alexrabarts-term_extraction 0.1.0
Sign up to get free protection for your applications and to get access to all the features.
- data/VERSION.yml +4 -0
- data/lib/term_extraction.rb +12 -0
- data/lib/term_extraction/yahoo.rb +51 -0
- data/lib/term_extraction/zemanta.rb +49 -0
- data/test/fixtures/yahoo.xml +3 -0
- data/test/fixtures/zemanta.xml +104 -0
- data/test/term_extraction_test.rb +15 -0
- data/test/test_helper.rb +13 -0
- metadata +72 -0
data/VERSION.yml
ADDED
@@ -0,0 +1,51 @@
|
|
1
|
+
require 'nokogiri'
|
2
|
+
|
3
|
+
class TermExtraction
|
4
|
+
class Yahoo < TermExtraction
|
5
|
+
def terms
|
6
|
+
terms = []
|
7
|
+
|
8
|
+
unless data.nil?
|
9
|
+
data.search('//s:Result', ns).each do |n|
|
10
|
+
terms << n.text
|
11
|
+
end
|
12
|
+
end
|
13
|
+
|
14
|
+
terms
|
15
|
+
end
|
16
|
+
|
17
|
+
class << self
|
18
|
+
def canonical_name
|
19
|
+
'yahoo'
|
20
|
+
end
|
21
|
+
end
|
22
|
+
|
23
|
+
private
|
24
|
+
def ns
|
25
|
+
{'s' => 'urn:yahoo:cate'}
|
26
|
+
end
|
27
|
+
|
28
|
+
def gateway
|
29
|
+
'http://search.yahooapis.com/ContentAnalysisService/V1/termExtraction'
|
30
|
+
end
|
31
|
+
|
32
|
+
def url
|
33
|
+
uri = Addressable::URI.parse(gateway)
|
34
|
+
uri.query_values = {
|
35
|
+
# TODO: Change appid to the BMP one
|
36
|
+
'appid' => @api_key,
|
37
|
+
'output' => 'xml',
|
38
|
+
'context' => @context
|
39
|
+
}
|
40
|
+
uri
|
41
|
+
end
|
42
|
+
|
43
|
+
def remote_xml
|
44
|
+
open(url).read
|
45
|
+
end
|
46
|
+
|
47
|
+
def data
|
48
|
+
@data ||= Nokogiri::XML.parse(remote_xml)
|
49
|
+
end
|
50
|
+
end
|
51
|
+
end
|
@@ -0,0 +1,49 @@
|
|
1
|
+
require 'net/http'
|
2
|
+
require 'nokogiri'
|
3
|
+
|
4
|
+
class TermExtraction
|
5
|
+
class Zemanta < TermExtraction
|
6
|
+
def terms
|
7
|
+
terms = []
|
8
|
+
|
9
|
+
data.search('//name').each do |n|
|
10
|
+
terms << n.text
|
11
|
+
end
|
12
|
+
|
13
|
+
terms
|
14
|
+
end
|
15
|
+
|
16
|
+
class << self
|
17
|
+
def canonical_name
|
18
|
+
'zemanta'
|
19
|
+
end
|
20
|
+
end
|
21
|
+
|
22
|
+
private
|
23
|
+
def gateway
|
24
|
+
'http://api.zemanta.com/services/rest/0.0/'
|
25
|
+
end
|
26
|
+
|
27
|
+
def url
|
28
|
+
URI.parse(gateway)
|
29
|
+
end
|
30
|
+
|
31
|
+
def post_params
|
32
|
+
{
|
33
|
+
'method' =>'zemanta.suggest',
|
34
|
+
'api_key' => @api_key,
|
35
|
+
'return_images' => 0,
|
36
|
+
'text' => @context,
|
37
|
+
'format' => 'xml'
|
38
|
+
}
|
39
|
+
end
|
40
|
+
|
41
|
+
def remote_xml
|
42
|
+
Net::HTTP.post_form(url, post_params).body
|
43
|
+
end
|
44
|
+
|
45
|
+
def data
|
46
|
+
@data ||= Nokogiri::XML.parse(remote_xml)
|
47
|
+
end
|
48
|
+
end
|
49
|
+
end
|
@@ -0,0 +1,3 @@
|
|
1
|
+
<?xml version="1.0"?>
|
2
|
+
<ResultSet xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns="urn:yahoo:cate" xsi:schemaLocation="urn:yahoo:cate http://api.search.yahoo.com/ContentAnalysisService/V1/TermExtractionResponse.xsd"><Result>gears of war</Result><Result>gears</Result></ResultSet>
