bookshark 1.0.0.beta.5 → 1.0.0.pre.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: a592c6e055f501c19f4a7dea23bad79446b6b28c
-  data.tar.gz: 797820d896b398e6294f9804a4a42f151682a027
+  metadata.gz: 7e2e0077beb0dadb4ef9df9fcce7b7d55d54e8a6
+  data.tar.gz: 712b2769781e8b1ea55ecb25a501d886a939e307
 SHA512:
-  metadata.gz: 1f8f7b1b0b0f7312549964153afe8198aab00b0acabdb9693c5e9f7c8242e90a19357fb3507397df6cd5b3278cb1ceb22d2b45dfb4646851e698b5cf2bfeb96d
-  data.tar.gz: 73e4f60098595c1e5b20101bf66c5d24d3cc2e2659c5776a7973635a9789149e5a6d2a2a4a69ff7af126314f25f9b2e31824fd131879f0b3b5939eff55462604
+  metadata.gz: a2348fa9b757abe44be69c0d21e5c61bc41e87eaae80d11bafff115201f40793d2724d6fe5f518c3c3f26308696febd4c772f8a7e989f085949ea6ab584c7264
+  data.tar.gz: 221cdadd0bcaf5df156cb668465b6a60e7471a5a650ddcce703a602a646dac19dee008a4de088ed037695289d7fa983792d1a5397ea40cfe843a39bf50155bb0

data/README.md

@@ -13,7 +13,7 @@ The representation of bibliographic metadata in JSON is inspired by [BibJSON](ht
 Add this line to your application's Gemfile:
 
 ```ruby
-gem 'bookshark', "~> 1.0.0.beta"
+gem 'bookshark', "~> 1.0.0.pre"
 ```
 
 And then execute:
@@ -145,29 +145,47 @@ The expected result of a book extraction is something like this:
         ]
       },
       "publisher": {
-        "name": "Εκδοτικός Οίκος Α. Α. Λιβάνη",
+        "text": "Εκδοτικός Οίκος Α. Α. Λιβάνη",
         "b_id": "271"
       },
-      "publication_year": "2006",
-      "pages": "326",
       "isbn": "960-14-1157-7",
       "isbn_13": "978-960-14-1157-6",
-      "status": "Κυκλοφορεί",
-      "price": "16,31",
       "award": [
+
       ],
       "description": "Τι είναι πιο επικίνδυνο, ένα όπλο ή μια πισίνα; Τι κοινό έχουν οι δάσκαλοι με τους παλαιστές του σούμο;...",
       "category": [
         {
           "ddc": "330",
-          "text": "Οικονομία",
+          "name": "Οικονομία",
           "b_id": "142"
         }
       ],
-      "b_id": "103788"
+      "b_id": "103788",
+      "publication": {
+        "year": "2006",
+        "version": "1",
+        "place": "Αθήνα"
+      },
+      "format": "Βιβλίο",
+      "original_language": "αγγλικά",
+      "original_title": "Freakonomics",
+      "price": "16,31",
+      "availability": "Κυκλοφορεί",
+      "last_update": "27/1/2006",
+      "series": {
+        "name": "Οικονομία",
+        "volume": null
+      },
+      "physical_description": {
+        "pages": "326",
+        "size": "21x14",
+        "cover_type": "Μαλακό εξώφυλλο"
+      }
     }
   ]
 }
+
 ```
 Here is a [Book Sample](https://gist.github.com/dklisiaris/a6f3d6f37806186f3c79) extracted with eager option enabled.
