htmltools 1.10
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- data/INSTALL +58 -0
- data/README +162 -0
- data/demo/degolive.rb +89 -0
- data/demo/ebaySearch.rb +93 -0
- data/demo/xpath.rb +62 -0
- data/lib/html/element.rb +323 -0
- data/lib/html/rexml-nodepath.rb +49 -0
- data/lib/html/sgml-parser.rb +372 -0
- data/lib/html/stparser.rb +280 -0
- data/lib/html/tags.rb +288 -0
- data/lib/html/tree.rb +140 -0
- data/lib/html/xmltree.rb +173 -0
- data/lib/html/xpath.rb +72 -0
- data/test/suite.rb +5 -0
- data/test/tc_html-element.rb +73 -0
- data/test/tc_html-tree.rb +201 -0
- data/test/tc_source-parser.rb +160 -0
- data/test/tc_stacking-parser.rb +196 -0
- data/test/tc_xpath.rb +87 -0
- metadata +58 -0
data/test/tc_xpath.rb
ADDED
@@ -0,0 +1,87 @@
|
|
1
|
+
#!/usr/bin/ruby
|
2
|
+
# Copyright:: Copyright (C) 2002, Ned Konz <ned@bike-nomad.com>
|
3
|
+
# License:: Same as Ruby's
|
4
|
+
# CVS ID: $Id: tc_xpath.rb,v 1.6 2004/09/10 17:29:37 jhannes Exp $
|
5
|
+
#
|
6
|
+
# This is the test file for the HTMLTree XPath interface.
|
7
|
+
#
|
8
|
+
require 'html/tree'
|
9
|
+
require 'html/xpath'
|
10
|
+
require 'test/unit'
|
11
|
+
require 'html/rexml-nodepath'
|
12
|
+
|
13
|
+
|
14
|
+
$sample1 = %q{
|
15
|
+
<html>
|
16
|
+
<head>
|
17
|
+
</head>
|
18
|
+
<body>
|
19
|
+
<ul compact="compact">
|
20
|
+
<li>An item</li>
|
21
|
+
</ul>
|
22
|
+
<hr>
|
23
|
+
<img src="http://wherever">
|
24
|
+
<p class='123'>A < paragraph
|
25
|
+
<p class='234'>Another paragraph
|
26
|
+
</body>
|
27
|
+
</html>
|
28
|
+
}
|
29
|
+
|
30
|
+
class XPathTestCase < Test::Unit::TestCase
|
31
|
+
attr_accessor :p
|
32
|
+
|
33
|
+
def setup
|
34
|
+
@p = HTMLTree::Parser.new(true, true)
|
35
|
+
end
|
36
|
+
|
37
|
+
def test_basic
|
38
|
+
p.feed("<html></html>")
|
39
|
+
t = p.tree
|
40
|
+
assert_same(HTMLTree::Document, t.class)
|
41
|
+
d = t.as_rexml_document
|
42
|
+
assert_same(REXML::Document, d.class)
|
43
|
+
end
|
44
|
+
|
45
|
+
def test_match1
|
46
|
+
p.feed($sample1)
|
47
|
+
d = p.tree.as_rexml_document
|
48
|
+
m = d.root.get_elements( '//p' )
|
49
|
+
assert_equal(2, m.size)
|
50
|
+
assert_equal('p', m.first.name)
|
51
|
+
end
|
52
|
+
|
53
|
+
def test_match2
|
54
|
+
p.feed($sample1)
|
55
|
+
m = match('/html/body/p')
|
56
|
+
assert_equal(2, m.size)
|
57
|
+
assert_equal('p', m.first.name)
|
58
|
+
assert_equal('123', m[0].attributes['class'])
|
59
|
+
assert_equal('234', m[1].attributes['class'])
|
60
|
+
end
|
61
|
+
|
62
|
+
def test_match_all
|
63
|
+
p.feed($sample1)
|
64
|
+
html = p.tree.html_node
|
65
|
+