|
3
|
+
<!-- ws04.search.scd.yahoo.com uncompressed/chunked Thu Feb 26 21:04:16 PST 2009 -->
|
@@ -0,0 +1,104 @@
|
|
1
|
+
<?xml version="1.0"?>
|
2
|
+
<rsp>
|
3
|
+
<status>ok</status>
|
4
|
+
<articles>
|
5
|
+
<article>
|
6
|
+
<url>http://www.crunchgear.com/2009/02/24/nvidia-based-imacs-coming-soon/</url>
|
7
|
+
<confidence>0.033153</confidence>
|
8
|
+
<published_datetime>2009-02-24T18:00:45Z</published_datetime>
|
9
|
+
<zemified>0</zemified>
|
10
|
+
<title>NVIDIA-based iMacs coming soon?</title>
|
11
|
+
</article><article>
|
12
|
+
<url>http://www.tuaw.com/2009/02/24/rumor-new-imacs-around-the-bend/</url>
|
13
|
+
<confidence>0.028595</confidence>
|
14
|
+
<published_datetime>2009-02-24T22:00:00Z</published_datetime>
|
15
|
+
<zemified>0</zemified>
|
16
|
+
<title>Rumor: New iMacs around the bend</title>
|
17
|
+
</article><article>
|
18
|
+
<url>http://www.engadget.com/2009/02/24/a-few-new-rumors-point-to-two-new-nvidia-packing-imacs/</url>
|
19
|
+
<confidence>0.027595</confidence>
|
20
|
+
<published_datetime>2009-02-24T13:03:00Z</published_datetime>
|
21
|
+
<zemified>0</zemified>
|
22
|
+
<title>A few new rumors point to two new NVIDIA-packing iMacs</title>
|
23
|
+
</article><article>
|
24
|
+
<url>http://www.techmeme.com/090126/p31</url>
|
25
|
+
<confidence>0.019056</confidence>
|
26
|
+
<published_datetime>2009-01-26T17:00:22Z</published_datetime>
|
27
|
+
<zemified>0</zemified>
|
28
|
+
<title>Chip complex delaying Apple's new iMac line, says analyst (Zach Spear/AppleInsider)</title>
|
29
|
+
</article><article>
|
30
|
+
<url>http://i.gizmodo.com/5143476/apple-warns-resellers-of-reduced-imac-availability-new-models-on-the-way</url>
|
31
|
+
<confidence>0.018611</confidence>
|
32
|
+
<published_datetime>2009-01-31T16:30:00Z</published_datetime>
|
33
|
+
<zemified>0</zemified>
|
34
|
+
<title>Apple Warns Resellers of Reduced iMac Availability: New Models On the Way? [Apple]</title>
|
35
|
+
</article><article>
|
36
|
+
<url>http://cultofmac.com/analyst-new-imacs-delayed-for-chips-snow-leopard/7549</url>
|
37
|
+
<confidence>0.01685</confidence>
|
38
|
+
<published_datetime>2009-01-26T16:43:44Z</published_datetime>
|
39
|
+
<zemified>0</zemified>
|
40
|
+
<title>Analyst: New iMacs Delayed For Chips, Snow Leopard</title>
|
41
|
+
</article><article>
|
42
|
+
<url>http://www.ubergizmo.com/15/archives/2008/12/new_imac_allinone_pc_confirmed.html</url>
|
43
|
+
<confidence>0.016311</confidence>
|
44
|
+
<published_datetime>2008-12-24T01:06:56Z</published_datetime>
|
45
|
+
<zemified>0</zemified>
|
46
|
+
<title>New iMac All-In-One PC Confirmed</title>
|
47
|
+
</article><article>
|
48
|
+
<url>http://www.labnol.org/gadgets/exchange-windows-pc-for-apple-mac/5775/</url>
|
49
|
+
<confidence>0.016238</confidence>
|
50
|
+