 

data/bookshark.gemspec

@@ -21,7 +21,7 @@ Gem::Specification.new do |spec|
   spec.require_paths = ["lib"]
 
   spec.add_dependency "nokogiri", "~> 1.6", ">= 1.6.6"
-  spec.add_dependency "sanitize", "~> 3.1"
+  spec.add_dependency "sanitize", "~> 4.0"
   spec.add_dependency "json", "~> 1.8"
   spec.add_dependency "htmlentities", "~> 4.3"
 

data/lib/bookshark.rb

@@ -4,6 +4,7 @@ require 'bookshark/storage/file_manager'
 require 'bookshark/extractors/author_extractor'
 require 'bookshark/extractors/category_extractor'
 require 'bookshark/extractors/book_extractor'
+require 'bookshark/extractors/bibliographical_book_extractor'
 require 'bookshark/extractors/publisher_extractor'
 require 'bookshark/extractors/search'
 
@@ -91,6 +92,23 @@ module Bookshark
       return response            
     end
 
+    
+    # def bibliographical_book(options = {})
+    #   bibliographical_book_extractor = Biblionet::Extractors::BibliographicalBookExtractor.new
+ 
+    #   uri = "http://www.biblionet.gr/main.asp?page=results&Titlesid=#{options[:id]}"
+    #   options[:format]  ||= @format
+ 
+    #   book = bibliographical_book_extractor.load_and_extract_book(uri)
+ 
+    #   response = {}      
+    #   response[:book] = !book.nil? ? [book] : []
+    #   response = change_format(response, options[:format])
+    #   response = bibliographical_book_extractor.decode_text(response)      
+    # end    
+    
+    # puts Bookshark::Extractor.new(format: 'pretty_json').bibliographical_book(id: 103788)
+
     def category(options = {})
       uri = process_options(options, __method__)
       options[:format] ||= @format      
@@ -231,7 +249,7 @@ module Bookshark
           local_path    = "html_book_pages/#{((id-1)/1000)}/book_#{id}.html"
         when 'category'
           url_method    = 'index' 
-          local_path    = "html_ddc_pages/#{((id-1)/1000)}/ddc_#{id}.html"       
+          local_path    = "html_ddc_pages/#{((id-1)/1000)}/ddc_#{id}.html"               
         else
           puts "Called from unknown method. Probably its rspec."
         end      

data/lib/bookshark/extractors/base.rb

@@ -141,6 +141,10 @@ module Biblionet
       # +encoded_text+:: the text which contains encoded entities
       #
       def decode_text(encoded_text)
+        self.class.decode_text(encoded_text)
+      end
+
+      def self.decode_text(encoded_text)
         # encoded_text = File.read(encoded_file_path)
         coder = HTMLEntities.new
         coder.decode(encoded_text)

data/lib/bookshark/extractors/bibliographical_book_extractor.rb

@@ -0,0 +1,172 @@
+#!/bin/env ruby
+# encoding: utf-8
+
+require_relative 'base'
+
+
+module Biblionet
+  module Extractors   
+    
+    class BibliographicalBookExtractor < Base 
+      attr_reader :bibliographical_book
+
+      def initialize(uri=nil)
+        super(uri)        
+        extract_bibliographical_book unless uri.nil? or @page.nil?        
+      end
+
+      def load_and_extract_book(uri=nil)
+        load_page(uri)
+        extract_bibliographical_book unless uri.nil? or @page.nil?
+      end  
+
+      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
+
+    end
+
+    class BibliographicalBookDataExtractor
+      attr_reader :nodeset
+
+      def initialize(document)
+        # No need to operate on whole page. Just on part containing the book.
+        content_re = /<!-- CONTENT START -->.*<!-- CONTENT END -->/m
+        if (content_re.match(document)).nil?
+          puts document
+        end
+        content = content_re.match(document)[0] unless (content_re.match(document)).nil?
+        
+        # If content is nil, there is something wrong with the html, so return nil
+        if content.nil?
