textutils 0.9.0 → 0.9.1
Sign up to get free protection for your applications and to get access to all the features.
- data/Manifest.txt +2 -0
- data/lib/textutils.rb +3 -0
- data/lib/textutils/core_ext/file.rb +15 -0
- data/lib/textutils/core_ext/time.rb +19 -0
- data/lib/textutils/helper/title_helper.rb +43 -4
- data/lib/textutils/utils.rb +0 -15
- data/lib/textutils/version.rb +1 -1
- data/test/test_title_helper.rb +31 -1
- metadata +13 -11
data/Manifest.txt
CHANGED
data/lib/textutils.rb
CHANGED
|
@@ -42,6 +42,9 @@ require 'textutils/helper/address_helper'
|
|
|
42
42
|
require 'textutils/helper/value_helper'
|
|
43
43
|
|
|
44
44
|
require 'textutils/utils'
|
|
45
|
+
require 'textutils/core_ext/file'
|
|
46
|
+
require 'textutils/core_ext/time'
|
|
47
|
+
|
|
45
48
|
require 'textutils/reader/code_reader'
|
|
46
49
|
require 'textutils/reader/hash_reader'
|
|
47
50
|
require 'textutils/reader/hash_reader_v2'
|
|
@@ -0,0 +1,15 @@
|
|
|
1
|
+
# encoding: utf-8
|
|
2
|
+
|
|
3
|
+
class File
|
|
4
|
+
def self.read_utf8( path )
|
|
5
|
+
text = open( path, 'r:bom|utf-8' ) do |file|
|
|
6
|
+
file.read
|
|
7
|
+
end
|
|
8
|
+
|
|
9
|
+
# NB: for convenience: convert fancy unicode dashes/hyphens to plain ascii hyphen-minus
|
|
10
|
+
text = TextUtils.convert_unicode_dashes_to_plain_ascii( text, path: path )
|
|
11
|
+
|
|
12
|
+
text
|
|
13
|
+
end
|
|
14
|
+
end # class File
|
|
15
|
+
|
|
@@ -0,0 +1,19 @@
|
|
|
1
|
+
# encoding: utf-8
|
|
2
|
+
|
|
3
|
+
|
|
4
|
+
class Time
|
|
5
|
+
|
|
6
|
+
def self.cet( str ) # central european time (cet) + central european summer time (cest)
|
|
7
|
+
ActiveSupport::TimeZone['Vienna'].parse( str )
|
|
8
|
+
end
|
|
9
|
+
|
|
10
|
+
def self.eet( str ) # eastern european time (eet) + 2 hours
|
|
11
|
+
ActiveSupport::TimeZone['Bucharest'].parse( str )
|
|
12
|
+
end
|
|
13
|
+
|
|
14
|
+
def self.cst( str ) # central standard time (cst) - 6 hours
|
|
15
|
+
ActiveSupport::TimeZone['Mexico City'].parse( str )
|
|
16
|
+
end
|
|
17
|
+
|
|
18
|
+
end # class Time
|
|
19
|
+
|
|
@@ -45,7 +45,7 @@ module TextUtils
|
|
|
45
45
|
|
|
46
46
|
def strip_whitespaces( title )
|
|
47
47
|
# remove all whitespace and punctuation
|
|
48
|
-
title.gsub( /[ \t_\-\.!()\[\]'"
|
|
48
|
+
title.gsub( /[ \t_\-\.!()\[\]'"’\/]/, '' )
|
|
49
49
|
end
|
|
50
50
|
|
|
51
51
|
def strip_special_chars( title )
|
|
@@ -81,21 +81,28 @@ module TextUtils
|
|
|
81
81
|
## see http://en.wikipedia.org/wiki/List_of_XML_and_HTML_character_entity_references for more
|
|
82
82
|
|
|
83
83
|
## todo: add unicode codepoint name
|
|
84
|
-
|
|
84
|
+
|
|
85
|
+
|
|
85
86
|
alternatives = [
|
|
86
87
|
['ß', 'ss'],
|
|
88
|
+
|
|
87
89
|
['æ', 'ae'],
|
|
88
90
|
['ä', 'ae'],
