Class: Biblionet::Extractors::BibliographicalBookExtractor

Inherits:
Base
  • Object
show all
Defined in:
lib/bookshark/extractors/bibliographical_book_extractor.rb

Instance Attribute Summary collapse

Attributes inherited from Base

#biblionet_id, #filepath, #page, #url

Instance Method Summary collapse

Methods inherited from Base

#decode_text, decode_text, #load_page, #load_page_from_file, #load_page_from_url, #present?, #save_page

Methods included from FileManager

#list_directories, #list_files, #save_to

Constructor Details

#initialize(uri = nil) ⇒ BibliographicalBookExtractor

Returns a new instance of BibliographicalBookExtractor.



13
14
15
16
# File 'lib/bookshark/extractors/bibliographical_book_extractor.rb', line 13

def initialize(uri=nil)
  super(uri)
  extract_bibliographical_book unless uri.nil? or @page.nil?
end

Instance Attribute Details

#bibliographical_bookObject (readonly)

Returns the value of attribute bibliographical_book.



11
12
13
# File 'lib/bookshark/extractors/bibliographical_book_extractor.rb', line 11

def bibliographical_book
  @bibliographical_book
end

Instance Method Details

#extract_bibliographical_book(biblionet_id = @biblionet_id, book_page = @page) ⇒ Object



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
# File 'lib/bookshark/extractors/bibliographical_book_extractor.rb', line 23

def extract_bibliographical_book(biblionet_id=@biblionet_id, book_page=@page)
  # log = Logger.new(File.new(File.dirname(__dir__).to_s + "/logs/book_parsing.log",'a+'))
  log = Logger.new(STDOUT)

  page = BibliographicalBookDataExtractor.new(book_page)

  # End extraction if BookDataExtractor couldnt create a nodeset
  return nil if page.nodeset.nil?

  bibliographical_book_hash = Hash.new

  extracted_details = page.details

  bibliographical_book_hash[:original_language] = extracted_details[:original_language]
  bibliographical_book_hash[:original_title]    = extracted_details[:original_title]
  bibliographical_book_hash[:last_update]       = extracted_details[:last_update]
  bibliographical_book_hash[:cover_type]        = extracted_details[:cover_type]
  bibliographical_book_hash[:availability]      = extracted_details[:availability]
  bibliographical_book_hash[:price]             = extracted_details[:price]

  bibliographical_book_hash[:series]            = extracted_details[:series]
  bibliographical_book_hash[:physical_size]     = extracted_details[:physical_size]

  bibliographical_book_hash[:format]            = extracted_details[:format]

  bibliographical_book_hash[:publisher]         = extracted_details[:publisher]
  bibliographical_book_hash[:publication]       = extracted_details[:publication]

  return @bibliographical_book = bibliographical_book_hash
end

#load_and_extract_book(uri = nil) ⇒ Object



18
19
20
21
# File 'lib/bookshark/extractors/bibliographical_book_extractor.rb', line 18

def load_and_extract_book(uri=nil)
  load_page(uri)
  extract_bibliographical_book unless uri.nil? or @page.nil?
end