+          @nodeset = nil
+        else
+          @nodeset = Nokogiri::HTML(content) 
+        end        
+      end    
+
+      def size
+        size_regex = /\d+x\d+/
+      end
+
+      def series
+        series_regex        = /(?<=\()\p{Word}+( \p{Word}+)* · \d+(?=\))/
+        series_name_regex   = /\p{Word}+( \p{Word}+)*(?= ·)/
+        series_volume_regex = /(?<=· )\d+/
+      end
+
+      def details
+        details_hash        = {}
+        isbn_regex          = /(?<= )\d+-\d+-\d+-\d+(?= |,)/
+        isbn_13_regex       = /\d+-\d+-\d+-\d+-\d+/
+        last_update_regex   = /\d{1,2}\/\d{1,2}\/\d{2,4}/
+        cover_type_regex    = /(?<=\()\p{Word}+( \p{Word}+)?(?=\))/
+        availability_regex  = /(?<=\[).+(?=\])/
+        price_regex         = /(?<=€ )\d+,\d*/
+
+        @nodeset.xpath("//span[@class='small'][1]").inner_html.split('<br>').each do |detail|
+          detail = BibliographicalBookExtractor.decode_text(detail)
+
+          if detail.start_with? "Γλώσσα πρωτοτύπου:"
+            original_language = detail.gsub(/Γλώσσα πρωτοτύπου:/, "").strip
+            details_hash[:original_language] = original_language
+          elsif detail.start_with? "Τίτλος πρωτοτύπου:"
+            original_title = detail.gsub(/Τίτλος πρωτοτύπου:/, "").strip
+            details_hash[:original_title] = original_title
+          end
+          
+          details_hash[:isbn]         = detail[isbn_regex] if detail =~ isbn_regex                 
+
+          details_hash[:isbn_13]      = detail[isbn_13_regex] if detail =~ isbn_13_regex
+
+          details_hash[:last_update]  = detail[last_update_regex] if detail =~ last_update_regex 
+
+          details_hash[:cover_type]   = detail[cover_type_regex] if detail =~ cover_type_regex  
+
+          details_hash[:availability] = detail[availability_regex] if detail =~ availability_regex
+
+          details_hash[:price]        = detail[price_regex] if detail =~ price_regex 
+          
+        end
+
+        pre_details_text = @nodeset.xpath("//span[@class='small'][1]/preceding::text()").text
+        pre_details_text = BibliographicalBookExtractor.decode_text(pre_details_text)
+
+        series_regex        = /(?<=\()\p{Word}+( \S)?( \p{Word}+( \S)?)* · \d+(?=\))/
+        series_regex_no_vol = /(?<=\()\p{Word}+( \S)?( \p{Word}+( \S)?)*(?=\))/
+        series_name_regex   = /\p{Word}+( \S)?( \p{Word}+( \S)?)*(?= ·)/
+        series_volume_regex = /(?<=· )\d+/
+        physical_size_regex = /\d+x\d+/        
+
+        series_hash = {}
+        if pre_details_text =~ series_regex
+          series = pre_details_text[series_regex]
+          series_hash[:name]    = series[series_name_regex] if series =~ series_name_regex
+          series_hash[:volume]  = series[series_volume_regex] if series =~ series_volume_regex          
+        elsif pre_details_text =~ series_regex_no_vol
+          series = pre_details_text[series_regex_no_vol]
+          series_hash[:name]    = series
+          series_hash[:volume]  = nil
+        end
+
+        details_hash[:series] = series_hash
+        
+        details_hash[:physical_size] = (pre_details_text =~ physical_size_regex) ? pre_details_text[physical_size_regex] : nil
+
+        format_regex = /(?<=\[).+(?=\])/
+
+        after_title_text = @nodeset.xpath("//a[@class='booklink' and @href[contains(.,'/book/') ]][1]").first.next_sibling.text.strip
+        format = after_title_text[format_regex] if after_title_text =~ format_regex
+
+        details_hash[:format] = format.nil? ? 'Βιβλίο' : format
+
+        publisher_node = @nodeset.xpath("//a[@class='booklink' and @href[contains(.,'/com/') ]][1]").first
+        publisher_hash = {}
+        publisher_hash[:text] = publisher_node.text
+        publisher_hash[:b_id] = (publisher_node[:href].split("/"))[2]
+
+        pre_publisher_text = BibliographicalBookExtractor.decode_text(publisher_node.previous_sibling.text)
+        after_publisher_text = BibliographicalBookExtractor.decode_text(publisher_node.next_sibling.text)
+
+        publication_hash = {}
+        publication_hash[:year] = after_publisher_text[/(?<=, )\d+(?=\.)/]
