craft 0.0.2 → 0.1.0
Sign up to get free protection for your applications and to get access to all the features.
- data/.travis.yml +5 -0
- data/CHANGES.md +8 -0
- data/Gemfile +9 -0
- data/README.md +6 -1
- data/lib/craft.rb +57 -21
- data/lib/craft/version.rb +1 -1
- data/spec/craft_spec.rb +93 -20
- metadata +6 -4
data/.travis.yml
ADDED
data/CHANGES.md
ADDED
@@ -0,0 +1,8 @@
|
|
1
|
+
v0.1.0
|
2
|
+
-----------
|
3
|
+
|
4
|
+
- When nesting, make parent accessible to child.
|
5
|
+
- Craft#attributes returns the attributes.
|
6
|
+
- Craft.stub stubs a static or dynamic value that doesn't need to be parsed in
|
7
|
+
the document.
|
8
|
+
- Transforms are executed in the context of the crafted object.
|
data/Gemfile
CHANGED
data/README.md
CHANGED
@@ -1,4 +1,4 @@
|
|
1
|
-
# Craft
|
1
|
+
# Craft [![Build Status](https://secure.travis-ci.org/papercavalier/craft.png)](http://travis-ci.org/papercavalier/craft)
|
2
2
|
|
3
3
|
Craft XML and HTML into objects.
|
4
4
|
|
@@ -16,6 +16,9 @@ class Page < Craft
|
|
16
16
|
|
17
17
|
# Perform transforms on returned nodes
|
18
18
|
many :images, 'img', lambda { |img| img.attr('src').upcase }
|
19
|
+
|
20
|
+
# Stub attributes that don't need to be parsed
|
21
|
+
stub :spidered_at, lambda { Time.now }
|
19
22
|
end
|
20
23
|
|
21
24
|
page = Page.parse open('http://www.google.com')
|
@@ -24,6 +27,8 @@ page.title #=> 'Google'
|
|
24
27
|
page.links #=> ['http://www.google.com/imghp?hl=en&tab=wi', ...]
|
25
28
|
page.images #=> ['/LOGOS/2012/MOBY_DICK12-HP.JPG']
|
26
29
|
|
30
|
+
page.attributes #=> { :title => 'Google', :links => ... }
|
31
|
+
|
27
32
|
class Script < Craft
|
28
33
|
one :body, 'text()'
|
29
34
|
end
|
data/lib/craft.rb
CHANGED
@@ -7,6 +7,7 @@ require 'nokogiri'
|
|
7
7
|
#
|
8
8
|
# module Transformations
|
9
9
|
# IntegerTransform = lambda { |n| Integer n.text }
|
10
|
+
# Timestamp = lambda { Time.now }
|
10
11
|
# end
|
11
12
|
#
|
12
13
|
# class Person < Craft
|
@@ -15,46 +16,48 @@ require 'nokogiri'
