microformats2 1.0.0
Sign up to get free protection for your applications and to get access to all the features.
- data/.autotest +23 -0
- data/.gemtest +0 -0
- data/History.txt +6 -0
- data/Manifest.txt +8 -0
- data/README.txt +41 -0
- data/Rakefile +12 -0
- data/bin/microformats2 +3 -0
- data/lib/microformats2.rb +165 -0
- data/test/test_microformats2.rb +163 -0
- metadata +105 -0
data/.autotest
ADDED
@@ -0,0 +1,23 @@
|
|
1
|
+
# -*- ruby -*-
|
2
|
+
|
3
|
+
require 'autotest/restart'
|
4
|
+
|
5
|
+
# Autotest.add_hook :initialize do |at|
|
6
|
+
# at.extra_files << "../some/external/dependency.rb"
|
7
|
+
#
|
8
|
+
# at.libs << ":../some/external"
|
9
|
+
#
|
10
|
+
# at.add_exception 'vendor'
|
11
|
+
#
|
12
|
+
# at.add_mapping(/dependency.rb/) do |f, _|
|
13
|
+
# at.files_matching(/test_.*rb$/)
|
14
|
+
# end
|
15
|
+
#
|
16
|
+
# %w(TestA TestB).each do |klass|
|
17
|
+
# at.extra_class_map[klass] = "test/test_misc.rb"
|
18
|
+
# end
|
19
|
+
# end
|
20
|
+
|
21
|
+
# Autotest.add_hook :run_command do |at|
|
22
|
+
# system "rake build"
|
23
|
+
# end
|
data/.gemtest
ADDED
File without changes
|
data/History.txt
ADDED
data/Manifest.txt
ADDED
data/README.txt
ADDED
@@ -0,0 +1,41 @@
|
|
1
|
+
# Microformats
|
2
|
+
|
3
|
+
http://github.com/veganstraightedge/microformats2
|
4
|
+
|
5
|
+
## DESCRIPTION
|
6
|
+
|
7
|
+
Generic Microformats 2 Extractor
|
8
|
+
|
9
|
+
## FEATURES/PROBLEMS
|
10
|
+
|
11
|
+
* parses and extracts [Microformats 2](http://microformats.org/wiki/microformats-2) syntax
|
12
|
+
* needs more test cases
|
13
|
+
* needs better docs
|
14
|
+
|
15
|
+
## SYNOPSIS
|
16
|
+
|
17
|
+
Microformats2.parse(File.open("http://iamshane.html"))
|
18
|
+
|
19
|
+
## REQUIREMENTS
|
20
|
+
|
21
|
+
* Hoe
|
22
|
+
* Nokogiri
|
23
|
+
|
24
|
+
## INSTALL
|
25
|
+
|
26
|
+
sudo gem install microformats2
|
27
|
+
|
28
|
+
## DEVELOPERS
|
29
|
+
|
30
|
+
After checking out the source, run:
|
31
|
+
|
32
|
+
rake newb
|
33
|
+
|
34
|
+
This task will install any missing dependencies,
|
35
|
+
run the tests/specs, and generate the RDoc.
|
36
|
+
|
37
|
+
## LICENSE
|
38
|
+
|
39
|
+
PUBLIC DOMAIN.
|
40
|
+
Your heart is as free as the air you breathe.
|
41
|
+
The ground you stand on is liberated territory.