|
|
89
|
-
['ā', 'a' ], # e.g. Liepājas
|
|
91
|
+
['ā', 'a' ], # e.g. Liepājas, Kāṭhmāḍaũ
|
|
90
92
|
['á', 'a' ], # e.g. Bogotá, Králové
|
|
93
|
+
['à', 'a' ], # e.g. Thành Phố Hồ Chí Minh [Saigon]
|
|
91
94
|
['ã', 'a' ], # e.g São Paulo
|
|
92
95
|
['ă', 'a' ], # e.g. Chișinău
|
|
93
96
|
['â', 'a' ], # e.g Goiânia
|
|
94
97
|
['å', 'a' ], # e.g. Vålerenga
|
|
95
98
|
['ą', 'a' ], # e.g. Śląsk
|
|
99
|
+
|
|
96
100
|
['ç', 'c' ], # e.g. São Gonçalo, Iguaçu, Neftçi
|
|
97
101
|
['ć', 'c' ], # e.g. Budućnost
|
|
98
102
|
['č', 'c' ], # e.g. Tradiční, Výčepní
|
|
103
|
+
|
|
104
|
+
['ḍ', 'd' ], # e.g. Kāṭhmāḍaũ [Kathmandu]
|
|
105
|
+
|
|
99
106
|
['é', 'e' ], # e.g. Vélez, Králové
|
|
100
107
|
['è', 'e' ], # e.g. Rivières
|
|
101
108
|
['ê', 'e' ], # e.g. Grêmio
|
|
@@ -103,31 +110,53 @@ module TextUtils
|
|
|
103
110
|
['ĕ', 'e' ], # e.g. Svĕtlý
|
|
104
111
|
['ė', 'e' ], # e.g. Vėtra
|
|
105
112
|
['ë', 'e' ], # e.g. Skënderbeu
|
|
113
|
+
|
|
106
114
|
['ğ', 'g' ], # e.g. Qarabağ
|
|
115
|
+
|
|
116
|
+
['ḥ', 'h' ], # e.g. Ad-Dawḥah [Doha]
|
|
117
|
+
|
|
107
118
|
['ì', 'i' ], # e.g. Potosì
|
|
108
119
|
['í', 'i' ], # e.g. Ústí
|
|
109
120
|
['ï', 'i' ], # e.g. El Djazaïr
|
|
121
|
+
['ī', 'i' ], # e.g. Al-Iskandarīyah [Alexandria]
|
|
122
|
+
|
|
110
123
|
['ł', 'l' ], # e.g. Wisła, Wrocław
|
|
111
124
|
['ñ', 'n' ], # e.g. Porteño
|
|
112
125
|
['ň', 'n' ], # e.g. Plzeň, Třeboň
|
|
126
|
+
|
|
113
127
|
['ö', 'oe'],
|
|
114
128
|
['ő', 'o' ], # e.g. Győri
|
|
115
129
|
['ó', 'o' ], # e.g. Colón, Łódź, Kraków
|
|
116
130
|
['õ', 'o' ], # e.g. Nõmme
|
|
117
131
|
['ô', 'o' ], # e.g. Amazônia (pt)
|
|
118
|
-
['ō', 'o' ], # e.g.
|
|
132
|
+
['ō', 'o' ], # e.g. Tōkyō, Pishōr
|
|
133
|
+
['ŏ', 'o' ], # e.g. P'yŏngyang [Pyongyang]
|
|
119
134
|
['ø', 'o' ], # e.g. Fuglafjørdur, København
|
|
135
|
+
['ố', 'o' ], # e.g. Thành Phố Hồ Chí Minh [Saigon]
|
|
136
|
+
['ồ', 'o' ], # e.g. Thành Phố Hồ Chí Minh [Saigon]
|
|
137
|
+
['ộ', 'o' ], # e.g. Hà Nội [Hanoi]
|
|
138
|
+
|
|
120
139
|
['ř', 'r' ], # e.g. Třeboň
|
|
140
|
+
|
|
121
141
|
['ș', 's' ], # e.g. Chișinău, București
|
|
122
142
|
['ş', 's' ], # e.g. Beşiktaş
|
|
123
143
|
['š', 's' ], # e.g. Košice
|
|
144
|
+
['ṣ', 's' ], # e.g. Al-Mawṣil [Mosul]
|
|
145
|
+
|
|
124
146
|
['ť', 't' ], # e.g. Měšťan
|
|
147
|
+
['ṭ', 't' ], # e.g. Al-Kharṭūm [Khartoum], Kāṭhmāḍaũ
|
|
148
|
+
|
|
125
149
|
['ü', 'ue'],
|
|
126
150
|
['ú', 'u' ], # e.g. Fútbol
|
|
127
151
|
['ù', 'u' ], # e.g. Xyauyù (it)
|
|
128
152
|
['ū', 'u' ], # e.g. Sūduva
|
|
129
153
|
['ů', 'u' ], # e.g. Sládkův
|
|
154
|
+
['ũ', 'u' ], # e.g. Kāṭhmāḍaũ [Kathmandu]
|
|
155
|
+
|
|
130
156
|
['ı', 'u' ], # e.g. Bakı # use u?? (Baku) why-why not?