<published_datetime>2008-12-03T08:17:35Z</published_datetime>
|
51
|
+
<zemified>0</zemified>
|
52
|
+
<title>Exchange Your Windows PC or TV for an Apple iMac</title>
|
53
|
+
</article><article>
|
54
|
+
<url>http://cultofmac.com/class-action-lawsuit-over-imac-display-problems/6562</url>
|
55
|
+
<confidence>0.015608</confidence>
|
56
|
+
<published_datetime>2009-01-02T16:17:06Z</published_datetime>
|
57
|
+
<zemified>0</zemified>
|
58
|
+
<title>Class-Action Lawsuit Over iMac Display Problems</title>
|
59
|
+
</article><article>
|
60
|
+
<url>http://www.crunchgear.com/2008/12/29/new-imacs-to-include-hotter-components-new-cooling/</url>
|
61
|
+
<confidence>0.015093</confidence>
|
62
|
+
<published_datetime>2008-12-29T23:50:04Z</published_datetime>
|
63
|
+
<zemified>0</zemified>
|
64
|
+
<title>New iMacs to include hotter components, new cooling?</title>
|
65
|
+
</article>
|
66
|
+
</articles><markup>
|
67
|
+
<text>apple imac</text>
|
68
|
+
</markup><signature><div class="zemanta-pixie"><a class="zemanta-pixie-a" href="http://reblog.zemanta.com/zemified/3695b55a-d68e-4328-bf86-c3f01be8b2da/" title="Zemified by Zemanta"><img class="zemanta-pixie-img" src="http://img.zemanta.com/reblog_e.png?x-id=3695b55a-d68e-4328-bf86-c3f01be8b2da" alt="Reblog this post [with Zemanta]" /></a></div></signature>
|
69
|
+
<keywords>
|
70
|
+
<keyword>
|
71
|
+
<confidence>0.404748</confidence>
|
72
|
+
<scheme>general</scheme>
|
73
|
+
<name>Apple</name>
|
74
|
+
</keyword><keyword>
|
75
|
+
<confidence>0.277753</confidence>
|
76
|
+
<scheme>general</scheme>
|
77
|
+
<name>IMac</name>
|
78
|
+
</keyword><keyword>
|
79
|
+
<confidence>0.094392</confidence>
|
80
|
+
<scheme>general</scheme>
|
81
|
+
<name>Rumor</name>
|
82
|
+
</keyword><keyword>
|
83
|
+
<confidence>0.076399</confidence>
|
84
|
+
<scheme>general</scheme>
|
85
|
+
<name>Hardware</name>
|
86
|
+
</keyword><keyword>
|
87
|
+
<confidence>0.072186</confidence>
|
88
|
+
<scheme>general</scheme>
|
89
|
+
<name>Nvidia</name>
|
90
|
+
</keyword><keyword>
|
91
|
+
<confidence>0.069183</confidence>
|
92
|
+
<scheme>general</scheme>
|
93
|
+
<name>Macintosh</name>
|
94
|
+
</keyword><keyword>
|
95
|
+
<confidence>0.050487</confidence>
|
96
|
+
<scheme>general</scheme>
|
97
|
+
<name>Desktops</name>
|
98
|
+
</keyword><keyword>
|
99
|
+
<confidence>0.039205</confidence>
|
100
|
+
<scheme>general</scheme>
|
101
|
+
<name>AllInOne</name>
|
102
|
+
</keyword>
|
103
|
+
</keywords><rid>3695b55a-d68e-4328-bf86-c3f01be8b2da</rid>
|
104
|
+
</rsp>
|
@@ -0,0 +1,15 @@
|
|
1
|
+
require File.dirname(__FILE__) + '/test_helper'
|
2
|
+
|
3
|
+
class TermExtractionTest < Test::Unit::TestCase
|
4
|
+
should 'return correct terms from Yahoo!' do
|
5
|
+
yahoo = TermExtraction::Yahoo.new
|
6
|
+
yahoo.stubs(:remote_xml).returns(read_xml_fixture('yahoo'))
|
7
|
+
assert_equal yahoo.terms, ['gears of war', 'gears']
|
8
|
+
end
|
9