|
data/Rakefile
ADDED
data/bin/microformats2
ADDED
@@ -0,0 +1,165 @@
|
|
1
|
+
require 'nokogiri'
|
2
|
+
require 'time'
|
3
|
+
require 'date'
|
4
|
+
|
5
|
+
module Microformats2
|
6
|
+
VERSION = "1.0.0"
|
7
|
+
|
8
|
+
class LoadError < StandardError; end
|
9
|
+
|
10
|
+
def self.parse(html)
|
11
|
+
raise LoadError, "argument must be a String or File" unless [String, File].include?(html.class)
|
12
|
+
|
13
|
+
doc = Nokogiri::HTML(html)
|
14
|
+
microformats = Hash.new{|hash, key| hash[key] = Array.new}
|
15
|
+
doc.css("*[class^=h-]").each do |microformat|
|
16
|
+
constant_name = classify(microformat.attribute("class").to_s.gsub("-","_"))
|
17
|
+
|
18
|
+
if Object.const_defined?(constant_name)
|
19
|
+
klass = Object.const_get(constant_name)
|
20
|
+
else
|
21
|
+
klass = Class.new
|
22
|
+
Object.const_set constant_name, klass
|
23
|
+
end
|
24
|
+
|
25
|
+
obj = klass.new
|
26
|
+
|
27
|
+
# Add any properties to the object
|
28
|
+
add_properties(microformat, obj)
|
29
|
+
add_urls(microformat, obj)
|
30
|
+
add_dates(microformat, obj)
|
31
|
+
add_times(microformat, obj)
|
32
|
+
|
33
|
+
microformats[constant_name.downcase.to_sym] << obj
|
34
|
+
end
|
35
|
+
|
36
|
+
return microformats
|
37
|
+
end
|
38
|
+
|
39
|
+
def self.add_method(obj, method_name)
|
40
|
+
unless obj.respond_to?(method_name)
|
41
|
+
obj.class.class_eval { attr_accessor method_name }
|
42
|
+
end
|
43
|
+
|
44
|
+
obj
|
45
|
+
end
|
46
|
+
|
47
|
+
def self.add_properties(mf, obj)
|
48
|
+
%w(p n e i).each do |letter|
|
49
|
+
mf.css("*[class|=#{letter}]").each do |property|
|
50
|
+
property.attribute("class").to_s.split.each do |css_class|
|
51
|
+
if css_class =~ /^[pnei]/
|
52
|
+
css_class = css_class[2..-1].gsub("-","_")
|
53
|
+
method_name = css_class.gsub("-","_")
|
54
|
+
value = property.text.gsub(/\n+/, " ").gsub(/\s+/, " ").strip
|
55
|
+
|
56
|
+
add_method(obj, method_name)
|
57
|
+
|
58
|
+
if cur = obj.send(method_name)
|
59
|
+
if cur.kind_of? Array
|
60
|
+
cur << value
|
61
|
+
else
|
62
|
+
obj.send("#{method_name}=", [cur, value])
|
63
|
+
end
|
64
|
+
else
|
65
|
+
obj.send("#{method_name}=", value)
|
66
|
+
end
|
67
|
+
end
|
68
|
+
end
|
69
|
+
end
|
70
|
+
end
|
71
|
+
end
|
72
|
+
|
73
|
+
def self.add_urls(mf, obj)
|
74
|
+
mf.css("*[class*=u-]").each do |property|
|
75
|
+
property.attribute("class").to_s.split.each do |css_class|
|
76
|
+
if css_class =~ /^u/
|
77
|
+
css_class = css_class[2..-1].gsub("-","_")
|
78
|
+
method_name = css_class.gsub("-","_")
|
79
|
+
value = property.attribute("href").to_s
|
80
|
+
|
81
|
+
add_method(obj, method_name)
|
82
|
+
|
83
|
+
if cur = obj.send(method_name)
|
84
|
+
if cur.kind_of? Array
|
85
|
+
cur << value
|
86
|
+
else
|
87
|
+
obj.send("#{method_name}=", [cur, value])
|
88
|
+
end
|
89
|
+
else
|
90
|
+
obj.send("#{method_name}=", value)
|
91
|
+
end
|
92
|
+
end
|
93
|
+
end
|
94
|
+
end
|
95
|
+
end
|
96
|
+
|
97
|
+
def self.add_dates(mf, obj)
|
98
|
+
mf.css("*[class*=d-]").each do |property|
|
99
|
+
property.attribute("class").to_s.split.each do |css_class|
|
100
|
+
if css_class =~ /^d/
|
101
|
+
css_class = css_class[2..-1].gsub("-","_")
|
102
|
+
method_name = css_class.gsub("-","_")