|
|
157
|
+
|
|
158
|
+
['x̌', 'x'], # e.g. Pex̌awar [Peshawar]
|
|
159
|
+
|
|
131
160
|
['ý', 'y' ], # e.g. Nefitrovaný
|
|
132
161
|
['ź', 'z' ], # e.g. Łódź
|
|
133
162
|
['ž', 'z' ], # e.g. Domžale, Petržalka
|
|
@@ -135,15 +164,25 @@ module TextUtils
|
|
|
135
164
|
|
|
136
165
|
['Á', 'a' ], # e.g. Águila (es)
|
|
137
166
|
['Č', 'c' ], # e.g. České
|
|
167
|
+
|
|
168
|
+
['Ḥ', 'h' ], # e.g. Ḥalab [Aleppo]
|
|
169
|
+
['Ḫ', 'h' ], # e.g. Ḫamīs Mušayṭ
|
|
138
170
|
['İ', 'i' ], # e.g. İnter
|
|
139
171
|
['Í', 'i' ], # e.g. ÍBV
|
|
140
172
|
['Ł', 'l' ], # e.g. Łódź
|
|
173
|
+
|
|
141
174
|
['Ö', 'oe' ], # e.g. Örebro
|
|
175
|
+
['Ō', 'o' ], # e.g. Ōsaka [Osaka]
|
|
142
176
|
['Ø', 'o' ], # e.g. Nogne Ø Imperial Stout (no)
|
|
177
|
+
|
|
143
178
|
['Ř', 'r' ], # e.g. Řezák
|
|
179
|
+
|
|
144
180
|
['Ś', 's' ], # e.g. Śląsk
|
|
145
181
|
['Š', 's' ], # e.g. MŠK
|
|
146
182
|
['Ş', 's' ], # e.g. Şüvälan
|
|
183
|
+
['Ṣ', 's' ], # e.g. Ṣan'ā' [Sana'a]
|
|
184
|
+
|
|
185
|
+
['Ṭ', 't' ], # e.g. Ṭarābulus [Tripoli]
|
|
147
186
|
['Ú', 'u' ], # e.g. Ústí, Újpest
|
|
148
187
|
['Ž', 'z' ], # e.g. Žilina
|
|
149
188
|
['Ż', 'z' ] # e.g. Żywiec (polish)
|
data/lib/textutils/utils.rb
CHANGED
|
@@ -9,21 +9,6 @@ end
|
|
|
9
9
|
|
|
10
10
|
|
|
11
11
|
|
|
12
|
-
class File
|
|
13
|
-
def self.read_utf8( path )
|
|
14
|
-
text = open( path, 'r:bom|utf-8' ) do |file|
|
|
15
|
-
file.read
|
|
16
|
-
end
|
|
17
|
-
|
|
18
|
-
# NB: for convenience: convert fancy unicode dashes/hyphens to plain ascii hyphen-minus
|
|
19
|
-
text = TextUtils.convert_unicode_dashes_to_plain_ascii( text, path: path )
|
|
20
|
-
|
|
21
|
-
text
|
|
22
|
-
end
|
|
23
|
-
end # class File
|
|
24
|
-
|
|
25
|
-
|
|
26
|
-
|
|
27
12
|
def title_esc_regex( title_unescaped )
|
|
28
13
|
puts "*** warn: depreceated fn call: use TextUtils.title_esc_regex() or include TextUtils::TitleHelpers"
|
|
29
14
|
TextUtils.title_esc_regex( title_unescaped )
