|
59
|
+
if (dir == 0) {
|
60
|
+
from = enc_utf8;
|
61
|
+
to = enc;
|
62
|
+
} else {
|
63
|
+
from = enc;
|
64
|
+
to = enc_utf8;
|
65
|
+
}
|
66
|
+
if ( rb_str_equal(enc, enc_iso8859_1) == Qfalse && rb_str_equal(enc, enc_utf8) == Qfalse) {
|
67
|
+
rb_funcall(str, rb_intern("force_encoding"), 1, enc_utf8);
|
68
|
+
return rb_funcall(rb_cIconv, rb_intern("conv"), 3, to, from, str);
|
69
|
+
} else {
|
70
|
+
return str;
|
71
|
+
}
|
72
|
+
}
|
73
|
+
|
74
|
+
/*
|
75
|
+
* call-seq:
|
76
|
+
* hunspell.suggest(misspeledword) -> ary
|
77
|
+
*
|
78
|
+
* Returns a list of suggestions.
|
79
|
+
*/
|
80
|
+
static VALUE wrap_suggest(VALUE self, VALUE word) {
|
81
|
+
VALUE str = recode_if_needed(self, StringValue(word), 0);
|
82
|
+
VALUE res;
|
83
|
+
char** slst = NULL;
|
84
|
+
int i, count = 0;
|
85
|
+
Hunhandle *handle = get_handle(self);
|
86
|
+
|
87
|
+
count = Hunspell_suggest(handle, &slst, RSTRING_PTR(str));
|
88
|
+
|
89
|
+
res = rb_ary_new2(count);
|
90
|
+
for (i=0; i<count; ++i) {
|
91
|
+
rb_ary_push(res, recode_if_needed(self, rb_str_new2(slst[i]), 1));
|
92
|
+
free(slst[i]);
|
93
|
+
}
|
94
|
+
|
95
|
+
if (slst) {
|
96
|
+
free(slst);
|
97
|
+
}
|
98
|
+
|
99
|
+
return res;
|
100
|
+
}
|
101
|
+
|
102
|
+
/*
|
103
|
+
* call-seq:
|
104
|
+
* hunspell.valid?(word) -> bool
|
105
|
+
*
|
106
|
+
* Checks if the word is in the dictionary.
|
107
|
+
*/
|
108
|
+
static VALUE wrap_check(VALUE self, VALUE word) {
|
109
|
+
VALUE str = recode_if_needed(self, StringValue(word), 0);
|
110
|
+
Hunhandle *handle = get_handle(self);
|
111
|
+
int rc = Hunspell_spell(handle, RSTRING_PTR(str));
|
112
|
+
return rc == 0 ? Qfalse : Qtrue;
|
113
|
+
}
|
114
|
+
|
115
|
+
|
116
|
+
void Init_hunspell(void) {
|
117
|
+
rb_cHunspell = rb_define_class("Hunspell", rb_cObject);
|
118
|
+
rb_define_method(rb_cHunspell, "initialize", dict_init, 2);
|
119
|
+
//rb_define_method(rb_cHunspell, "check?", wrap_check, 1);
|
120
|
+
rb_define_method(rb_cHunspell, "valid?", wrap_check, 1);
|
121
|
+
rb_define_method(rb_cHunspell, "suggest", wrap_suggest, 1);
|
122
|
+
rb_define_method(rb_cHunspell, "encoding", wrap_encoding, 0);
|
123
|
+
rb_require("iconv");
|
124
|
+
rb_cIconv = rb_const_get(rb_cObject, rb_intern("Iconv"));
|
125
|
+
enc_iso8859_1 = rb_str_new2("ISO8859-1");
|
126
|
+
enc_utf8 = rb_str_new2("UTF-8");
|
127
|
+
rb_global_variable(&enc_iso8859_1);
|
128
|
+
rb_global_variable(&enc_utf8);
|
129
|
+
}
|
@@ -0,0 +1,44 @@
|
|
1
|
+
# Generated by jeweler
|
2
|
+
# DO NOT EDIT THIS FILE DIRECTLY
|
3
|
+
# Instead, edit Jeweler::Tasks in Rakefile, and run 'rake gemspec'
|
4
|
+
# -*- encoding: utf-8 -*-
|
5
|
+
|
6
|
+
Gem::Specification.new do |s|
|
7
|
+
s.name = "hunspell-i18n"
|
8
|
+
s.version = "0.2.0"
|
9
|
+
|
10
|
+
s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
|
11
|
+
s.authors = ["Roman Shterenzon"]
|
12
|
+
s.date = "2012-01-21"
|
13
|
+
s.description = "Ruby bindings for libhunspell-1.2 with i18n support"
|
14
|
+
s.email = "romanbsd@yahoo.com"
|
15
|
+
s.extensions = ["ext/extconf.rb"]
|
16
|
+
s.extra_rdoc_files = [
|
17
|
+
"README.md"
|
18
|
+
]
|
19
|
+
s.files = [
|
20
|
+
".rspec",
|
21
|
+
"README.md",
|
22
|
+
"Rakefile",
|
23
|
+
"VERSION",
