ruby-jdict 0.0.7 → 0.0.8
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/LICENSING +28 -28
- data/README.md +18 -20
- data/Rakefile +41 -30
- data/examples/query.rb +19 -22
- data/lib/ruby-jdict.rb +14 -0
- data/lib/{constants.rb → ruby-jdict/constants.rb} +73 -64
- data/lib/ruby-jdict/convert.rb +33 -0
- data/lib/ruby-jdict/dictionary.rb +59 -0
- data/lib/ruby-jdict/index.rb +151 -0
- data/lib/ruby-jdict/indexer/dictionary_indexer.rb +28 -0
- data/lib/ruby-jdict/indexer/libxml_dictionary_indexer.rb +164 -0
- data/lib/ruby-jdict/indexer/nokogiri_dictionary_indexer.rb +60 -0
- data/lib/ruby-jdict/jdict.rb +2 -0
- data/lib/ruby-jdict/models/entry.rb +64 -0
- data/lib/ruby-jdict/models/sense.rb +81 -0
- data/lib/ruby-jdict/version.rb +3 -3
- data/spec/convert_spec.rb +27 -0
- data/spec/dictionary_spec.rb +113 -113
- data/spec/entry_spec.rb +25 -0
- data/spec/fixtures/feeds/sample_entry.xml +32 -32
- data/spec/index_spec.rb +82 -84
- data/spec/spec_helper.rb +49 -49
- metadata +35 -36
- data/examples/lst.txt +0 -4
- data/lib/configuration.rb +0 -34
- data/lib/dictionaries/jmdict.rb +0 -38
- data/lib/dictionary.rb +0 -90
- data/lib/downloader.rb +0 -42
- data/lib/entry.rb +0 -101
- data/lib/index.rb +0 -305
- data/lib/jdict.rb +0 -20
- data/lib/kana.rb +0 -4
- data/lib/kanji.rb +0 -4
- data/lib/sense.rb +0 -28
- data/lib/unicode.rb +0 -63
- data/spec/configuration_spec.rb +0 -20
- data/spec/jmdict_spec.rb +0 -19
data/spec/spec_helper.rb
CHANGED
@@ -1,50 +1,50 @@
|
|
1
|
-
require 'rubygems' #use gems
|
2
|
-
require 'bundler/setup' #load up the bundled environment
|
3
|
-
# require 'spec' #test framework
|
4
|
-
|
5
|
-
$DEBUG = true
|
6
|
-
|
7
|
-
BASE_PATH = File.dirname(__FILE__) + '/..'
|
8
|
-
INDEX_PATH = BASE_PATH + '/test_index'
|
9
|
-
JMDICT_PATH = BASE_PATH + '/dictionaries/JMdict'
|
10
|
-
|
11
|
-
##
|
12
|
-
# rSpec Hash additions.
|
13
|
-
#
|
14
|
-
# From
|
15
|
-
# * http://wincent.com/knowledge-base/Fixtures_considered_harmful%3F
|
16
|
-
# * Neil Rahilly
|
17
|
-
|
18
|
-
class Hash
|
19
|
-
|
20
|
-
##
|
21
|
-
# Filter keys out of a Hash.
|
22
|
-
#
|
23
|
-
# { :a => 1, :b => 2, :c => 3 }.except(:a)
|
24
|
-
# => { :b => 2, :c => 3 }
|
25
|
-
|
26
|
-
def except(*keys)
|
27
|
-
self.reject { |k,v| keys.include?(k || k.to_sym) }
|
28
|
-
end
|
29
|
-
|
30
|
-
##
|
31
|
-
# Override some keys.
|
32
|
-
#
|
33
|
-
# { :a => 1, :b => 2, :c => 3 }.with(:a => 4)
|
34
|
-
# => { :a => 4, :b => 2, :c => 3 }
|
35
|
-
|
36
|
-
def with(overrides = {})
|
37
|
-
self.merge overrides
|
38
|
-
end
|
39
|
-
|
40
|
-
##
|
41
|
-
# Returns a Hash with only the pairs identified by +keys+.
|
42
|
-
#
|
43
|
-
# { :a => 1, :b => 2, :c => 3 }.only(:a)
|
44
|
-
# => { :a => 1 }
|
45
|
-
|
46
|
-
def only(*keys)
|
47
|
-
self.reject { |k,v| !keys.include?(k || k.to_sym) }
|
48
|
-
end
|
49
|
-
|
1
|
+
require 'rubygems' #use gems
|
2
|
+
require 'bundler/setup' #load up the bundled environment
|
3
|
+
# require 'spec' #test framework
|
4
|
+
|
5
|
+
$DEBUG = true
|
6
|
+
|
7
|
+
BASE_PATH = File.dirname(__FILE__) + '/..'