|
data/lib/textutils/version.rb
CHANGED
data/test/test_title_helper.rb
CHANGED
|
@@ -38,7 +38,37 @@ class TestTitleHelper < MiniTest::Unit::TestCase
|
|
|
38
38
|
[ '$Alianz$ Arena', 'alianzarena' ],
|
|
39
39
|
[ 'Arena Amazônia', 'arenaamazonia' ],
|
|
40
40
|
[ 'Tōkyō [Tokyo]', 'tokyo' ],
|
|
41
|
-
[ '
|
|
41
|
+
[ 'Ōsaka [Osaka]', 'osaka' ],
|
|
42
|
+
[ 'El Djazaïr [Algiers]', 'eldjazair' ],
|
|
43
|
+
[ 'Al-Kharṭūm [Khartoum]', 'alkhartum' ],
|
|
44
|
+
[ 'Ṭarābulus [Tripoli]', 'tarabulus' ],
|
|
45
|
+
[ 'Al-Iskandarīyah [Alexandria]', 'aliskandariyah' ],
|
|
46
|
+
[ 'Pex̌awar', 'pexawar'],
|
|
47
|
+
[ 'Pishōr', 'pishor' ],
|
|
48
|
+
[ 'Pishāwar', 'pishawar' ],
|
|
49
|
+
[ 'Islām ābād', 'islamabad' ],
|
|
50
|
+
[ 'Thành Phố Hồ Chí Minh [Saigon]', 'thanhphohochiminh' ],
|
|
51
|
+
[ 'Hà Nội [Hanoi]', 'hanoi' ],
|
|
52
|
+
[ 'Donets’k', 'donetsk' ],
|
|
53
|
+
[ 'Baghdād [Baghdad]', 'baghdad'],
|
|
54
|
+
[ 'Al-Mawṣil [Mosul]', 'almawsil'],
|
|
55
|
+
[ 'Al-Baṣrah [Basra]', 'albasrah'],
|
|
56
|
+
[ 'Arbīl [Erbil]', 'arbil' ],
|
|
57
|
+
[ 'Kirkūk [Kirkuk]', 'kirkuk' ],
|
|
58
|
+
[ 'Tehrān [Tehran]', 'tehran' ],
|
|
59
|
+
[ 'Eṣfahān [Isfahan]', 'esfahan' ],
|
|
60
|
+
[ 'Shīrāz [Shiraz]', 'shiraz' ],
|
|
61
|
+
[ 'Tabrīz [Tabriz]', 'tabriz' ],
|
|
62
|
+
[ 'Ahvāz [Ahvaz]', 'ahvaz' ],
|
|
63
|
+
[ 'Ad-Dawḥah [Doha]', 'addawhah'],
|
|
64
|
+
[ 'Ḥalab [Aleppo]', 'halab'],
|
|
65
|
+
[ 'Al-Madīnah [Medina]', 'almadinah'],
|
|
66
|
+
[ 'Ad-Dammām [Dammam]', 'addammam' ],
|
|
67
|
+
[ 'Aṭ-Ṭā’if', 'attaif'],
|
|
68
|
+
[ 'Ḫamīs Mušayṭ', 'hamismusayt'],
|
|
69
|
+
[ "Ṣan'ā' [Sana'a]", 'sana'],
|
|
70
|
+
[ "P'yŏngyang [Pyongyang]", 'pyongyang' ],
|
|
71
|
+
[ 'Kāṭhmāḍaũ [Kathmandu]', 'kathmadau' ]
|
|
42
72
|
]
|
|
43
73
|
|
|
44
74
|
txt_io.each do |txt|
|
metadata
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
|
2
2
|
name: textutils
|
|
3
3
|
version: !ruby/object:Gem::Version
|
|
4
|
-
version: 0.9.
