hypermicrodata 0.1.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +7 -0
- data/.gitignore +17 -0
- data/.travis.yml +8 -0
- data/Gemfile +8 -0
- data/LICENSE.txt +22 -0
- data/README.md +100 -0
- data/Rakefile +10 -0
- data/bin/hypermicrodata.rb +25 -0
- data/hypermicrodata.gemspec +28 -0
- data/lib/hypermicrodata.rb +37 -0
- data/lib/hypermicrodata/document.rb +27 -0
- data/lib/hypermicrodata/extract.rb +22 -0
- data/lib/hypermicrodata/item.rb +113 -0
- data/lib/hypermicrodata/itemprop_parser.rb +114 -0
- data/lib/hypermicrodata/link.rb +7 -0
- data/lib/hypermicrodata/property.rb +27 -0
- data/lib/hypermicrodata/rails/html_based_json_renderer.rb +35 -0
- data/lib/hypermicrodata/serializer/base.rb +24 -0
- data/lib/hypermicrodata/serializer/hal.rb +47 -0
- data/lib/hypermicrodata/serializer/jsonld.rb +44 -0
- data/lib/hypermicrodata/serializer/uber.rb +100 -0
- data/lib/hypermicrodata/submit_button.rb +105 -0
- data/lib/hypermicrodata/version.rb +3 -0
- data/lib/uberous/uber.rb +104 -0
- data/test/data/example.html +22 -0
- data/test/data/example_itemref.html +16 -0
- data/test/data/example_with_no_itemscope.html +22 -0
- data/test/test_helper.rb +3 -0
- data/test/test_itemref.rb +19 -0
- data/test/test_json.rb +15 -0
- data/test/test_parse.rb +36 -0
- metadata +139 -0
data/lib/uberous/uber.rb
ADDED
|
@@ -0,0 +1,104 @@
|
|
|
1
|
+
module Uberous
|
|
2
|
+
class Uber
|
|
3
|
+
attr_reader :data_collection, :error_data_collection, :version
|
|
4
|
+
def initialize(data_collection = [], error_data_collection = [], version = '1.0')
|
|
5
|
+
@data_collection = data_collection
|
|
6
|
+
@error_data_collection = error_data_collection
|
|
7
|
+
@version = version
|
|
8
|
+
end
|
|
9
|
+
|
|
10
|
+
def add_data(data)
|
|
11
|
+
@data_collection << data
|
|
12
|
+
end
|
|
13
|
+
|
|
14
|
+
def add_error_data(data)
|
|
15
|
+
@error_data_collection << data
|
|
16
|
+
end
|
|
17
|
+
|
|
18
|
+
def add_link(rel, url, options = {})
|
|
19
|
+
link = Data.new(options.merge(rel: rel, url: url))
|
|
20
|
+
add_data(link)
|
|
21
|
+
end
|
|
22
|
+
|
|
23
|
+
def to_hash
|
|
24
|
+
data = @data_collection.map(&:to_hash)
|
|
25
|
+
error = @error_data_collection.map(&:to_hash)
|
|
26
|
+
uber = { version: @version }
|
|
27
|
+
uber[:data] = data unless data.empty?
|
|
28
|
+
uber[:error] = error unless error.empty?
|
|
29
|
+
{ uber: uber }
|
|
30
|
+
end
|
|
31
|
+
end
|
|
32
|
+
|
|
33
|
+
class Data
|
|
34
|
+
ATTR_NAMES = [:id, :name, :rel, :url, :action, :transclude, :model, :sending, :accepting, :value].freeze
|
|
35
|
+
attr_reader :data_collection
|
|
36
|
+
def initialize(attrs = {})
|
|
37
|
+
@attrs = ATTR_NAMES.each_with_object({}) {|name, h| h[name] = attrs[name] if attrs.has_key?(name) } # slice
|
|
38
|
+
@data_collection = Array(attrs[:data_collection])
|
|
39
|
+
end
|
|
40
|
+
|
|
41
|
+
ATTR_NAMES.each do |attr_name|
|
|
42
|
+
define_method(attr_name) do
|
|
43
|
+
@attrs[attr_name]
|
|
44
|
+
end
|
|
45
|
+
define_method("#{attr_name}=") do |value|
|
|
46
|
+
@attrs[attr_name] = value
|
|
47
|
+
end
|
|
48
|
+
end
|
|
49
|
+
|
|
50
|
+
def add_data(data)
|
|
51
|
+
@data_collection << data
|
|
52
|
+
end
|
|
53
|
+
|
|
54
|
+
def add_link(rel, url, options = {})
|
|
55
|
+
link = Data.new(options.merge(rel: rel, url: url))
|
|
56
|
+
add_data(link)
|
|
57
|
+
end
|
|
58
|
+
|
|
59
|
+
def to_hash
|
|
60
|
+
hash = @attrs.dup
|
|
61
|
+
data = @data_collection.map(&:to_hash)
|
|
62
|
+
hash[:data] = data unless data.empty?