|
8
|
+
INDEX_PATH = BASE_PATH + '/test_index'
|
9
|
+
JMDICT_PATH = BASE_PATH + '/dictionaries/JMdict'
|
10
|
+
|
11
|
+
##
|
12
|
+
# rSpec Hash additions.
|
13
|
+
#
|
14
|
+
# From
|
15
|
+
# * http://wincent.com/knowledge-base/Fixtures_considered_harmful%3F
|
16
|
+
# * Neil Rahilly
|
17
|
+
|
18
|
+
class Hash
|
19
|
+
|
20
|
+
##
|
21
|
+
# Filter keys out of a Hash.
|
22
|
+
#
|
23
|
+
# { :a => 1, :b => 2, :c => 3 }.except(:a)
|
24
|
+
# => { :b => 2, :c => 3 }
|
25
|
+
|
26
|
+
def except(*keys)
|
27
|
+
self.reject { |k,v| keys.include?(k || k.to_sym) }
|
28
|
+
end
|
29
|
+
|
30
|
+
##
|
31
|
+
# Override some keys.
|
32
|
+
#
|
33
|
+
# { :a => 1, :b => 2, :c => 3 }.with(:a => 4)
|
34
|
+
# => { :a => 4, :b => 2, :c => 3 }
|
35
|
+
|
36
|
+
def with(overrides = {})
|
37
|
+
self.merge overrides
|
38
|
+
end
|
39
|
+
|
40
|
+
##
|
41
|
+
# Returns a Hash with only the pairs identified by +keys+.
|
42
|
+
#
|
43
|
+
# { :a => 1, :b => 2, :c => 3 }.only(:a)
|
44
|
+
# => { :a => 1 }
|
45
|
+
|
46
|
+
def only(*keys)
|
47
|
+
self.reject { |k,v| !keys.include?(k || k.to_sym) }
|
48
|
+
end
|
49
|
+
|
50
50
|
end
|
metadata
CHANGED
@@ -1,85 +1,85 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: ruby-jdict
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.0.
|
4
|
+
version: 0.0.8
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- Ian Pickering
|
8
8
|
autorequire:
|
9
9
|
bindir: bin
|
10
10
|
cert_chain: []
|
11
|
-
date:
|
11
|
+
date: 2019-02-06 00:00:00.000000000 Z
|
12
12
|
dependencies:
|
13
13
|
- !ruby/object:Gem::Dependency
|
14
|
-
name:
|
14
|
+
name: rake
|
15
15
|
requirement: !ruby/object:Gem::Requirement
|
16
16
|
requirements:
|
17
17
|
- - "~>"
|
18
18
|
- !ruby/object:Gem::Version
|
19
|
-
version:
|
19
|
+
version: '12.3'
|
20
20
|
type: :runtime
|
21
21
|
prerelease: false
|
22
22
|
version_requirements: !ruby/object:Gem::Requirement
|
23
23
|
requirements:
|
24
24
|
- - "~>"
|
25
25
|
- !ruby/object:Gem::Version
|
26
|
-
version:
|
26
|
+
version: '12.3'
|
27
27
|
- !ruby/object:Gem::Dependency
|
28
|
-
name:
|
28
|
+
name: nokogiri
|
29
29
|
requirement: !ruby/object:Gem::Requirement
|
30
30
|
requirements:
|
31
31
|
- - "~>"
|
32
32
|
- !ruby/object:Gem::Version
|
33
|
-
version: 1.
|
33
|
+
version: '1.10'
|
34
34
|
type: :runtime
|
35
35
|
prerelease: false
|
36
36
|
version_requirements: !ruby/object:Gem::Requirement
|
37
37
|
requirements:
|
38
38
|
- - "~>"
|
39
39
|
- !ruby/object:Gem::Version
|
40
|
-
version: 1.
|
40
|
+
version: '1.10'
|
41
41
|
- !ruby/object:Gem::Dependency
|
42
|
-
name:
|
42
|
+
name: amalgalite
|
43
43
|
requirement: !ruby/object:Gem::Requirement
|
44
44
|
requirements:
|
45
45
|
- - "~>"
|
46
46
|
- !ruby/object:Gem::Version
|
47
|
-
version: 1.
