scrubyt 0.1.0 → 0.1.9

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.
Files changed (19) hide show
  1. data/CHANGELOG +34 -0
  2. data/COPYING +340 -0
  3. data/README +34 -5
  4. data/Rakefile +6 -5
  5. data/lib/scrubyt.rb +1 -0
  6. data/lib/scrubyt/constraint.rb +12 -24
  7. data/lib/scrubyt/constraint_adder.rb +3 -17
  8. data/lib/scrubyt/export.rb +33 -17
  9. data/lib/scrubyt/extractor.rb +74 -23
  10. data/lib/scrubyt/filter.rb +52 -37
  11. data/lib/scrubyt/pattern.rb +74 -30
  12. data/lib/scrubyt/post_processor.rb +58 -0
  13. data/lib/scrubyt/result.rb +2 -2
  14. data/lib/scrubyt/result_dumper.rb +6 -0
  15. data/lib/scrubyt/xpathutils.rb +52 -15
  16. data/test/unittests/constraint_test.rb +0 -3
  17. data/test/unittests/extractor_test.rb +11 -13
  18. data/test/unittests/xpathutils_test.rb +31 -31
  19. metadata +8 -5
data/test/unittests/xpathutils_test.rb CHANGED
@@ -94,15 +94,15 @@ class XPathUtilsTest < Test::Unit::TestCase
94
94
  end
95
95
 
96
96
  def test_find_index
97
- assert_equal(Scrubyt::XPathUtils.find_index(@a), 0)
98
- assert_equal(Scrubyt::XPathUtils.find_index(@b), 0)
99
- assert_equal(Scrubyt::XPathUtils.find_index(@c), 0)
100
- assert_equal(Scrubyt::XPathUtils.find_index(@d), 0)
101
- assert_equal(Scrubyt::XPathUtils.find_index(@n_1), 0)
102
- assert_equal(Scrubyt::XPathUtils.find_index(@n_2), 1)
103
- assert_equal(Scrubyt::XPathUtils.find_index(@n_3), 2)
104
- assert_equal(Scrubyt::XPathUtils.find_index(@n_4), 3)
105
- assert_equal(Scrubyt::XPathUtils.find_index(@r), 0)
97
+ assert_equal(Scrubyt::XPathUtils.find_index(@a), 1)
98
+ assert_equal(Scrubyt::XPathUtils.find_index(@b), 1)
99
+ assert_equal(Scrubyt::XPathUtils.find_index(@c), 1)
100
+ assert_equal(Scrubyt::XPathUtils.find_index(@d), 1)
101
+ assert_equal(Scrubyt::XPathUtils.find_index(@n_1), 1)
102
+ assert_equal(Scrubyt::XPathUtils.find_index(@n_2), 2)
103
+ assert_equal(Scrubyt::XPathUtils.find_index(@n_3), 3)
104
+ assert_equal(Scrubyt::XPathUtils.find_index(@n_4), 4)
105
+ assert_equal(Scrubyt::XPathUtils.find_index(@r), 1)
106
106
  end
107
107
 
108
108
  def test_generate_XPath
@@ -121,32 +121,32 @@ class XPathUtilsTest < Test::Unit::TestCase
121
121
  end
122
122
 