+        publication_hash[:version] = pre_details_text[/(?<=- )\d+(?=η)/]
+        publication_hash[:place] = pre_details_text[/(?<=- )\p{Word}+( \S)?( \p{Word}+( \S)?)*(?= :)/]
+
+        details_hash[:publisher] = publisher_hash
+        details_hash[:publication] = publication_hash
+
+        details_hash
+      end
+
+    end
+
+  end
+end

data/lib/bookshark/extractors/book_extractor.rb

@@ -2,6 +2,7 @@
 # encoding: utf-8
 
 require_relative 'base'
+require_relative 'bibliographical_book_extractor'
 require 'sanitize'
 
 module Biblionet
@@ -169,12 +170,12 @@ module Biblionet
 
         details_hash = proccess_details(details)
 
-        book_hash[:publication_year] = details_hash[:publication_year]
-        book_hash[:pages]            = details_hash[:pages]
+        # book_hash[:publication_year] = details_hash[:publication_year]
+        # book_hash[:pages]            = details_hash[:pages]
         book_hash[:isbn]             = details_hash[:isbn]
         book_hash[:isbn_13]          = details_hash[:isbn_13].nil? ? nil : details_hash[:isbn_13]
-        book_hash[:status]           = details_hash[:status]
-        book_hash[:price]            = details_hash[:price]
+        # book_hash[:status]           = details_hash[:status]
+        # book_hash[:price]            = details_hash[:price]
         book_hash[:award]            = page.awards
 
 
@@ -192,7 +193,34 @@ module Biblionet
 
 
         book_hash[:category]   = ddcs
-        book_hash[:b_id] = biblionet_id 
+        book_hash[:b_id] = biblionet_id
+
+        uri = "http://www.biblionet.gr/main.asp?page=results&Titlesid=#{biblionet_id}"
+
+        bibliographical_book_extractor = Biblionet::Extractors::BibliographicalBookExtractor.new
+        bibliographical_details = bibliographical_book_extractor.load_and_extract_book(uri)      
+
+        book_hash[:publisher]         = bibliographical_details[:publisher]
+        book_hash[:publication]       = bibliographical_details[:publication]   
+
+        book_hash[:format]            = bibliographical_details[:format]     
+
+        book_hash[:original_language] = bibliographical_details[:original_language]
+        book_hash[:original_title]    = bibliographical_details[:original_title]
+
+        book_hash[:price]             = bibliographical_details[:price]      
+        book_hash[:availability]      = bibliographical_details[:availability]
+        book_hash[:last_update]       = bibliographical_details[:last_update]
+        
+        book_hash[:series]            = bibliographical_details[:series]        
+
+        physical_description_hash = {}
+        physical_description_hash[:pages]      = details_hash[:pages]
+        physical_description_hash[:size]       = bibliographical_details[:physical_size]
+        physical_description_hash[:cover_type] = bibliographical_details[:cover_type]
+
+        book_hash[:physical_description] = physical_description_hash
+        
 
         return @book = book_hash  
       end      

data/lib/bookshark/extractors/publisher_extractor.rb

@@ -99,6 +99,7 @@ module Biblionet
         headquarters_hash             = Hash[headquarters_hash.map {|k, v| [mappings[k], v] }]
         headquarters_hash[:telephone] = [headquarters_hash[:telephone]] unless headquarters_hash[:telephone].kind_of?(Array) or headquarters_hash[:telephone].nil?
         headquarters_hash[:website]   = headquarters_hash[:website].split(',').map(&:strip) if (headquarters_hash[:website] and headquarters_hash[:website].include? ',')
+        headquarters_hash[:address]   = [headquarters_hash[:address]] unless headquarters_hash[:address].kind_of?(Array) 
 
         return headquarters_hash                
       end

data/lib/bookshark/version.rb

@@ -1,3 +1,3 @@
 module Bookshark
-  VERSION = "1.0.0.beta.5"
+  VERSION = "1.0.0.pre.1"
 end

data/spec/bookshark_spec.rb

@@ -206,4 +206,8 @@ describe Bookshark::Extractor do
     end
   end
 
+  describe 'Biblionet::Extractors::BibliographicalBook' do 
+    
+  end
+
 end

data/spec/test_data/book_103788.json

@@ -23,15 +23,11 @@
         ]
       },
       "publisher": {
-        "name": "Εκδοτικός Οίκος Α. Α. Λιβάνη",