|
47
|
+
version: '1.6'
|
48
48
|
type: :runtime
|
49
49
|
prerelease: false
|
50
50
|
version_requirements: !ruby/object:Gem::Requirement
|
51
51
|
requirements:
|
52
52
|
- - "~>"
|
53
53
|
- !ruby/object:Gem::Version
|
54
|
-
version: 1.
|
54
|
+
version: '1.6'
|
55
55
|
- !ruby/object:Gem::Dependency
|
56
|
-
name:
|
56
|
+
name: rsync
|
57
57
|
requirement: !ruby/object:Gem::Requirement
|
58
58
|
requirements:
|
59
|
-
- - "
|
59
|
+
- - "~>"
|
60
60
|
- !ruby/object:Gem::Version
|
61
|
-
version: '0'
|
62
|
-
type: :
|
61
|
+
version: '1.0'
|
62
|
+
type: :runtime
|
63
63
|
prerelease: false
|
64
64
|
version_requirements: !ruby/object:Gem::Requirement
|
65
65
|
requirements:
|
66
|
-
- - "
|
66
|
+
- - "~>"
|
67
67
|
- !ruby/object:Gem::Version
|
68
|
-
version: '0'
|
68
|
+
version: '1.0'
|
69
69
|
- !ruby/object:Gem::Dependency
|
70
70
|
name: rspec
|
71
71
|
requirement: !ruby/object:Gem::Requirement
|
72
72
|
requirements:
|
73
73
|
- - "~>"
|
74
74
|
- !ruby/object:Gem::Version
|
75
|
-
version: 3.
|
75
|
+
version: '3.8'
|
76
76
|
type: :development
|
77
77
|
prerelease: false
|
78
78
|
version_requirements: !ruby/object:Gem::Requirement
|
79
79
|
requirements:
|
80
80
|
- - "~>"
|
81
81
|
- !ruby/object:Gem::Version
|
82
|
-
version: 3.
|
82
|
+
version: '3.8'
|
83
83
|
description:
|
84
84
|
email:
|
85
85
|
- ipickering2@gmail.com
|
@@ -90,29 +90,28 @@ files:
|
|
90
90
|
- LICENSING
|
91
91
|
- README.md
|
92
92
|
- Rakefile
|
93
|
-
- examples/lst.txt
|
94
93
|
- examples/query.rb
|
95
|
-
- lib/
|
96
|
-
- lib/constants.rb
|
97
|
-
- lib/
|
98
|
-
- lib/dictionary.rb
|
99
|
-
- lib/
|
100
|
-
- lib/
|
101
|
-
- lib/
|
102
|
-
- lib/jdict.rb
|
103
|
-
- lib/
|
104
|
-
- lib/
|
94
|
+
- lib/ruby-jdict.rb
|
95
|
+
- lib/ruby-jdict/constants.rb
|
96
|
+
- lib/ruby-jdict/convert.rb
|
97
|
+
- lib/ruby-jdict/dictionary.rb
|
98
|
+
- lib/ruby-jdict/index.rb
|
99
|
+
- lib/ruby-jdict/indexer/dictionary_indexer.rb
|
100
|
+
- lib/ruby-jdict/indexer/libxml_dictionary_indexer.rb
|
101
|
+
- lib/ruby-jdict/indexer/nokogiri_dictionary_indexer.rb
|
102
|
+
- lib/ruby-jdict/jdict.rb
|
103
|
+
- lib/ruby-jdict/models/entry.rb
|
104
|
+
- lib/ruby-jdict/models/sense.rb
|
105
105
|
- lib/ruby-jdict/version.rb
|
106
|
-
-
|
107
|
-
- lib/unicode.rb
|
108
|
-
- spec/configuration_spec.rb
|
106
|
+
- spec/convert_spec.rb
|
109
107
|
- spec/dictionary_spec.rb
|
108
|
+
- spec/entry_spec.rb
|
110
109
|
- spec/fixtures/feeds/sample_entry.xml
|
111
110
|
- spec/index_spec.rb
|
112
|
-
- spec/jmdict_spec.rb
|
113
111
|
- spec/spec_helper.rb
|
114
112
|
homepage: https://github.com/Ruin0x11/ruby-jdict
|
115
|
-
licenses:
|
113
|
+
licenses:
|
114
|
+
- MIT
|
116
115
|
metadata: {}
|
117
116
|
post_install_message:
|
118
117
|
rdoc_options: []
|
@@ -130,7 +129,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
|
|
130
129
|
version: '0'
|
131
130
|
requirements: []
|
132
131
|
rubyforge_project:
|
133
|
-
rubygems_version: 2.
