busk-ruby-readability 1.0.0 → 1.0.1
Sign up to get free protection for your applications and to get access to all the features.
- data/lib/readability.rb +13 -11
- metadata +17 -5
data/lib/readability.rb
CHANGED
@@ -6,7 +6,7 @@ module Readability
|
|
6
6
|
TEXT_LENGTH_THRESHOLD = 25
|
7
7
|
RETRY_LENGTH = 250
|
8
8
|
|
9
|
-
attr_accessor :options, :html
|
9
|
+
attr_accessor :options, :html, :best_candidate
|
10
10
|
|
11
11
|
def initialize(input, options = {})
|
12
12
|
@input = input
|
@@ -83,17 +83,18 @@ module Readability
|
|
83
83
|
end
|
84
84
|
|
85
85
|
def select_best_candidate(candidates)
|
86
|
-
|
86
|
+
@best_candidate ||= begin
|
87
|
+
sorted_candidates = candidates.values.sort { |a, b| b[:content_score] <=> a[:content_score] }
|
87
88
|
|
88
|
-
|
89
|
-
|
90
|
-
|
91
|
-
|
92
|
-
|
93
|
-
best_candidate = sorted_candidates.first || { :elem => @html.css("body").first, :content_score => 0 }
|
94
|
-
debug("Best candidate #{best_candidate[:elem].name}##{best_candidate[:elem][:id]}.#{best_candidate[:elem][:class]} with score #{best_candidate[:content_score]}")
|
89
|
+
debug("Top 5 candidates:")
|
90
|
+
sorted_candidates[0...5].each do |candidate|
|
91
|
+
debug("Candidate #{candidate[:elem].andand.name}##{candidate[:elem][:id]}.#{candidate[:elem][:class]} with score #{candidate[:content_score]}")
|
92
|
+
end
|
95
93
|
|
96
|
-
|
94
|
+
best_candidate = sorted_candidates.first || { :elem => @html.css("body").first, :content_score => 0 }
|
95
|
+
#debug("Best candidate #{best_candidate[:elem].andand.name} with score #{best_candidate[:content_score]}")
|
96
|
+
best_candidate
|
97
|
+
end
|
97
98
|
end
|
98
99
|
|
99
100
|
def get_link_density(elem)
|
@@ -281,7 +282,8 @@ module Readability
|
|
281
282
|
el.attributes.each { |a, x| el.delete(a) unless @options[:attributes] && @options[:attributes].include?(a.to_s) }
|
282
283
|
# Otherwise, replace the element with its contents
|
283
284
|
else
|
284
|
-
|
285
|
+
# keep getting whiny nils with nokogiri
|
286
|
+
el.swap(el.text) rescue nil
|
285
287
|
end
|
286
288
|
|
287
289
|
end
|
metadata
CHANGED
@@ -1,7 +1,13 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: busk-ruby-readability
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
|
4
|
+
hash: 21
|
5
|
+
prerelease: false
|
6
|
+
segments:
|
7
|
+
- 1
|
8
|
+
- 0
|
9
|
+
- 1
|
10
|
+
version: 1.0.1
|
5
11
|
platform: ruby
|
6
12
|
authors: []
|
7
13
|
|
@@ -9,7 +15,7 @@ autorequire:
|
|
9
15
|
bindir: bin
|
10
16
|
cert_chain: []
|
11
17
|
|
12
|
-
date: 2010-09-
|
18
|
+
date: 2010-09-20 00:00:00 -03:00
|
13
19
|
default_executable:
|
14
20
|
dependencies: []
|
15
21
|
|
@@ -33,21 +39,27 @@ rdoc_options: []
|
|
33
39
|
require_paths:
|
34
40
|
- lib
|
35
41
|
required_ruby_version: !ruby/object:Gem::Requirement
|
42
|
+
none: false
|
36
43
|
requirements:
|
37
44
|
- - ">="
|
38
45
|
- !ruby/object:Gem::Version
|
46
|
+
hash: 3
|
47
|
+
segments:
|
48
|
+
- 0
|
39
49
|
version: "0"
|
40
|
-
version:
|
41
50
|
required_rubygems_version: !ruby/object:Gem::Requirement
|
51
|
+
none: false
|
42
52
|
requirements:
|
43
53
|
- - ">="
|
44
54
|
- !ruby/object:Gem::Version
|
55
|
+
hash: 3
|
56
|
+
segments:
|
57
|
+
- 0
|
45
58
|
version: "0"
|
46
|
-
version:
|
47
59
|
requirements: []
|
48
60
|
|
49
61
|
rubyforge_project:
|
50
|
-
rubygems_version: 1.3.
|
62
|
+
rubygems_version: 1.3.7
|
51
63
|
signing_key:
|
52
64
|
specification_version: 3
|
53
65
|
summary: A rewrite of original ruby-readability
|