|
24
|
+
"ext/extconf.rb",
|
25
|
+
"ext/hunspell_wrap.c",
|
26
|
+
"hunspell-i18n.gemspec",
|
27
|
+
"spec/hunspell_spec.rb",
|
28
|
+
"spec/spec_helper.rb"
|
29
|
+
]
|
30
|
+
s.homepage = "http://github.com/romanbsd/hunspell"
|
31
|
+
s.require_paths = ["lib"]
|
32
|
+
s.rubygems_version = "1.8.15"
|
33
|
+
s.summary = "Ruby bindings for libhunspell-1.2 with i18n support"
|
34
|
+
|
35
|
+
if s.respond_to? :specification_version then
|
36
|
+
s.specification_version = 3
|
37
|
+
|
38
|
+
if Gem::Version.new(Gem::VERSION) >= Gem::Version.new('1.2.0') then
|
39
|
+
else
|
40
|
+
end
|
41
|
+
else
|
42
|
+
end
|
43
|
+
end
|
44
|
+
|
@@ -0,0 +1,33 @@
|
|
1
|
+
require 'spec_helper'
|
2
|
+
|
3
|
+
describe Hunspell do
|
4
|
+
|
5
|
+
before :all do
|
6
|
+
@speller = Hunspell.new('en_US.aff', 'en_US.dic')
|
7
|
+
end
|
8
|
+
|
9
|
+
it "should check if a word is valid" do
|
10
|
+
@speller.should be_valid('dog')
|
11
|
+
@speller.should_not be_valid('dxg')
|
12
|
+
end
|
13
|
+
|
14
|
+
it "survives GC" do
|
15
|
+
10.times do
|
16
|
+
@speller.suggest('dxg')
|
17
|
+
end
|
18
|
+
GC.start
|
19
|
+
10.times do
|
20
|
+
@speller.suggest('dxg')
|
21
|
+
end
|
22
|
+
end
|
23
|
+
|
24
|
+
it "should suggest alternate spellings for words" do
|
25
|
+
@speller.suggest('arbitrage').should == %w[
|
26
|
+
arbitrage
|
27
|
+
arbitrages
|
28
|
+
arbitrager
|
29
|
+
arbitraged
|
30
|
+
arbitrate
|
31
|
+
]
|
32
|
+
end
|
33
|
+
end
|
data/spec/spec_helper.rb
ADDED
@@ -0,0 +1 @@
|
|
1
|
+
require 'hunspell'
|
metadata
ADDED
@@ -0,0 +1,62 @@
|
|
1
|
+
--- !ruby/object:Gem::Specification
|
2
|
+
name: hunspell-i18n
|
3
|
+
version: !ruby/object:Gem::Version
|
4
|
+
prerelease:
|
5
|
+
version: 0.2.0
|
6
|
+
platform: ruby
|
7
|
+
authors:
|
8
|
+
- Roman Shterenzon
|
9
|
+
autorequire:
|
10
|
+
bindir: bin
|
11
|
+
cert_chain: []
|
12
|
+
|
13
|
+
date: 2012-01-21 00:00:00 Z
|
14
|
+
dependencies: []
|
15
|
+
|
16
|
+
description: Ruby bindings for libhunspell-1.2 with i18n support
|
17
|
+
email: romanbsd@yahoo.com
|
18
|
+
executables: []
|
19
|
+
|
20
|
+
extensions:
|
21
|
+
- ext/extconf.rb
|
22
|
+
extra_rdoc_files:
|
23
|
+
- README.md
|
24
|
+
files:
|
25
|
+
- .rspec
|
26
|
+
- README.md
|
27
|
+
- Rakefile
|
28
|
+
- VERSION
|
29
|
+
- ext/extconf.rb
|
30
|
+
- ext/hunspell_wrap.c
|
31
|
+
- hunspell-i18n.gemspec
|
32
|
+
- spec/hunspell_spec.rb
|
33
|
+
- spec/spec_helper.rb
|
34
|
+
homepage: http://github.com/romanbsd/hunspell
|
35
|
+
licenses: []
|
36
|
+
|
37
|
+
post_install_message:
|
38
|
+
rdoc_options: []
|
39
|
+
|
40
|
+
require_paths:
|
41
|
+
- lib
|
42
|
+
required_ruby_version: !ruby/object:Gem::Requirement
|
43
|
+
none: false
|
44
|
+
requirements:
|
45
|
+
- - ">="
|
46
|
+
- !ruby/object:Gem::Version
|
47
|
+
version: "0"
|
48
|
+
required_rubygems_version: !ruby/object:Gem::Requirement
|
49
|
+
none: false
|
50
|
+
requirements:
|
51
|
+
- - ">="
|
52
|
+
- !ruby/object:Gem::Version
|
53
|
+
version: "0"
|
54
|
+
requirements: []
|
55
|
+
|
56
|
+
rubyforge_project:
|
57
|
+
rubygems_version: 1.8.15
|
58
|
+
signing_key:
|
59
|
+
specification_version: 3
|
60
|
+
summary: Ruby bindings for libhunspell-1.2 with i18n support
|
61
|
+
test_files: []
|
62
|
+
|