|
132
|
+
rubygems_version: 2.6.14.3
|
134
133
|
signing_key:
|
135
134
|
specification_version: 4
|
136
135
|
summary: Ruby gem for accessing Jim Breen's Japanese dictionaries
|
data/examples/lst.txt
DELETED
data/lib/configuration.rb
DELETED
@@ -1,34 +0,0 @@
|
|
1
|
-
require 'constants'
|
2
|
-
|
3
|
-
module JDict
|
4
|
-
class Configuration
|
5
|
-
attr_accessor :dictionary_path, :num_results, :language, :debug,
|
6
|
-
:jmdict_url, :jmdict_rsync_url
|
7
|
-
|
8
|
-
BASE_PATH = ENV["HOME"]
|
9
|
-
DICT_PATH = File.join(BASE_PATH, '.dicts')
|
10
|
-
|
11
|
-
JMDICT_URL = "ftp://ftp.monash.edu.au/pub/nihongo/JMdict.gz"
|
12
|
-
JMDICT_RSYNC_URL = "rsync://ftp.monash.edu.au/nihongo/JMdict"
|
13
|
-
|
14
|
-
def initialize
|
15
|
-
# directory containing dictionary files
|
16
|
-
@dictionary_path = DICT_PATH
|
17
|
-
|
18
|
-
# maximum results to return from searching
|
19
|
-
@num_results = 50
|
20
|
-
|
21
|
-
# language to return search results in
|
22
|
-
@language = JDict::JMDictConstants::Languages::ENGLISH
|
23
|
-
|
24
|
-
# limit number of entries indexed, rebuild index on instantiation
|
25
|
-
@debug = false
|
26
|
-
|
27
|
-
# url to retrieve JMDict from
|
28
|
-
@jmdict_url = JMDICT_URL
|
29
|
-
|
30
|
-
# url to rsync JMDict from
|
31
|
-
@jmdict_rsync_url = JMDICT_RSYNC_URL
|
32
|
-
end
|
33
|
-
end
|
34
|
-
end
|
data/lib/dictionaries/jmdict.rb
DELETED
@@ -1,38 +0,0 @@
|
|
1
|
-
require 'jdict'
|
2
|
-
require 'dictionary'
|
3
|
-
require 'downloader'
|
4
|
-
|
5
|
-
module JDict
|
6
|
-
class JMDict < Dictionary
|
7
|
-
def dict_url
|
8
|
-
JDict.config.jmdict_url
|
9
|
-
end
|
10
|
-
|
11
|
-
def dict_file
|
12
|
-
'JMDict'
|
13
|
-
end
|
14
|
-
|
15
|
-
private
|
16
|
-
|
17
|
-
def initialize
|
18
|
-
dict_path = JDict.config.dictionary_path
|
19
|
-
@downloader = JMDictDownloader.new
|
20
|
-
|
21
|
-
super(dict_path)
|
22
|
-
end
|
23
|
-
end
|
24
|
-
|
25
|
-
class JMDictDownloader < Downloader
|
26
|
-
def download(dir)
|
27
|
-
url = JDict.config.jmdict_url
|
28
|
-
full_path = retrieve_file(url, dir)
|
29
|
-
gunzip(full_path)
|
30
|
-
end
|
31
|
-
|
32
|
-
def sync(dir)
|
33
|
-
dict_path = File.join(dir, 'JMdict')
|
34
|
-
rsync_url = JDict.config.jmdict_rsync_url
|
35
|
-
rsync(rsync_url, dict_path)
|
36
|
-
end
|
37
|
-
end
|
38
|
-
end
|
data/lib/dictionary.rb
DELETED
@@ -1,90 +0,0 @@
|
|
1
|
-
require 'jdict'
|
2
|
-
require 'index'
|
3
|
-
|
4
|
-
module JDict
|
5
|
-
class Dictionary
|
6
|
-
def initialize(path)
|
7
|
-
@dictionary_path = File.join(path, self.dict_file)
|
8
|
-
@entries = []
|
9
|
-
|
10
|
-
prompt_for_download unless File.exists? @dictionary_path
|
11
|
-
|
12
|
-
@index = DictIndex.new(@dictionary_path)
|
13
|
-
end
|
14
|
-
|
15
|
-
def size
|
16
|
-
@entries.size
|
17
|
-
end
|
18
|
-
|
19
|
-
def loaded?