|
103
|
+
value = DateTime.parse((property.attribute("title") || property.text).to_s)
|
104
|
+
|
105
|
+
add_method(obj, method_name)
|
106
|
+
|
107
|
+
if cur = obj.send(method_name)
|
108
|
+
if cur.kind_of? Array
|
109
|
+
cur << value
|
110
|
+
else
|
111
|
+
obj.send("#{method_name}=", [cur, value])
|
112
|
+
end
|
113
|
+
else
|
114
|
+
obj.send("#{method_name}=", value)
|
115
|
+
end
|
116
|
+
end
|
117
|
+
end
|
118
|
+
end
|
119
|
+
end
|
120
|
+
|
121
|
+
def self.add_times(mf, obj)
|
122
|
+
mf.css("*[class*=t-]").each do |property|
|
123
|
+
property.attribute("class").to_s.split.each do |css_class|
|
124
|
+
if css_class =~ /^t/
|
125
|
+
css_class = css_class[2..-1].gsub("-","_")
|
126
|
+
method_name = css_class.gsub("-","_")
|
127
|
+
value = Time.parse((property.attribute("title") || property.text).to_s)
|
128
|
+
|
129
|
+
add_method(obj, method_name)
|
130
|
+
|
131
|
+
if cur = obj.send(method_name)
|
132
|
+
if cur.kind_of? Array
|
133
|
+
cur << value
|
134
|
+
else
|
135
|
+
obj.send("#{method_name}=", [cur, value])
|
136
|
+
end
|
137
|
+
else
|
138
|
+
obj.send("#{method_name}=", value)
|
139
|
+
end
|
140
|
+
end
|
141
|
+
end
|
142
|
+
end
|
143
|
+
end
|
144
|
+
|
145
|
+
# Thank you Rails Developers for your unitentional contribution to this project
|
146
|
+
# File activesupport/lib/active_support/inflector/inflections.rb, line 206
|
147
|
+
def self.classify(str)
|
148
|
+
# strip out any leading schema name
|
149
|
+
camelize(singularize(str.to_s.sub(/.*\./, '')))
|
150
|
+
end
|
151
|
+
|
152
|
+
# File activesupport/lib/active_support/inflector/inflections.rb, line 148
|
153
|
+
def self.singularize(word)
|
154
|
+
result = word.to_s.dup
|
155
|
+
end
|
156
|
+
|
157
|
+
# File activesupport/lib/active_support/inflector/methods.rb, line 28
|
158
|
+
def self.camelize(lower_case_and_underscored_word, first_letter_in_uppercase = true)
|
159
|
+
if first_letter_in_uppercase
|
160
|
+
lower_case_and_underscored_word.to_s.gsub(/\/(.?)/) { "::#{$1.upcase}" }.gsub(/(?:^|_)(.)/) { $1.upcase }
|
161
|
+
else
|
162
|
+
lower_case_and_underscored_word.to_s[0].chr.downcase + camelize(lower_case_and_underscored_word)[1..-1]
|
163
|
+
end
|
164
|
+
end
|
165
|
+
end
|
@@ -0,0 +1,163 @@
|
|
1
|
+
require "test/unit"
|
2
|
+
require "microformats2"
|
3
|
+
|
4
|
+
class TestMicroformats2 < Test::Unit::TestCase
|
5
|
+
def test_throw_exception_on_non_string_params
|
6
|
+
assert_raise Microformats2::LoadError do
|
7
|
+
Microformats2.parse(nil)
|
8
|
+
end
|
9
|
+
end
|
10
|
+
|
11
|
+
def test_returns_hash_of_microformat_objects
|
12
|
+
result = Microformats2.parse("A String")
|
13
|
+
assert_equal Hash, result.class
|
14
|
+
end
|
15
|
+
|
16
|
+
def test_only_parse_microformats
|
17
|
+
result = Microformats2.parse("<html><body><p>Something</p></body></html>")
|
18
|
+
assert_equal 0, result.size
|
19
|
+
end
|
20
|
+
|
21
|
+
def test_extracts_hcard_from_an_html_file
|
22
|
+
hcard = <<-END
|
23
|
+
<html>
|
24
|
+
<head>
|
25
|
+
<title>Simple hCard</title>
|
26
|
+
</head>
|
27
|
+
|
28
|
+
<body>
|
29
|
+
<h1 class="h-card">Chris</h1>
|
30
|
+
</body>
|
31
|
+
</html>
|
32
|
+
END
|
33
|
+
