Class: ContentFocus::HTML

Inherits:
Object
  • Object
show all
Defined in:
lib/content_focus/html.rb

Overview

Static content fragments are things like: title, about, author, content of an article, etc.

Instance Method Summary collapse

Constructor Details

#initialize(html) ⇒ HTML

Returns a new instance of HTML.



9
10
11
 
# File 'lib/content_focus/html.rb', line 9

def initialize(html)
  @doc = Hpricot(html)
end

Instance Method Details

#static_fragment(options = {}) ⇒ Object

Based on the title, find a common chunk of HTML that is the most relevant This is to extract atomic/permanent content



20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
 
# File 'lib/content_focus/html.rb', line 20

def static_fragment(options = {})
  fragments = self.static_fragments(options)
    
  return nil if fragments == nil || fragments.empty?

  if fragments.size == 1
    return fragments.first
  end

  # Find common ancestors
  fragments_by_parents = {}
  fragments.each do |fragment|
    next unless fragment[:parent]
    fragments_by_parents[fragment[:parent]] ||= []
    fragments_by_parents[fragment[:parent]] << fragment
  end

  # Find the top parent
  top_fragments = []
  top_parent_fragments_count = 0
  fragments_by_parents.each do |parent,fr|
    if fr.size > top_parent_fragments_count
      top_parent_fragments_count = fr.size
      top_fragments = fr
    end
  end

  # Failed?
  if !top_fragments || top_fragments.empty?
    return fragments.first
  end

  # Create a combined fragment with combined score
  element = top_fragments.first[:element]
  combined_fragment = {:score => 0, :element => element.parent, :inner_text => element.parent.inner_text, :parent => element.parent ? element.parent.object_id : nil}
  top_fragments.each { |f| combined_fragment[:score] = combined_fragment[:score] + f[:score] }

  # De-value the body tag
  if combined_fragment[:element].name == 'body'
    combined_fragment[:score] = top_fragments.size
  end

  # Add combined fragment to pool and re-order by score.
  fragments << combined_fragment
  fragments.sort! { |b,a| a[:score] <=> b[:score] }

  #puts fragments.collect { |f| ["#{f[:element].parent ? f[:element].parent.name : nil}:#{f[:parent]}", f[:element].name + '(' + f[:score].to_s + '): ', f[:element].attributes] }.inspect

  fragments.first
end

#static_fragments(options = {}) ⇒ Object

Get all relevant div/span/td/body/p blocks from the HTML page - based on the <title> This is to extract atomic/permanent content



73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
 
# File 'lib/content_focus/html.rb', line 73

def static_fragments(options = {})
  title_elements = (@doc/"title")
  return html if !title_elements || title_elements.empty?
  title_inner_text = title_elements.first.inner_text
  keywords = Linguistics::Tagger.keywords_for_caption(title_inner_text)
  blocks = []

  # First, find the smallest blocks, but bigger than the title
  (@doc/"div|span|td|body|p|dd|ul").each do |element|
  
    next if element_with_negative_identifier(element)
  
    inner_text = ''
    element.children.each do |child|
      inner_text << child.to_s if child.is_a?(Hpricot::Text)
    end
    inner_text.downcase!
    next if inner_text.size <= title_inner_text.size
  
    # Check the occurance of keyword in block, skip if none
    num_matches = 0
    keywords.each { |k| num_matches+=1 if inner_text.split(/\s+/).include?(k) }
    next if num_matches == 0
  
    # Calculate a score based on keyword matches times positive naming of id/class
    score = num_matches
    identifier = nil
    if (identifier = element_with_positive_identifier(element))
      score = score * 2;
    end
  
    blocks << {:score => score, :element => element, :inner_text => inner_text, :parent => element.parent ? element.parent.object_id : nil, :identifier => identifier}
  end

  big_block_identifiers = {}

  # Finding big blocks with both matches and positive identifiers
  (@doc/"div|span|table|td|body|p|dd|ul").each do |element|
  
    next if element_with_negative_identifier(element)
  
    # Need to log identifier statistics
    identifier = nil
    if (identifier = element_with_positive_identifier(element))
      big_block_identifiers[identifier] ||= 0
      big_block_identifiers[identifier] += 1
    else
      next
    end
  
    inner_text = element.inner_text
    inner_text.downcase!
    next if inner_text.size <= title_inner_text.size
  
    # Check the occurance of keyword in block, skip if none
    num_matches = 0
    keywords.each { |k| num_matches+=1 if inner_text.split(/\s+/).include?(k) }
  
    #puts "#{element.name}(#{element.inner_text.size}/#{title_inner_text.size}, score:#{num_matches} * #{element_with_positive_identifier(element)}): " + element.attributes['class'].to_s
  
    next if num_matches == 0
  
    # Calculate a score based on keyword matches times positive naming of id/class
    score = num_matches
    if identifier
      score = score * 3;
    end
  
    blocks << {:score => score, :element => element, :inner_text => inner_text, :parent => element.parent ? element.parent : nil, :identifier => identifier}
  end

  # De-value the identifiers that are repeated
  blocks.each do |block|
    if block[:identifier] && big_block_identifiers[block[:identifier]].to_i > 1
      block[:score] = block[:score] / 3;
    end
  end

  # Order those blocks by top matches
  blocks.sort! { |b,a| a[:score] <=> b[:score] }
  blocks.reject! { |b| b[:score] == 0 }
  blocks
end

#static_text ⇒ Object 



13
14
15
16
 
# File 'lib/content_focus/html.rb', line 13

def static_text
  fragment = self.static_fragment
  fragment ? fragment[:element].inner_text.strip! : nil
end