|
|
15
16
|
# one :name, 'div.name'
|
16
17
|
# one :age, 'div.age', IntegerTransform
|
17
18
|
# many :friends, 'li.friend', Person
|
19
|
+
# stub :created_at, Timestamp
|
18
20
|
# end
|
19
21
|
#
|
20
22
|
class Craft
|
21
23
|
class << self
|
22
|
-
#
|
23
|
-
|
24
|
-
alias call new
|
24
|
+
# Returns an Array of names for the attributes defined in the class.
|
25
|
+
attr :attribute_names
|
25
26
|
|
26
|
-
# Define
|
27
|
+
# Define an attribute that extracts a collection of values from a parsed
|
27
28
|
# document.
|
28
29
|
#
|
29
|
-
# name - The Symbol name of the
|
30
|
+
# name - The Symbol name of the attribute.
|
30
31
|
# paths - One or more String XPath of CSS queries. An optional Proc
|
31
32
|
# transformation on the extracted value may be appended. If none is
|
32
33
|
# appended, the default transformation returns the stripped String
|
33
34
|
# value of the node.
|
34
35
|
#
|
35
|
-
# Returns
|
36
|
+
# Returns nothing.
|
36
37
|
def many(name, *paths)
|
37
|
-
transform =
|
38
|
+
transform = pop_transform_from_paths paths
|
39
|
+
@attribute_names << name
|
38
40
|
|
39
41
|
define_method name do
|
40
|
-
@node.search(*paths).map { |node|
|
42
|
+
@node.search(*paths).map { |node| instance_exec node, &transform }
|
41
43
|
end
|
42
44
|
end
|
43
45
|
|
44
|
-
# Define
|
46
|
+
# Define an attribute that extracts a single value from a parsed document.
|
45
47
|
#
|
46
|
-
# name - The Symbol name of the
|
48
|
+
# name - The Symbol name of the attribute.
|
47
49
|
# paths - One or more String XPath of CSS queries. An optional Proc
|
48
50
|
# transformation on the extracted value may be appended. If none is
|
49
51
|
# appended, the default transformation returns the stripped String
|
50
52
|
# value of the node.
|
51
53
|
#
|
52
|
-
# Returns
|
54
|
+
# Returns nothing.
|
53
55
|
def one(name, *paths)
|
54
|
-
transform =
|
56
|
+
transform = pop_transform_from_paths paths
|
57
|
+
@attribute_names << name
|
55
58
|
|
56
59
|
define_method name do
|
57
|
-
|
60
|
+
instance_exec @node.at(*paths), &transform
|
58
61
|
end
|
59
62
|
end
|
60
63
|
|
@@ -67,25 +70,58 @@ class Craft
|
|
67
70
|
new Nokogiri body
|
68
71
|
end
|
69
72
|
|
73
|
+
# Define an attribute that returns a value without parsing the document.
|
74
|
+
#
|
75
|
+
# name - The Symbol name of the attribute.
|
76
|
+
# value - Some value the attribute should return. If given a Proc, the
|
77
|
+
# value will be generated dynamically (default: nil).
|
78
|
+
#
|
79
|
+
# Returns nothing.
|
80
|
+
def stub(name, value = nil)
|
81
|
+
@attribute_names << name
|
82
|
+
|
83
|
+
define_method name do
|
84
|
+
value.respond_to?(:call) ? instance_exec(&value) : value
|
85
|
+
end
|
86
|
+
end
|
87
|
+
|
88
|
+
def to_proc
|
89
|
+
klass = self
|
90
|
+
->(node) { klass.new node, self }
|
91
|
+
end
|
92
|
+
|
70
93
|
private
|
71
94
|
|
72
|
-
def
|
73
|
-
|
95
|
+
def inherited(subclass)
|
96
|
+
subclass.instance_variable_set :@attribute_names, []
|
97
|
+
end
|
98
|
+
|
99
|
+
def pop_transform_from_paths(array)
|
100
|
+
if array.last.respond_to? :to_proc
|
74
101
|
array.pop
|
75
102
|
else
|
76
|
-
|
77
|
-
def self.call(node)
|
78
|
-
node.text.strip if node
|
79
|
-
end
|
80
|
-
end
|
103
|
+
->(node) { node.text.strip if node }
|
81
104
|
end
|
82
105
|
end
|
83
106
|
end
|
84
107
|
|
108
|
+
attr :parent
|
109
|
+
|
85
110
|
# Craft a new object.
|
86
111
|
#
|
87
112
|
# node - A Nokogiri::XML::Node.
|
88
|
-
def initialize(node)
|
113
|
+
def initialize(node, parent = nil)
|
89
114
|
@node = node
|
115
|
+
@parent = parent
|
116
|
+
end
|
117
|
+
|
118
|
+
# Returns the Hash attributes.
|
119
|
+
def attributes
|
120
|
+
Hash[attribute_names.map { |key| [key, self.send(key)] }]