result = Microformats2.parse(File.open(File.join(File.dirname(__FILE__), "hcard.html")))
|
34
|
+
assert_equal HCard, result[:hcard].first.class
|
35
|
+
end
|
36
|
+
|
37
|
+
def test_extracts_hcard_from_html
|
38
|
+
hcard = <<-END
|
39
|
+
<html>
|
40
|
+
<head>
|
41
|
+
<title>Simple hCard</title>
|
42
|
+
</head>
|
43
|
+
|
44
|
+
<body>
|
45
|
+
<h1 class="h-card">Chris</h1>
|
46
|
+
</body>
|
47
|
+
</html>
|
48
|
+
END
|
49
|
+
result = Microformats2.parse(hcard)
|
50
|
+
assert_equal HCard, result[:hcard].first.class
|
51
|
+
end
|
52
|
+
|
53
|
+
def test_constructs_properties_from_hcard
|
54
|
+
hcard = <<-END
|
55
|
+
<html>
|
56
|
+
<head>
|
57
|
+
<title>Simple hCard</title>
|
58
|
+
</head>
|
59
|
+
|
60
|
+
<body>
|
61
|
+
<h1 class="h-card">
|
62
|
+
<a class="p-fn u-url" href="http://factoryjoe.com/">
|
63
|
+
<span class="p-given-name">Chris</span>
|
64
|
+
<abbr class="p-additional-name">R.</abbr>
|
65
|
+
<span class="p-family-name">Messina</span>
|
66
|
+
</a>
|
67
|
+
</h1>
|
68
|
+
</body>
|
69
|
+
</html>
|
70
|
+
END
|
71
|
+
result = Microformats2.parse(hcard)
|
72
|
+
mycard = result[:hcard].first
|
73
|
+
|
74
|
+
assert_equal "Chris", mycard.given_name
|
75
|
+
assert_equal "R.", mycard.additional_name
|
76
|
+
assert_equal "Messina", mycard.family_name
|
77
|
+
assert_equal "Chris R. Messina", mycard.fn
|
78
|
+
end
|
79
|
+
|
80
|
+
def test_constructs_dates
|
81
|
+
hcard = <<-END
|
82
|
+
<html>
|
83
|
+
<head>
|
84
|
+
<title>Simple hCard</title>
|
85
|
+
</head>
|
86
|
+
|
87
|
+
<body>
|
88
|
+
<h1 class="h-card">
|
89
|
+
<span class="d-bday">1979-09-18</span>
|
90
|
+
<span class="d-epoch" title="1970-01-01">EPOCH!</span>
|
91
|
+
</h1>
|
92
|
+
</body>
|
93
|
+
</html>
|
94
|
+
END
|
95
|
+
result = Microformats2.parse(hcard)
|
96
|
+
mycard = result[:hcard].first
|
97
|
+
|
98
|
+
assert_equal DateTime.parse("1979-09-18"), mycard.bday
|
99
|
+
assert_equal DateTime.parse("1970-01-01"), mycard.epoch
|
100
|
+
end
|
101
|
+
|
102
|
+
def test_constructs_times
|
103
|
+
hcard = <<-END
|
104
|
+
<html>
|
105
|
+
<head>
|
106
|
+
<title>Simple hCard</title>
|
107
|
+
</head>
|
108
|
+
|
109
|
+
<body>
|
110
|
+
<h1 class="h-card">
|
111
|
+
<span class="t-start">09:30</span>
|
112
|
+
<span class="t-end" title="6:00">Leaving time</span>
|
113
|
+
</h1>
|
114
|
+
</body>
|
115
|
+
</html>
|
116
|
+
END
|
117
|
+
result = Microformats2.parse(hcard)
|
118
|
+
mycard = result[:hcard].first
|
119
|
+
|
120
|
+
assert_equal Time.parse("09:30"), mycard.start
|
121
|
+
assert_equal Time.parse("06:00"), mycard.end
|
122
|
+
end
|
123
|
+
|
124
|
+
def test_ignores_pattern_matches_not_at_the_beginning_of_class
|
125
|
+
hcard = <<-END
|
126
|
+
<html>
|
127
|
+
<head>
|
128
|
+
<title>Simple hCard</title>
|
129
|
+
</head>
|
130
|
+
|
131
|
+
<body>
|
132
|
+
<h1 class="h-card">
|
133
|
+
<span class="p-n-x">Chris</span>
|
134
|
+
</h1>
|
135
|
+
</body>
|
136
|
+
</html>
|
137
|
+
END
|
138
|
+
result = Microformats2.parse(hcard)
|
139
|
+
mycard = result[:hcard].first
|
140
|
+
|
141
|
+
assert_equal "Chris", mycard.n_x
|
142
|
+
assert mycard.n_x.is_a?(String)
|
143
|
+
end
|
144
|
+
|
145
|
+
def test_constructs_urls_from_hcard
|
146
|
+
hcard = <<-END
|
147