+        "text": "Εκδοτικός Οίκος Α. Α. Λιβάνη",
         "b_id": "271"
       },
-      "publication_year": "2006",
-      "pages": "326",
       "isbn": "960-14-1157-7",
       "isbn_13": "978-960-14-1157-6",
-      "status": "Κυκλοφορεί",
-      "price": "16,31",
       "award": [
 
       ],
@@ -43,7 +39,27 @@
           "b_id": "142"
         }
       ],
-      "b_id": "103788"
+      "b_id": "103788",
+      "publication": {
+        "year": "2006",
+        "version": "1",
+        "place": "Αθήνα"
+      },
+      "format": "Βιβλίο",
+      "original_language": "αγγλικά",
+      "original_title": "Freakonomics",
+      "price": "16,31",
+      "availability": "Κυκλοφορεί",
+      "last_update": "27/1/2006",
+      "series": {
+        "name": "Οικονομία",
+        "volume": null
+      },
+      "physical_description": {
+        "pages": "326",
+        "size": "21x14",
+        "cover_type": "Μαλακό εξώφυλλο"
+      }
     }
   ]
 }

data/spec/test_data/eager_book_184923.json

@@ -71,12 +71,8 @@
           "b_id": "112"
         }
       ],
-      "publication_year": "2012",
-      "pages": "345",
       "isbn": "978-960-524-394-4",
       "isbn_13": null,
-      "status": "Κυκλοφορεί",
-      "price": "16,00",
       "award": [
 
       ],
@@ -209,7 +205,27 @@
           }
         }
       ],
-      "b_id": "184923"
+      "b_id": "184923",
+      "publication": {
+        "year": "2012",
+        "version": "1",
+        "place": "Ηράκλειο Κρήτης"
+      },
+      "format": "Βιβλίο",
+      "original_language": null,
+      "original_title": null,
+      "price": "16,00",
+      "availability": "Κυκλοφορεί",
+      "last_update": "12/12/2012",
+      "series": {
+        "name": "Εκλαΐκευση της Επιστήμης",
+        "volume": null
+      },
+      "physical_description": {
+        "pages": "345",
+        "size": "24x17",
+        "cover_type": "Μαλακό εξώφυλλο"
+      }
     }
   ]
-}
+}

data/spec/test_data/search_01.json

@@ -13,15 +13,11 @@
       "contributors": {
       },
       "publisher": {
-        "name": "Λογοσοφία",
+        "text": "Λογοσοφία",
         "b_id": "7628"
       },
-      "publication_year": "2007",
-      "pages": "181",
       "isbn": "978-960-89288-3-1",
       "isbn_13": null,
-      "status": "Κυκλοφορεί",
-      "price": "13,52",
       "award": [
 
       ],
@@ -33,7 +29,25 @@
           "b_id": "3"
         }
       ],
-      "b_id": "119000"
+      "b_id": "119000",
+      "publication": {
+        "year": "2007",
+        "version": "1",
+        "place": "Αθήνα"
+      },
+      "format": "Βιβλίο",
+      "original_language": null,
+      "original_title": null,
+      "price": "13,52",
+      "availability": "Κυκλοφορεί",
+      "last_update": "9/5/2007",
+      "series": {
+      },
+      "physical_description": {
+        "pages": "181",
+        "size": "21x14",
+        "cover_type": "Μαλακό εξώφυλλο"
+      }
     },
     {
       "title": "Σημεία και τέρατα της οικονομίας",
@@ -58,15 +72,11 @@
         ]
       },
       "publisher": {
-        "name": "Εκδοτικός Οίκος Α. Α. Λιβάνη",
+        "text": "Εκδοτικός Οίκος Α. Α. Λιβάνη",
         "b_id": "271"
       },
-      "publication_year": "2006",
-      "pages": "326",
       "isbn": "960-14-1157-7",
       "isbn_13": "978-960-14-1157-6",
-      "status": "Κυκλοφορεί",
-      "price": "16,31",
       "award": [
 
       ],
@@ -78,7 +88,27 @@
           "b_id": "142"
         }
       ],
-      "b_id": "103788"
+      "b_id": "103788",
+      "publication": {
+        "year": "2006",
+        "version": "1",
+        "place": "Αθήνα"
+      },
+      "format": "Βιβλίο",
+      "original_language": "αγγλικά",
+      "original_title": "Freakonomics",
+      "price": "16,31",
+      "availability": "Κυκλοφορεί",
+      "last_update": "27/1/2006",
+      "series": {
+        "name": "Οικονομία",
+        "volume": null
+      },
+      "physical_description": {
+        "pages": "326",
+        "size": "21x14",
+        "cover_type": "Μαλακό εξώφυλλο"
+      }
     },
     {
       "title": "Και άλλα σημεία και τέρατα από την ιστορία",
@@ -103,15 +133,11 @@
         ]
       },