|
121
|
+
end
|
122
|
+
|
123
|
+
# Returns an Array of names for the attributes on this object.
|
124
|
+
def attribute_names
|
125
|
+
self.class.attribute_names
|
90
126
|
end
|
91
127
|
end
|
data/lib/craft/version.rb
CHANGED
data/spec/craft_spec.rb
CHANGED
@@ -3,45 +3,89 @@ require 'minitest/autorun'
|
|
3
3
|
require 'craft'
|
4
4
|
|
5
5
|
describe Craft do
|
6
|
-
let
|
7
|
-
|
8
|
-
|
6
|
+
let(:html) { '<html><ul><li>1</li><li>2</li>' }
|
7
|
+
let(:klass) { Class.new Craft }
|
8
|
+
let(:instance) { klass.parse html }
|
9
9
|
|
10
|
-
|
11
|
-
|
12
|
-
|
10
|
+
describe '.attribute_names' do
|
11
|
+
it 'is empty by default' do
|
12
|
+
klass.attribute_names.must_equal []
|
13
|
+
end
|
13
14
|
|
14
|
-
|
15
|
-
|
15
|
+
it 'does not reference other attribute names' do
|
16
|
+
klass.stub :foo
|
17
|
+
other = Class.new(Craft) { stub :bar }
|
18
|
+
klass.attribute_names.wont_equal other.attribute_names
|
19
|
+
end
|
16
20
|
end
|
17
21
|
|
18
22
|
describe '.many' do
|
19
23
|
it 'extracts nodes' do
|
20
|
-
klass.many
|
24
|
+
klass.many :foo, 'li'
|
21
25
|
instance.foo.must_equal %w(1 2)
|
22
26
|
end
|
23
27
|
|
24
28
|
it 'transforms' do
|
25
|
-
klass.many
|
29
|
+
klass.many :foo, 'li', ->(node) { node.text.to_i }
|
26
30
|
instance.foo.must_equal [1, 2]
|
27
31
|
end
|
32
|
+
|
33
|
+
it 'transforms in scope' do
|
34
|
+
klass.many :foo, 'li', ->(node) { bar }
|
35
|
+
klass.send(:define_method, :bar) { 'bar' }
|
36
|
+
instance.foo.must_equal ['bar', 'bar']
|
37
|
+
end
|
38
|
+
|
39
|
+
it 'stores attribute name' do
|
40
|
+
klass.many :foo, 'li'
|
41
|
+
klass.attribute_names.must_include :foo
|
42
|
+
end
|
43
|
+
|
44
|
+
it 'nests' do
|
45
|
+
klass.many :foo, 'ul', Class.new(Craft)
|
46
|
+
instance.foo.each { |attr| attr.must_be_kind_of Craft }
|
47
|
+
end
|
48
|
+
|
49
|
+
it 'has a parent when nested' do
|
50
|
+
klass.many :foo, 'li', Class.new(Craft)
|
51
|
+
instance.foo.each { |attr| attr.parent.must_equal instance }
|
52
|
+
end
|
28
53
|
end
|
29
54
|
|
30
55
|
describe '.one' do
|
31
56
|
it 'extracts a node' do
|
32
|
-
klass.one
|
57
|
+
klass.one :foo, 'li'
|
33
58
|
instance.foo.must_equal '1'
|
34
59
|
end
|
35
60
|
|
36
61
|
it 'transforms' do
|
37
|
-
klass.one
|
62
|
+
klass.one :foo, 'li', ->(node) { node.text.to_i }
|
38
63
|
instance.foo.must_equal 1
|
39
64
|
end
|
40
65
|
|
66
|
+
it 'transforms in scope' do
|
67
|
+
klass.one :foo, 'li', ->(node) { bar }
|
68
|
+
klass.send(:define_method, :bar) { 'bar' }
|
69
|
+
instance.foo.must_equal 'bar'
|
70
|
+
end
|
71
|
+
|
72
|
+
it 'stores attribute name' do
|
73
|
+
klass.one :foo, 'li'
|
74
|
+
klass.attribute_names.must_include :foo
|
75
|
+
end
|
76
|
+
|
77
|
+
it 'nests' do
|
78
|
+
klass.one :foo, 'ul', Class.new(Craft)
|
79
|
+
instance.foo.must_be_kind_of Craft