|
+
<html>
|
148
|
+
<head>
|
149
|
+
<title>Simple hCard</title>
|
150
|
+
</head>
|
151
|
+
|
152
|
+
<body>
|
153
|
+
<h1 class="h-card">
|
154
|
+
<a class="p-fn u-url" href="http://factoryjoe.com/">Chris</a>
|
155
|
+
</h1>
|
156
|
+
</body>
|
157
|
+
</html>
|
158
|
+
END
|
159
|
+
result = Microformats2.parse(hcard)
|
160
|
+
mycard = result[:hcard].first
|
161
|
+
assert_equal "http://factoryjoe.com/", mycard.url
|
162
|
+
end
|
163
|
+
end
|
metadata
ADDED
@@ -0,0 +1,105 @@
|
|
1
|
+
--- !ruby/object:Gem::Specification
|
2
|
+
name: microformats2
|
3
|
+
version: !ruby/object:Gem::Version
|
4
|
+
hash: 23
|
5
|
+
prerelease:
|
6
|
+
segments:
|
7
|
+
- 1
|
8
|
+
- 0
|
9
|
+
- 0
|
10
|
+
version: 1.0.0
|
11
|
+
platform: ruby
|
12
|
+
authors:
|
13
|
+
- Shane Becker
|
14
|
+
autorequire:
|
15
|
+
bindir: bin
|
16
|
+
cert_chain: []
|
17
|
+
|
18
|
+
date: 2011-06-25 00:00:00 Z
|
19
|
+
dependencies:
|
20
|
+
- !ruby/object:Gem::Dependency
|
21
|
+
name: nokogiri
|
22
|
+
prerelease: false
|
23
|
+
requirement: &id001 !ruby/object:Gem::Requirement
|
24
|
+
none: false
|
25
|
+
requirements:
|
26
|
+
- - ">="
|
27
|
+
- !ruby/object:Gem::Version
|
28
|
+
hash: 3
|
29
|
+
segments:
|
30
|
+
- 0
|
31
|
+
version: "0"
|
32
|
+
type: :runtime
|
33
|
+
version_requirements: *id001
|
34
|
+
- !ruby/object:Gem::Dependency
|
35
|
+
name: hoe
|
36
|
+
prerelease: false
|
37
|
+
requirement: &id002 !ruby/object:Gem::Requirement
|
38
|
+
none: false
|
39
|
+
requirements:
|
40
|
+
- - ~>
|
41
|
+
- !ruby/object:Gem::Version
|
42
|
+
hash: 17
|
43
|
+
segments:
|
44
|
+
- 2
|
45
|
+
- 9
|
46
|
+
version: "2.9"
|
47
|
+
type: :development
|
48
|
+
version_requirements: *id002
|
49
|
+
description: Generic Microformats 2 Extractor
|
50
|
+
email:
|
51
|
+
- veganstraightedge@gmail.com
|
52
|
+
executables:
|
53
|
+
- microformats2
|
54
|
+
extensions: []
|
55
|
+
|
56
|
+
extra_rdoc_files:
|
57
|
+
- History.txt
|
58
|
+
- Manifest.txt
|
59
|
+
- README.txt
|
60
|
+
files:
|
61
|
+
- .autotest
|
62
|
+
- History.txt
|
63
|
+
- Manifest.txt
|
64
|
+
- README.txt
|
65
|
+
- Rakefile
|
66
|
+
- bin/microformats2
|
67
|
+
- lib/microformats2.rb
|
68
|
+
- test/test_microformats2.rb
|
69
|
+
- .gemtest
|
70
|
+
homepage: http://github.com/veganstraightedge/microformats2
|
71
|
+
licenses: []
|
72
|
+
|
73
|
+
post_install_message:
|
74
|
+
rdoc_options:
|
75
|
+
- --main
|
76
|
+
- README.txt
|
77
|
+
require_paths:
|
78
|
+
- lib
|
79
|
+
required_ruby_version: !ruby/object:Gem::Requirement
|
80
|
+
none: false
|
81
|
+
requirements:
|
82
|
+
- - ">="
|
83
|
+
- !ruby/object:Gem::Version
|
84
|
+
hash: 3
|
85
|
+
segments:
|
86
|
+
- 0
|
87
|
+
version: "0"
|
88
|
+
required_rubygems_version: !ruby/object:Gem::Requirement
|
89
|
+
none: false
|
90
|
+
requirements:
|
91
|
+
- - ">="
|
92
|
+
- !ruby/object:Gem::Version
|
93
|
+
hash: 3
|
94
|
+
segments:
|
95
|
+
- 0
|
96
|
+
version: "0"
|
97
|
+
requirements: []
|
98
|
+
|
99
|
+
rubyforge_project: microformats2
|
100
|
+
rubygems_version: 1.8.5
|
101
|
+
signing_key:
|
102
|
+
specification_version: 3
|
103
|
+
summary: Generic Microformats 2 Extractor
|
104
|
+
test_files:
|
105
|
+
- test/test_microformats2.rb
|