|
|
4
|
+
version: 0.9.1
|
|
5
5
|
prerelease:
|
|
6
6
|
platform: ruby
|
|
7
7
|
authors:
|
|
@@ -13,7 +13,7 @@ date: 2014-03-24 00:00:00.000000000 Z
|
|
|
13
13
|
dependencies:
|
|
14
14
|
- !ruby/object:Gem::Dependency
|
|
15
15
|
name: props
|
|
16
|
-
requirement: &
|
|
16
|
+
requirement: &73970010 !ruby/object:Gem::Requirement
|
|
17
17
|
none: false
|
|
18
18
|
requirements:
|
|
19
19
|
- - ! '>='
|
|
@@ -21,10 +21,10 @@ dependencies:
|
|
|
21
21
|
version: '0'
|
|
22
22
|
type: :runtime
|
|
23
23
|
prerelease: false
|
|
24
|
-
version_requirements: *
|
|
24
|
+
version_requirements: *73970010
|
|
25
25
|
- !ruby/object:Gem::Dependency
|
|
26
26
|
name: logutils
|
|
27
|
-
requirement: &
|
|
27
|
+
requirement: &73969260 !ruby/object:Gem::Requirement
|
|
28
28
|
none: false
|
|
29
29
|
requirements:
|
|
30
30
|
- - ~>
|
|
@@ -32,10 +32,10 @@ dependencies:
|
|
|
32
32
|
version: '0.5'
|
|
33
33
|
type: :runtime
|
|
34
34
|
prerelease: false
|
|
35
|
-
version_requirements: *
|
|
35
|
+
version_requirements: *73969260
|
|
36
36
|
- !ruby/object:Gem::Dependency
|
|
37
37
|
name: activesupport
|
|
38
|
-
requirement: &
|
|
38
|
+
requirement: &73969070 !ruby/object:Gem::Requirement
|
|
39
39
|
none: false
|
|
40
40
|
requirements:
|
|
41
41
|
- - ! '>='
|
|
@@ -43,10 +43,10 @@ dependencies:
|
|
|
43
43
|
version: '0'
|
|
44
44
|
type: :runtime
|
|
45
45
|
prerelease: false
|
|
46
|
-
version_requirements: *
|
|
46
|
+
version_requirements: *73969070
|
|
47
47
|
- !ruby/object:Gem::Dependency
|
|
48
48
|
name: rdoc
|
|
49
|
-
requirement: &
|
|
49
|
+
requirement: &73968810 !ruby/object:Gem::Requirement
|
|
50
50
|
none: false
|
|
51
51
|
requirements:
|
|
52
52
|
- - ~>
|
|
@@ -54,10 +54,10 @@ dependencies:
|
|
|
54
54
|
version: '3.10'
|
|
55
55
|
type: :development
|
|
56
56
|
prerelease: false
|
|
57
|
-
version_requirements: *
|
|
57
|
+
version_requirements: *73968810
|
|
58
58
|
- !ruby/object:Gem::Dependency
|
|
59
59
|
name: hoe
|
|
60
|
-
requirement: &
|
|
60
|
+
requirement: &73953840 !ruby/object:Gem::Requirement
|
|
61
61
|
none: false
|
|
62
62
|
requirements:
|
|
63
63
|
- - ~>
|
|
@@ -65,7 +65,7 @@ dependencies:
|
|
|
65
65
|
version: '3.3'
|
|
66
66
|
type: :development
|
|
67
67
|
prerelease: false
|
|
68
|
-
version_requirements: *
|
|
68
|
+
version_requirements: *73953840
|
|
69
69
|
description: textutils - Text Filters, Helpers, Readers and More
|
|
70
70
|
email: ruby-talk@ruby-lang.org
|
|
71
71
|
executables: []
|
|
@@ -79,6 +79,8 @@ files:
|
|
|
79
79
|
- Rakefile
|
|
80
80
|
- lib/textutils.rb
|
|
81
81
|
- lib/textutils/classifier.rb
|
|
82
|
+
- lib/textutils/core_ext/file.rb
|
|
83
|
+
- lib/textutils/core_ext/time.rb
|
|
82
84
|
- lib/textutils/filter/code_filter.rb
|
|
83
85
|
- lib/textutils/filter/comment_filter.rb
|
|
84
86
|
- lib/textutils/filter/erb_django_filter.rb
|