123
123
  def test_generate_XPath_with_indices
124
- assert_equal(Scrubyt::XPathUtils.generate_XPath(@a, nil, true), "/a[0]")
125
- assert_equal(Scrubyt::XPathUtils.generate_XPath(@b, nil, true), "/a[0]/b[0]")
126
- assert_equal(Scrubyt::XPathUtils.generate_XPath(@c, nil, true), "/a[0]/b[0]/c[0]")
127
- assert_equal(Scrubyt::XPathUtils.generate_XPath(@d, nil, true), "/a[0]/b[0]/d[0]")
128
- assert_equal(Scrubyt::XPathUtils.generate_XPath(@e, nil, true), "/a[0]/b[0]/e[0]")
129
- assert_equal(Scrubyt::XPathUtils.generate_XPath(@f, nil, true), "/a[0]/b[0]/e[0]/f[0]")
130
- assert_equal(Scrubyt::XPathUtils.generate_XPath(@n_1, nil, true), "/a[0]/b[0]/e[0]/n[0]")
131
- assert_equal(Scrubyt::XPathUtils.generate_XPath(@n_2, nil, true), "/a[0]/b[0]/e[0]/n[1]")
132
- assert_equal(Scrubyt::XPathUtils.generate_XPath(@n_3, nil, true), "/a[0]/b[0]/e[0]/n[2]")
133
- assert_equal(Scrubyt::XPathUtils.generate_XPath(@n_4, nil, true), "/a[0]/b[0]/e[0]/n[3]")
134
- assert_equal(Scrubyt::XPathUtils.generate_XPath(@p, nil, true), "/a[0]/b[0]/e[0]/p[0]")
135
- assert_equal(Scrubyt::XPathUtils.generate_XPath(@r, nil, true), "/a[0]/b[0]/e[0]/n[2]/r[0]")
124
+ assert_equal(Scrubyt::XPathUtils.generate_XPath(@a, nil, true), "/a[1]")
125
+ assert_equal(Scrubyt::XPathUtils.generate_XPath(@b, nil, true), "/a[1]/b[1]")
126
+ assert_equal(Scrubyt::XPathUtils.generate_XPath(@c, nil, true), "/a[1]/b[1]/c[1]")
127
+ assert_equal(Scrubyt::XPathUtils.generate_XPath(@d, nil, true), "/a[1]/b[1]/d[1]")
128
+ assert_equal(Scrubyt::XPathUtils.generate_XPath(@e, nil, true), "/a[1]/b[1]/e[1]")
129
+ assert_equal(Scrubyt::XPathUtils.generate_XPath(@f, nil, true), "/a[1]/b[1]/e[1]/f[1]")
130
+ assert_equal(Scrubyt::XPathUtils.generate_XPath(@n_1, nil, true), "/a[1]/b[1]/e[1]/n[1]")
131
+ assert_equal(Scrubyt::XPathUtils.generate_XPath(@n_2, nil, true), "/a[1]/b[1]/e[1]/n[2]")
132
+ assert_equal(Scrubyt::XPathUtils.generate_XPath(@n_3, nil, true), "/a[1]/b[1]/e[1]/n[3]")
133
+ assert_equal(Scrubyt::XPathUtils.generate_XPath(@n_4, nil, true), "/a[1]/b[1]/e[1]/n[4]")
134
+ assert_equal(Scrubyt::XPathUtils.generate_XPath(@p, nil, true), "/a[1]/b[1]/e[1]/p[1]")
135
+ assert_equal(Scrubyt::XPathUtils.generate_XPath(@r, nil, true), "/a[1]/b[1]/e[1]/n[3]/r[1]")
136
136
  end
137
137
 