m = html.rexml_match('descendant::node()')
|
66
|
+
#m.each { |ea| puts ea.full_path + " --> " + ea.to_s }
|
67
|
+
assert_equal(11, m.size)
|
68
|
+
end
|
69
|
+
|
70
|
+
def test_find_by_attribute
|
71
|
+
p.feed($sample1)
|
72
|
+
assert_equal("Another paragraph", match("/html/body/p[@class = '234']/text()").to_s)
|
73
|
+
end
|
74
|
+
|
75
|
+
def test_show_attribute
|
76
|
+
p.feed("<html><body><ol>" +
|
77
|
+
"<li item='1'><a href='http://wrong.com'>test</a></li>" +
|
78
|
+
"<li item='2'><a href='http://test.com'>test</a></li>" +
|
79
|
+
"<li item='3'><a href='http://wrong.com'>test</a></li>" +
|
80
|
+
"</ol></body></html>")
|
81
|
+
assert_equal('http://test.com', match("//li[@item = '2']/a/@href")[0].value)
|
82
|
+
end
|
83
|
+
|
84
|
+
def match(xpath)
|
85
|
+
p.tree.rexml_match(xpath)
|
86
|
+
end
|
87
|
+
end
|
metadata
ADDED
@@ -0,0 +1,58 @@
|
|
1
|
+
--- !ruby/object:Gem::Specification
|
2
|
+
rubygems_version: 0.8.11
|
3
|
+
specification_version: 1
|
4
|
+
name: htmltools
|
5
|
+
version: !ruby/object:Gem::Version
|
6
|
+
version: "1.10"
|
7
|
+
date: 2006-07-25 00:00:00 +02:00
|
8
|
+
summary: This is a Ruby library for building trees representing HTML structure.
|
9
|
+
require_paths:
|
10
|
+
- lib
|
11
|
+
- lib
|
12
|
+
email: johannes@brodwall.com
|
13
|
+
homepage: http://ruby-htmltools.rubyforge.org/
|
14
|
+
rubyforge_project:
|
15
|
+
description:
|
16
|
+
autorequire: html/tree
|
17
|
+
default_executable:
|
18
|
+
bindir: bin
|
19
|
+
has_rdoc: false
|
20
|
+
required_ruby_version: !ruby/object:Gem::Version::Requirement
|
21
|
+
requirements:
|
22
|
+
-
|
23
|
+
- ">"
|
24
|
+
- !ruby/object:Gem::Version
|
25
|
+
version: 0.0.0
|
26
|
+
version:
|
27
|
+
platform: ruby
|
28
|
+
signing_key:
|
29
|
+
cert_chain:
|
30
|
+
authors:
|
31
|
+
- Johannes Brodwall
|
32
|
+
files:
|
33
|
+
- demo/degolive.rb
|
34
|
+
- demo/ebaySearch.rb
|
35
|
+
- demo/xpath.rb
|
36
|
+
- INSTALL
|
37
|
+
- lib/html/element.rb
|
38
|
+
- lib/html/rexml-nodepath.rb
|
39
|
+
- lib/html/sgml-parser.rb
|
40
|
+
- lib/html/stparser.rb
|
41
|
+
- lib/html/tags.rb
|
42
|
+
- lib/html/tree.rb
|
43
|
+
- lib/html/xmltree.rb
|
44
|
+
- lib/html/xpath.rb
|
45
|
+
- README
|
46
|
+
- test/tc_html-element.rb
|
47
|
+
- test/tc_html-tree.rb
|
48
|
+
- test/tc_stacking-parser.rb
|
49
|
+
- test/tc_xpath.rb
|
50
|
+
- test/tc_source-parser.rb
|
51
|
+
test_files:
|
52
|
+
- test/suite.rb
|
53
|
+
rdoc_options: []
|
54
|
+
extra_rdoc_files: []
|
55
|
+
executables: []
|
56
|
+
extensions: []
|
57
|
+
requirements: []
|
58
|
+
dependencies: []
|