|
|
63
|
+
hash
|
|
64
|
+
end
|
|
65
|
+
end
|
|
66
|
+
|
|
67
|
+
class Loader
|
|
68
|
+
attr_reader :uber
|
|
69
|
+
def initialize(hash)
|
|
70
|
+
uber_hash = hash['uber']
|
|
71
|
+
@uber_data_hashes = uber_hash.fetch('data', [])
|
|
72
|
+
@uber_error_data_hashes = uber_hash.fetch('error', [])
|
|
73
|
+
@uber = Uber.new([], [], uber_hash['version'])
|
|
74
|
+
|
|
75
|
+
load_data
|
|
76
|
+
load_error_data
|
|
77
|
+
end
|
|
78
|
+
|
|
79
|
+
def load_data
|
|
80
|
+
@uber_data_hashes.each do |uber_data_hash|
|
|
81
|
+
@uber.add_data(hash_to_data(uber_data_hash))
|
|
82
|
+
end
|
|
83
|
+
end
|
|
84
|
+
|
|
85
|
+
def load_error_data
|
|
86
|
+
@uber_error_data_hashes.each do |uber_error_data_hash|
|
|
87
|
+
@uber.add_data(hash_to_data(uber_error_data_hash))
|
|
88
|
+
end
|
|
89
|
+
end
|
|
90
|
+
|
|
91
|
+
def hash_to_data(hash)
|
|
92
|
+
data_hash = {}
|
|
93
|
+
hash.each do |k, v|
|
|
94
|
+
data_hash[(k.to_sym rescue k)] = v # symbolize_keys
|
|
95
|
+
end
|
|
96
|
+
child_data_hashes = data_hash.delete(:data) || []
|
|
97
|
+
data = Data.new(data_hash)
|
|
98
|
+
child_data_hashes.each do |child_data_hash|
|
|
99
|
+
data.add_data(hash_to_data(child_data_hash))
|
|
100
|
+
end
|
|
101
|
+
data
|
|
102
|
+
end
|
|
103
|
+
end
|
|
104
|
+
end
|
|
@@ -0,0 +1,22 @@
|
|
|
1
|
+
<!doctype html>
|
|
2
|
+
<html>
|
|
3
|
+
<!-- shameless -->
|
|
4
|
+
<head>
|
|
5
|
+
<title>Jason Ronallo</title>
|
|
6
|
+
</head>
|
|
7
|
+
|
|
8
|
+
<body>
|
|
9
|
+
<span itemscope itemtype="http://schema.org/Person"
|
|
10
|
+
itemid="http://ronallo.com#me">
|
|
11
|
+
<a itemprop="url" href="http://twitter.com/ronallo">
|
|
12
|
+
<span itemprop="name">Jason Ronallo</span>
|
|
13
|
+
</a> is the
|
|
14
|
+
<span itemprop="jobTitle">Associate Head of Digital Library Initiatives</span> at
|
|
15
|
+
<span itemprop="affiliation" itemscope itemtype="http://schema.org/Library" itemid="http://lib.ncsu.edu">
|
|
16
|
+
<span itemprop="name">
|
|
17
|
+
<a itemprop="url" href="http://www.lib.ncsu.edu">NCSU Libraries</a>
|
|
18
|
+
</span>
|
|
19
|
+
</span>.