138
138
  def test_generate_relative_XPath
139
139
  assert_nil(Scrubyt::XPathUtils.generate_relative_XPath(@a,@a))
140
- assert_equal(Scrubyt::XPathUtils.generate_relative_XPath(@b, @a), "/b[0]")
141
- assert_equal(Scrubyt::XPathUtils.generate_relative_XPath(@c, @a), "/b[0]/c[0]")
142
- assert_equal(Scrubyt::XPathUtils.generate_relative_XPath(@d, @a), "/b[0]/d[0]")
143
- assert_equal(Scrubyt::XPathUtils.generate_relative_XPath(@f, @a), "/b[0]/e[0]/f[0]")
144
- assert_equal(Scrubyt::XPathUtils.generate_relative_XPath(@n_1, @a), "/b[0]/e[0]/n[0]")
145
- assert_equal(Scrubyt::XPathUtils.generate_relative_XPath(@n_2, @a), "/b[0]/e[0]/n[1]")
146
- assert_equal(Scrubyt::XPathUtils.generate_relative_XPath(@n_3, @a), "/b[0]/e[0]/n[2]")
147
- assert_equal(Scrubyt::XPathUtils.generate_relative_XPath(@n_4, @a), "/b[0]/e[0]/n[3]")
148
- assert_equal(Scrubyt::XPathUtils.generate_relative_XPath(@r, @b), "/e[0]/n[2]/r[0]")
149
- assert_equal(Scrubyt::XPathUtils.generate_relative_XPath(@q, @e), "/n[2]/q[0]")
140
+ assert_equal(Scrubyt::XPathUtils.generate_relative_XPath(@b, @a), "/b[1]")
141
+ assert_equal(Scrubyt::XPathUtils.generate_relative_XPath(@c, @a), "/b[1]/c[1]")
142
+ assert_equal(Scrubyt::XPathUtils.generate_relative_XPath(@d, @a), "/b[1]/d[1]")
143
+ assert_equal(Scrubyt::XPathUtils.generate_relative_XPath(@f, @a), "/b[1]/e[1]/f[1]")
144
+ assert_equal(Scrubyt::XPathUtils.generate_relative_XPath(@n_1, @a), "/b[1]/e[1]/n[1]")
145
+ assert_equal(Scrubyt::XPathUtils.generate_relative_XPath(@n_2, @a), "/b[1]/e[1]/n[2]")
146
+ assert_equal(Scrubyt::XPathUtils.generate_relative_XPath(@n_3, @a), "/b[1]/e[1]/n[3]")
147
+ assert_equal(Scrubyt::XPathUtils.generate_relative_XPath(@n_4, @a), "/b[1]/e[1]/n[4]")
148
+ assert_equal(Scrubyt::XPathUtils.generate_relative_XPath(@r, @b), "/e[1]/n[3]/r[1]")
149
+ assert_equal(Scrubyt::XPathUtils.generate_relative_XPath(@q, @e), "/n[3]/q[1]")
150
150
 
151
151
  assert_nil(Scrubyt::XPathUtils.generate_relative_XPath(@r, @n_2))
152
152
  assert_nil(Scrubyt::XPathUtils.generate_relative_XPath(@q, @g))
metadata CHANGED
@@ -3,19 +3,19 @@ rubygems_version: 0.9.0
3
3
  specification_version: 1
4
4
  name: scrubyt
5
5
  version: !ruby/object:Gem::Version
6
- version: 0.1.0
7
- date: 2007-01-15 00:00:00 +01:00
6
+ version: 0.1.9
7
+ date: 2007-01-28 00:00:00 +01:00
8
8
  summary: A powerful Web-scraping framework
9
9
  require_paths:
10
10
  - lib
11
11
  email: peter@rubyrailways.com
12
- homepage: http://www.scrubyt.rubyforge.org
12
+ homepage: http://www.scrubyt.org
13
13
  rubyforge_project:
14
14
  description: scRUBYt! is an easy to learn and use, yet powerful and effective web scraping framework. It's most interesting part is a Web-scraping DSL built on HPricot and WWW::Mechanize, which allows to navigate to the page of interest, then extract and query data records with a few lines of code. It is hard to describe scRUBYt! in a few sentences - you have to see it for yourself!
15
15
  autorequire:
16
16
  default_executable:
17
17
  bindir: bin
18
- has_rdoc: false
18
+ has_rdoc: "true"
19
19
  required_ruby_version: !ruby/object:Gem::Version::Requirement
20
20
  requirements:
21
21
  - - ">"
@@ -30,6 +30,8 @@ authors:
30
30
  - Peter Szinek
31
31
  files:
32
32
  - README
33
+ - COPYING
34
+ - CHANGELOG
33
35
  - Rakefile
34
36
  - lib/scrubyt.rb
35
37
  - lib/scrubyt/constraint_adder.rb
@@ -40,10 +42,11 @@ files:
40
42
  - lib/scrubyt/filter.rb
41
43
  - lib/scrubyt/pattern.rb
42
44
  - lib/scrubyt/result.rb
45
+ - lib/scrubyt/post_processor.rb
43
46
  - lib/scrubyt/xpathutils.rb
44
47
  test_files:
45
- - test/unittests/filter_test.rb
46
48
  - test/unittests/input
49
+ - test/unittests/filter_test.rb
47
50
  - test/unittests/extractor_test.rb
48
51
  - test/unittests/xpathutils_test.rb
49
52
  - test/unittests/constraint_test.rb