Class: Nlg::Extractors::Base

Inherits:
Object
  • Object
show all
Defined in:
lib/bookshark/extractors/nlg/base.rb

Direct Known Subclasses

BookExtractor

Instance Attribute Summary collapse

Class Method Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(id = nil) ⇒ Base

Returns a new instance of Base.



18
19
20
 
# File 'lib/bookshark/extractors/nlg/base.rb', line 18

def initialize(id=nil)
  load_page(id)
end

Instance Attribute Details

#nlg_id ⇒ Object (readonly)

Returns the value of attribute nlg_id.



16
17
18
 
# File 'lib/bookshark/extractors/nlg/base.rb', line 16

def nlg_id
  @nlg_id
end

#page ⇒ Object (readonly)

Returns the value of attribute page.



16
17
18
 
# File 'lib/bookshark/extractors/nlg/base.rb', line 16

def page
  @page
end

#url ⇒ Object (readonly)

Returns the value of attribute url.



16
17
18
 
# File 'lib/bookshark/extractors/nlg/base.rb', line 16

def url
  @url
end

Class Method Details

.decode_text(encoded_text) ⇒ Object 



70
71
72
73
74
 
# File 'lib/bookshark/extractors/nlg/base.rb', line 70

def self.decode_text(encoded_text)
  # encoded_text = File.read(encoded_file_path)
  coder = HTMLEntities.new
  coder.decode(encoded_text)
end

Instance Method Details

#decode_text(encoded_text) ⇒ Object

Decodes text with escaped html entities and returns the decoded text.

Params:

encoded_text

the text which contains encoded entities



66
67
68
 
# File 'lib/bookshark/extractors/nlg/base.rb', line 66

def decode_text(encoded_text)
  self.class.decode_text(encoded_text)
end

#load_page(id = nil) ⇒ Object 



22
23
24
 
# File 'lib/bookshark/extractors/nlg/base.rb', line 22

def load_page(id=nil)
  load_page_by_id(id) unless id.nil?
end

#load_page_by_id(id) ⇒ Object 



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
 
# File 'lib/bookshark/extractors/nlg/base.rb', line 26

def load_page_by_id(id)
  begin
    @nlg_id = id unless id.nil? # id is expected to be the last number.
    @url = "http://nbib.nlg.gr/Record/#{@nlg_id}/Export?style=MARCXML"

    pp "Downloading page: #{@url}"

    Net::HTTP.start("nbib.nlg.gr") do |http|
      response = http.get("/Record/#{@nlg_id}/Export?style=MARCXML")
      pp response.content_type
      pp response.code
      raise EmptyPageError.new(@url) unless response.content_type == "text/xml" && response.code == "200"

      @page = response.body
    end

  rescue Errno::ENOENT => e
    pp "Page: #{@url} NOT FOUND."
    pp e
  rescue EmptyPageError => e
    pp "Page: #{@url} is EMPTY."
    pp e
    @page = nil
  rescue OpenURI::HTTPError => e
    pp e
    pp e.io.status
  rescue StandardError => e
    pp "Generic error #{e.class}. Will wait for 2 minutes and then try again."
    pp e
    sleep(120)
    retry
  end
end

#present?(value) ⇒ Boolean

Returns:

  • (Boolean) 


76
77
78
 
# File 'lib/bookshark/extractors/nlg/base.rb', line 76

def present?(value)
  return (not value.nil? and not value.empty?) ? true : false
end