|
|
20
|
+
</span>
|
|
21
|
+
</body>
|
|
22
|
+
</html>
|
|
@@ -0,0 +1,16 @@
|
|
|
1
|
+
<!doctype html>
|
|
2
|
+
<html>
|
|
3
|
+
<head>
|
|
4
|
+
<title>itemref example from Microdata spec</title>
|
|
5
|
+
</head>
|
|
6
|
+
|
|
7
|
+
<body>
|
|
8
|
+
<div itemscope id="amanda" itemref="a b"></div>
|
|
9
|
+
<p id="a">Name: <span itemprop="name">Amanda</span></p>
|
|
10
|
+
<div id="b" itemprop="band" itemscope itemref="c"></div>
|
|
11
|
+
<div id="c">
|
|
12
|
+
<p>Band: <span itemprop="name">Jazz Band</span></p>
|
|
13
|
+
<p>Size: <span itemprop="size">12</span> players</p>
|
|
14
|
+
</div>
|
|
15
|
+
</body>
|
|
16
|
+
</html>
|
|
@@ -0,0 +1,22 @@
|
|
|
1
|
+
<!doctype html>
|
|
2
|
+
<html>
|
|
3
|
+
<!-- shameless -->
|
|
4
|
+
<head>
|
|
5
|
+
<title>Jason Ronallo</title>
|
|
6
|
+
</head>
|
|
7
|
+
|
|
8
|
+
<body>
|
|
9
|
+
<span itemtype="http://schema.org/Person"
|
|
10
|
+
itemid="http://ronallo.com#me">
|
|
11
|
+
<a itemprop="url" href="http://twitter.com/ronallo">
|
|
12
|
+
<span itemprop="name">Jason Ronallo</span>
|
|
13
|
+
</a> is the
|
|
14
|
+
<span itemprop="jobTitle">Associate Head of Digital Library Initiatives</span> at
|
|
15
|
+
<span itemprop="affiliation" itemtype="http://schema.org/Library" itemid="http://lib.ncsu.edu">
|
|
16
|
+
<span itemprop="name">
|
|
17
|
+
<a itemprop="url" href="http://www.lib.ncsu.edu">NCSU Libraries</a>
|
|
18
|
+
</span>
|
|
19
|
+
</span>.
|
|
20
|
+
</span>
|
|
21
|
+
</body>
|
|
22
|
+
</html>
|
data/test/test_helper.rb
ADDED
|
@@ -0,0 +1,19 @@
|
|
|
1
|
+
require 'test_helper'
|
|
2
|
+
|
|
3
|
+
class TestItemref < Test::Unit::TestCase
|
|
4
|
+
|
|
5
|
+
def setup
|
|
6
|
+
@items = Hypermicrodata.get_items('test/data/example_itemref.html')
|
|
7
|
+
end
|
|
8
|
+
|
|
9
|
+
def test_top_item_name
|
|
10
|
+
assert_equal ['Amanda'], @items.first.properties['name']
|
|
11
|
+
end
|
|
12
|
+
|
|
13
|
+
def test_band_name_and_size
|
|
14
|
+
band = @items.first.properties['band'].first
|
|
15
|
+
assert_equal ['Jazz Band'], band.properties['name']
|
|
16
|
+
assert_equal ['12'], band.properties['size']
|
|
17
|
+
end
|
|
18
|
+
|
|
19
|
+
end
|
data/test/test_json.rb
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
1
|
+
require 'test_helper'
|
|
2
|
+
|
|
3
|
+
class TestJson < Test::Unit::TestCase
|
|
4
|
+
|
|
5
|
+
def setup
|
|
6
|
+
json = Hypermicrodata.to_json('test/data/example_itemref.html')
|
|
7
|
+
@top_item = JSON.parse(json)['items'].first
|
|
8
|
+
end
|
|
9
|
+
|
|
10
|
+
def test_json_without_itemtypes
|
|
11
|
+
assert !@top_item.has_key?('type')
|
|
12
|
+
assert !@top_item['properties']['band'].first.has_key?('type')
|
|
13
|
+
end
|
|
14
|
+
|
|
15
|
+
end
|
data/test/test_parse.rb
ADDED
|
@@ -0,0 +1,36 @@
|
|
|
1
|
+
require 'test_helper'
|
|
2
|
+
|
|
3
|
+
class TestParse < Test::Unit::TestCase
|
|
4
|
+
|
|
5
|
+
def setup
|
|
6
|
+
@items = Hypermicrodata.get_items('test/data/example.html')