|
20
|
-
@index.built?
|
21
|
-
end
|
22
|
-
|
23
|
-
def download(dir)
|
24
|
-
@downloader.download(dir)
|
25
|
-
end
|
26
|
-
|
27
|
-
# abstract method
|
28
|
-
def dict_url
|
29
|
-
""
|
30
|
-
end
|
31
|
-
|
32
|
-
def dict_file
|
33
|
-
""
|
34
|
-
end
|
35
|
-
|
36
|
-
def prompt_for_download
|
37
|
-
base_dir = File.dirname(@dictionary_path)
|
38
|
-
|
39
|
-
puts "Dictionary not found at #{@dictionary_path}.\n" \
|
40
|
-
"Would you like to download the dictionary from the URL\n" \
|
41
|
-
" #{self.dict_url}\n" \
|
42
|
-
"into the folder\n" \
|
43
|
-
" #{base_dir}? [y/N]"
|
44
|
-
|
45
|
-
response = case $stdin.getch
|
46
|
-
when "Y" then true
|
47
|
-
when "y" then true
|
48
|
-
else false
|
49
|
-
end
|
50
|
-
|
51
|
-
unless response
|
52
|
-
puts "Dictionary not downloaded."
|
53
|
-
exit
|
54
|
-
end
|
55
|
-
|
56
|
-
FileUtils.mkdir_p(base_dir)
|
57
|
-
|
58
|
-
puts "Downloading dictionary..."
|
59
|
-
download(base_dir)
|
60
|
-
puts "Download completed."
|
61
|
-
end
|
62
|
-
|
63
|
-
# Search this dictionary's index for the given string.
|
64
|
-
# @param query [String] the search query
|
65
|
-
# @return [Array(Entry)] the results of the search
|
66
|
-
def search(query, exact=false)
|
67
|
-
results = []
|
68
|
-
return results if query.empty?
|
69
|
-
|
70
|
-
results = @index.search(query, exact)
|
71
|
-
end
|
72
|
-
|
73
|
-
# Retrieves the definition of a part-of-speech from its abbreviation
|
74
|
-
# @param pos [String] the abbreviation for the part-of-speech
|
75
|
-
# @return [String] the full description of the part-of-speech
|
76
|
-
def get_pos(pos)
|
77
|
-
@index.get_pos(pos)
|
78
|
-
end
|
79
|
-
|
80
|
-
private
|
81
|
-
|
82
|
-
def load_index
|
83
|
-
if loaded?
|
84
|
-
Exception.new("Dictionary index is already loaded")
|
85
|
-
else
|
86
|
-
@index.build
|
87
|
-
end
|
88
|
-
end
|
89
|
-
end
|
90
|
-
end
|
data/lib/downloader.rb
DELETED
@@ -1,42 +0,0 @@
|
|
1
|
-
require 'open-uri'
|
2
|
-
require 'zlib'
|
3
|
-
require 'rsync'
|
4
|
-
|
5
|
-
module JDict
|
6
|
-
class Downloader
|
7
|
-
def download
|
8
|
-
|
9
|
-
end
|
10
|
-
|
11
|
-
def sync
|
12
|
-
|
13
|
-
end
|
14
|
-
|
15
|
-
private
|
16
|
-
|
17
|
-
def retrieve_file(url, dest_dir)
|
18
|
-
filename = File.basename(url)
|
19
|
-
full_path = File.join(dest_dir, filename)
|
20
|
-
File.write(full_path, open(url).read)
|
21
|
-
full_path
|
22
|
-
end
|
23
|
-
|
24
|
-
def gunzip(filename)
|
25
|
-
to_write = File.join(File.dirname(filename), File.basename(filename, ".gz"))
|
26
|
-
|
27
|
-
File.open(to_write, 'w') do |wri|
|
28
|
-
File.open(filename) do |f|
|
29
|
-
gz = Zlib::GzipReader.new(f)
|
30
|
-
wri.write(gz.read)
|
31
|
-
gz.close
|
32
|
-
end
|
33
|
-
end
|
34
|
-
|
35
|
-
File.unlink(filename)
|
36
|
-
end
|
37
|
-
|
38
|
-
def rsync(src, dest)
|
39
|
-
Rsync.run(src, dest)
|
40
|
-
end
|
41
|
-
end
|
42
|
-
end
|