|
+
|
10
|
+
should 'return correct terms from Zemanta' do
|
11
|
+
zemanta = TermExtraction::Zemanta.new
|
12
|
+
zemanta.stubs(:remote_xml).returns(read_xml_fixture('zemanta'))
|
13
|
+
assert_equal zemanta.terms, ['Apple', 'IMac', 'Rumor', 'Hardware', 'Nvidia', 'Macintosh', 'Desktops', 'AllInOne']
|
14
|
+
end
|
15
|
+
end
|
data/test/test_helper.rb
ADDED
@@ -0,0 +1,13 @@
|
|
1
|
+
require 'rubygems'
|
2
|
+
require 'test/unit'
|
3
|
+
require 'shoulda'
|
4
|
+
require 'mocha'
|
5
|
+
|
6
|
+
$LOAD_PATH.unshift(File.dirname(__FILE__))
|
7
|
+
require 'term_extraction'
|
8
|
+
|
9
|
+
class Test::Unit::TestCase
|
10
|
+
def read_xml_fixture(name)
|
11
|
+
File.read("test/fixtures/#{name}.xml")
|
12
|
+
end
|
13
|
+
end
|
metadata
ADDED
@@ -0,0 +1,72 @@
|
|
1
|
+
--- !ruby/object:Gem::Specification
|
2
|
+
name: alexrabarts-term_extraction
|
3
|
+
version: !ruby/object:Gem::Version
|
4
|
+
version: 0.1.0
|
5
|
+
platform: ruby
|
6
|
+
authors:
|
7
|
+
- alex
|
8
|
+
autorequire:
|
9
|
+
bindir: bin
|
10
|
+
cert_chain: []
|
11
|
+
|
12
|
+
date: 2009-02-27 00:00:00 -08:00
|
13
|
+
default_executable:
|
14
|
+
dependencies:
|
15
|
+
- !ruby/object:Gem::Dependency
|
16
|
+
name: nokogiri
|
17
|
+
type: :runtime
|
18
|
+
version_requirement:
|
19
|
+
version_requirements: !ruby/object:Gem::Requirement
|
20
|
+
requirements:
|
21
|
+
- - ">="
|
22
|
+
- !ruby/object:Gem::Version
|
23
|
+
version: 1.0.7
|
24
|
+
version:
|
25
|
+
description: Term extraction library
|
26
|
+
email: alexrabarts@gmail.com
|
27
|
+
executables: []
|
28
|
+
|
29
|
+
extensions: []
|
30
|
+
|
31
|
+
extra_rdoc_files: []
|
32
|
+
|
33
|
+
files:
|
34
|
+
- VERSION.yml
|
35
|
+
- lib/term_extraction
|
36
|
+
- lib/term_extraction/yahoo.rb
|
37
|
+
- lib/term_extraction/zemanta.rb
|
38
|
+
- lib/term_extraction.rb
|
39
|
+
- test/fixtures
|
40
|
+
- test/fixtures/yahoo.xml
|
41
|
+
- test/fixtures/zemanta.xml
|
42
|
+
- test/term_extraction_test.rb
|
43
|
+
- test/test_helper.rb
|
44
|
+
has_rdoc: true
|
45
|
+
homepage: http://github.com/alexrabarts/term_extraction
|
46
|
+
post_install_message:
|
47
|
+
rdoc_options:
|
48
|
+
- --inline-source
|
49
|
+
- --charset=UTF-8
|
50
|
+
require_paths:
|
51
|
+
- lib
|
52
|
+
required_ruby_version: !ruby/object:Gem::Requirement
|
53
|
+
requirements:
|
54
|
+
- - ">="
|
55
|
+
- !ruby/object:Gem::Version
|
56
|
+
version: "0"
|
57
|
+
version:
|
58
|
+
required_rubygems_version: !ruby/object:Gem::Requirement
|
59
|
+
requirements:
|
60
|
+
- - ">="
|
61
|
+
- !ruby/object:Gem::Version
|
62
|
+
version: "0"
|
63
|
+
version:
|
64
|
+
requirements: []
|
65
|
+
|
66
|
+
rubyforge_project:
|
67
|
+
rubygems_version: 1.2.0
|
68
|
+
signing_key:
|
69
|
+
specification_version: 2
|
70
|
+
summary: Provides access to term extraction APIs such as Yahoo! Term Extraction API and Zemanta.
|
71
|
+
test_files: []
|
72
|
+
|