|
|
7
|
+
end
|
|
8
|
+
|
|
9
|
+
def test_top_item_type
|
|
10
|
+
assert_equal ['http://schema.org/Person'], @items.first.type
|
|
11
|
+
end
|
|
12
|
+
|
|
13
|
+
def test_top_item_id
|
|
14
|
+
assert_equal "http://ronallo.com#me", @items.first.id
|
|
15
|
+
end
|
|
16
|
+
|
|
17
|
+
def test_top_item_properties
|
|
18
|
+
properties = @items.first.properties
|
|
19
|
+
assert_equal ["Jason Ronallo"], properties['name']
|
|
20
|
+
assert_equal ["http://twitter.com/ronallo"], properties['url']
|
|
21
|
+
assert_equal ["Associate Head of Digital Library Initiatives"], properties['jobTitle']
|
|
22
|
+
end
|
|
23
|
+
|
|
24
|
+
def test_nested_item
|
|
25
|
+
item = @items.first.properties['affiliation'][0]
|
|
26
|
+
assert_equal ['http://schema.org/Library'], item.type
|
|
27
|
+
assert_equal "http://lib.ncsu.edu", item.id
|
|
28
|
+
end
|
|
29
|
+
|
|
30
|
+
def test_nested_item_properties
|
|
31
|
+
properties = @items.first.properties['affiliation'][0].properties
|
|
32
|
+
assert_equal ['NCSU Libraries'], properties['name']
|
|
33
|
+
assert_equal ['http://www.lib.ncsu.edu'], properties['url']
|
|
34
|
+
end
|
|
35
|
+
|
|
36
|
+
end
|
metadata
ADDED
|
@@ -0,0 +1,139 @@
|
|
|
1
|
+
--- !ruby/object:Gem::Specification
|
|
2
|
+
name: hypermicrodata
|
|
3
|
+
version: !ruby/object:Gem::Version
|
|
4
|
+
version: 0.1.0
|
|
5
|
+
platform: ruby
|
|
6
|
+
authors:
|
|
7
|
+
- Jason Ronallo
|
|
8
|
+
- Toru KAWAMURA
|
|
9
|
+
autorequire:
|
|
10
|
+
bindir: bin
|
|
11
|
+
cert_chain: []
|
|
12
|
+
|
|
13
|
+
date: 2014-10-13 00:00:00 Z
|
|
14
|
+
dependencies:
|
|
15
|
+
- !ruby/object:Gem::Dependency
|
|
16
|
+
name: nokogiri
|
|
17
|
+
prerelease: false
|
|
18
|
+
requirement: &id001 !ruby/object:Gem::Requirement
|
|
19
|
+
requirements:
|
|
20
|
+
- &id002
|
|
21
|
+
- ">="
|
|
22
|
+
- !ruby/object:Gem::Version
|
|
23
|
+
version: "0"
|
|
24
|
+
type: :runtime
|
|
25
|
+
version_requirements: *id001
|
|
26
|
+
- !ruby/object:Gem::Dependency
|
|
27
|
+
name: mechanize
|
|
28
|
+
prerelease: false
|
|
29
|
+
requirement: &id003 !ruby/object:Gem::Requirement
|
|
30
|
+
requirements:
|
|
31
|
+
- *id002
|
|
32
|
+
type: :runtime
|
|
33
|
+
version_requirements: *id003
|
|
34
|
+
- !ruby/object:Gem::Dependency
|
|
35
|
+
name: halibut
|
|
36
|
+
prerelease: false
|
|
37
|
+
requirement: &id004 !ruby/object:Gem::Requirement
|
|
38
|
+
requirements:
|
|
39
|
+
- *id002
|
|
40
|
+
type: :runtime
|
|
41
|
+
version_requirements: *id004
|
|
42
|
+
- !ruby/object:Gem::Dependency
|
|
43
|
+
name: multi_json
|
|
44
|
+
prerelease: false
|
|
45
|
+
requirement: &id005 !ruby/object:Gem::Requirement
|
|
46
|
+
requirements:
|
|
47
|
+
- *id002
|
|
48
|
+
type: :runtime
|
|
49
|
+
version_requirements: *id005
|
|
50
|
+
- !ruby/object:Gem::Dependency
|
|
51
|
+
name: bundler
|
|
52
|
+
prerelease: false
|
|
53
|
+
requirement: &id006 !ruby/object:Gem::Requirement
|
|
54
|
+
requirements:
|
|
55
|
+
- - ~>
|
|
56
|
+
- !ruby/object:Gem::Version
|
|
57
|
+
version: "1.3"
|
|
58
|
+
type: :development
|
|
59
|
+
version_requirements: *id006
|
|
60
|
+
- !ruby/object:Gem::Dependency
|
|
61
|
+
name: rake
|
|
62
|
+
prerelease: false
|
|
63
|
+
requirement: &id007 !ruby/object:Gem::Requirement
|
|
64
|
+
requirements:
|
|
65
|
+
- *id002
|
|
66
|
+
type: :development
|
|
67
|
+
version_requirements: *id007
|
|
68
|
+
description: HTML5 Microdata extractor with Hypermedia
|
|
69
|
+
email:
|
|
70
|
+
- jronallo@gmail.com
|
|
71
|
+
- tkawa@4bit.net
|
|
72
|
+
executables:
|
|
73
|
+
- hypermicrodata.rb
|
|
74
|
+
extensions: []
|
|
75
|
+
|
|
76
|
+
extra_rdoc_files: []
|
|
77
|
+
|
|
78
|
+
files:
|
|
79
|
+
- .gitignore
|
|
80
|
+
- .travis.yml
|
|
81
|
+
- Gemfile
|
|
82
|
+
- LICENSE.txt
|
|
83
|
+
- README.md
|
|
84
|
+
- Rakefile
|
|
85
|
+
- bin/hypermicrodata.rb
|
|
86
|
+
- hypermicrodata.gemspec
|
|
87
|
+
- lib/hypermicrodata.rb
|
|
88
|
+
- lib/hypermicrodata/document.rb
|
|
89
|
+
- lib/hypermicrodata/extract.rb
|
|
90
|
+
- lib/hypermicrodata/item.rb
|
|
91
|
+
- lib/hypermicrodata/itemprop_parser.rb
|
|
92
|
+
- lib/hypermicrodata/link.rb
|
|
93
|
+
- lib/hypermicrodata/property.rb
|
|
94
|
+
- lib/hypermicrodata/rails/html_based_json_renderer.rb
|
|
95
|
+
- lib/hypermicrodata/serializer/base.rb
|
|
96
|
+
- lib/hypermicrodata/serializer/hal.rb
|
|
97
|
+
- lib/hypermicrodata/serializer/jsonld.rb
|
|
98
|
+
- lib/hypermicrodata/serializer/uber.rb
|
|
99
|
+
- lib/hypermicrodata/submit_button.rb
|
|
100
|
+
- lib/hypermicrodata/version.rb
|
|
101
|
+
- lib/uberous/uber.rb
|
|
102
|
+
- test/data/example.html
|
|
103
|
+
- test/data/example_itemref.html
|
|
104
|
+
- test/data/example_with_no_itemscope.html
|
|
105
|
+
- test/test_helper.rb
|
|
106
|
+
- test/test_itemref.rb
|
|
107
|
+
- test/test_json.rb
|
|
108
|
+
- test/test_parse.rb
|
|
109
|
+
homepage: https://github.com/tkawa/hypermicrodata
|
|
110
|
+
licenses:
|
|
111
|
+
- MIT
|
|
112
|
+
metadata: {}
|
|
113
|
+
|
|
114
|
+
post_install_message:
|
|
115
|
+
rdoc_options: []
|
|
116
|
+
|
|
117
|
+
require_paths:
|
|
118
|
+
- lib
|
|
119
|
+
required_ruby_version: !ruby/object:Gem::Requirement
|
|
120
|
+
requirements:
|
|
121
|
+
- *id002
|
|
122
|
+
required_rubygems_version: !ruby/object:Gem::Requirement
|
|
123
|
+
requirements:
|
|
124
|
+
- *id002
|
|
125
|
+
requirements: []
|
|
126
|
+
|
|
127
|
+
rubyforge_project:
|
|
128
|
+
rubygems_version: 2.0.14
|
|
129
|
+
signing_key:
|
|
130
|
+
specification_version: 4
|
|
131
|
+
summary: Ruby library for extracting HTML5 Microdata with Hypermedia
|
|
132
|
+
test_files:
|
|
133
|
+
- test/data/example.html
|
|
134
|
+
- test/data/example_itemref.html
|
|
135
|
+
- test/data/example_with_no_itemscope.html
|
|
136
|
+
- test/test_helper.rb
|
|
137
|
+
- test/test_itemref.rb
|
|
138
|
+
- test/test_json.rb
|
|
139
|
+
- test/test_parse.rb
|