|
80
|
+
end
|
81
|
+
|
82
|
+
it 'has a parent when nested' do
|
83
|
+
klass.one :foo, 'li', Class.new(Craft)
|
84
|
+
instance.foo.parent.must_equal instance
|
85
|
+
end
|
86
|
+
|
41
87
|
describe 'given no matches' do
|
42
|
-
before
|
43
|
-
klass.one 'foo', 'foo'
|
44
|
-
end
|
88
|
+
before { klass.one :foo, 'bar' }
|
45
89
|
|
46
90
|
it 'returns nil' do
|
47
91
|
instance.foo.must_be_nil
|
@@ -49,10 +93,39 @@ describe Craft do
|
|
49
93
|
end
|
50
94
|
end
|
51
95
|
|
52
|
-
|
53
|
-
|
54
|
-
|
55
|
-
|
56
|
-
|
96
|
+
describe '.stub' do
|
97
|
+
it 'returns nil by default' do
|
98
|
+
klass.stub :foo
|
99
|
+
instance.foo.must_be_nil
|
100
|
+
end
|
101
|
+
|
102
|
+
it 'returns a static value' do
|
103
|
+
klass.stub :foo, 1
|
104
|
+
instance.foo.must_equal 1
|
105
|
+
end
|
106
|
+
|
107
|
+
it 'returns a dynamic value' do
|
108
|
+
klass.stub :foo, -> { Time.now }
|
109
|
+
instance.foo.must_be_instance_of Time
|
110
|
+
end
|
111
|
+
|
112
|
+
it 'transforms in scope' do
|
113
|
+
klass.stub :foo, -> { bar }
|
114
|
+
klass.send(:define_method, :bar) { 'bar' }
|
115
|
+
instance.foo.must_equal 'bar'
|
116
|
+
end
|
117
|
+
|
118
|
+
it 'stores attribute name' do
|
119
|
+
klass.stub :foo
|
120
|
+
klass.attribute_names.must_include :foo
|
121
|
+
end
|
122
|
+
end
|
123
|
+
|
124
|
+
describe '#attributes' do
|
125
|
+
it 'returns attributes' do
|
126
|
+
klass.stub :foo
|
127
|
+
klass.one :bar, 'li'
|
128
|
+
instance.attributes.must_equal({ foo: nil, bar: '1' })
|
129
|
+
end
|
57
130
|
end
|
58
131
|
end
|
metadata
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: craft
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.0
|
4
|
+
version: 0.1.0
|
5
5
|
prerelease:
|
6
6
|
platform: ruby
|
7
7
|
authors:
|
@@ -10,7 +10,7 @@ authors:
|
|
10
10
|
autorequire:
|
11
11
|
bindir: bin
|
12
12
|
cert_chain: []
|
13
|
-
date: 2012-10-
|
13
|
+
date: 2012-10-25 00:00:00.000000000 Z
|
14
14
|
dependencies:
|
15
15
|
- !ruby/object:Gem::Dependency
|
16
16
|
name: nokogiri
|
@@ -36,6 +36,8 @@ extensions: []
|
|
36
36
|
extra_rdoc_files: []
|
37
37
|
files:
|
38
38
|
- .gitignore
|
39
|
+
- .travis.yml
|
40
|
+
- CHANGES.md
|
39
41
|
- Gemfile
|
40
42
|
- LICENSE.txt
|
41
43
|
- README.md
|
@@ -58,7 +60,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
|
|
58
60
|
version: '0'
|
59
61
|
segments:
|
60
62
|
- 0
|
61
|
-
hash:
|
63
|
+
hash: 2821744877133646519
|
62
64
|
required_rubygems_version: !ruby/object:Gem::Requirement
|
63
65
|
none: false
|
64
66
|
requirements:
|
@@ -67,7 +69,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
|
|
67
69
|
version: '0'
|
68
70
|
segments:
|
69
71
|
- 0
|
70
|
-
hash:
|
72
|
+
hash: 2821744877133646519
|
71
73
|
requirements: []
|
72
74
|
rubyforge_project:
|
73
75
|
rubygems_version: 1.8.23
|