       "publisher": {
-        "name": "Modern Times",
+        "text": "Modern Times",
         "b_id": "191"
       },
-      "publication_year": "2004",
-      "pages": "62",
       "isbn": "960-397-927-9",
       "isbn_13": "978-960-397-927-2",
-      "status": "Κυκλοφορεί",
-      "price": "15,14",
       "award": [
 
       ],
@@ -123,7 +149,25 @@
           "b_id": "2456"
         }
       ],
-      "b_id": "87815"
+      "b_id": "87815",
+      "publication": {
+        "year": "2004",
+        "version": null,
+        "place": "Αθήνα"
+      },
+      "format": "Βιβλίο",
+      "original_language": "αγγλικά",
+      "original_title": "Even more horrible history",
+      "price": "15,14",
+      "availability": "Κυκλοφορεί",
+      "last_update": null,
+      "series": {
+      },
+      "physical_description": {
+        "pages": "62",
+        "size": "28x22",
+        "cover_type": "Σκληρό εξώφυλλο"
+      }
     },
     {
       "title": "Σημεία και τέρατα από την ιστορία",
@@ -148,15 +192,11 @@
         ]
       },
       "publisher": {
-        "name": "Modern Times",
+        "text": "Modern Times",
         "b_id": "191"
       },
-      "publication_year": "2004",
-      "pages": "78",
       "isbn": "960-397-926-0",
       "isbn_13": "978-960-397-926-5",
-      "status": "Κυκλοφορεί",
-      "price": "15,14",
       "award": [
 
       ],
@@ -168,7 +208,25 @@
           "b_id": "2456"
         }
       ],
-      "b_id": "87812"
+      "b_id": "87812",
+      "publication": {
+        "year": "2004",
+        "version": null,
+        "place": "Αθήνα"
+      },
+      "format": "Βιβλίο",
+      "original_language": "αγγλικά",
+      "original_title": "Horrible history",
+      "price": "15,14",
+      "availability": "Κυκλοφορεί",
+      "last_update": null,
+      "series": {
+      },
+      "physical_description": {
+        "pages": "78",
+        "size": "28x22",
+        "cover_type": "Σκληρό εξώφυλλο"
+      }
     },
     {
       "title": "Σημεία και τέρατα",
@@ -183,15 +241,11 @@
       "contributors": {
       },
       "publisher": {
-        "name": "Κέδρος",
+        "text": "Κέδρος",
         "b_id": "21"
       },
-      "publication_year": "1994",
-      "pages": "126",
       "isbn": "960-04-0941-2",
       "isbn_13": "978-960-04-0941-3",
-      "status": "Κυκλοφορεί",
-      "price": "9,17",
       "award": [
 
       ],
@@ -203,7 +257,25 @@
           "b_id": "9"
         }
       ],
-      "b_id": "15839"
+      "b_id": "15839",
+      "publication": {
+        "year": "1994",
+        "version": "2",
+        "place": "Αθήνα"
+      },
+      "format": "Βιβλίο",
+      "original_language": null,
+      "original_title": null,
+      "price": "9,17",
+      "availability": "Κυκλοφορεί",
+      "last_update": "28/7/2010",
+      "series": {
+      },
+      "physical_description": {
+        "pages": "126",
+        "size": "21x14",
+        "cover_type": "Μαλακό εξώφυλλο"
+      }
     },
     {
       "title": "Επίσημη αγραμματοσύνη",
@@ -218,15 +290,11 @@
       "contributors": {
       },
       "publisher": {
-        "name": "Περιοδικό Πνευματική Ζωή",
+        "text": "Περιοδικό Πνευματική Ζωή",
         "b_id": "6770"
       },
-      "publication_year": null,
-      "pages": "48",
       "isbn": null,
       "isbn_13": null,
-      "status": "Κυκλοφορεί",
-      "price": null,
       "award": [
 
       ],
@@ -238,7 +306,27 @@
           "b_id": "1459"
         }
       ],
-      "b_id": "77381"
+      "b_id": "77381",
+      "publication": {
+        "year": null,
+        "version": null,
+        "place": "Αθήνα"
+      },
+      "format": "Βιβλίο",
+      "original_language": null,
+      "original_title": null,
+      "price": null,
+      "availability": "Κυκλοφορεί",
+      "last_update": null,
+      "series": {
+        "name": "Νεοελληνικά Αφιερώματα",
+        "volume": "21"
+      },
+      "physical_description": {
+        "pages": "48",
+        "size": "24x17",
+        "cover_type": "Μαλακό εξώφυλλο"
+      }
     },
     {
       "title": "Ο όλεθρος της πυραμίδος Καμπαλά",
@@ -259,15 +347,11 @@
         ]
       },
       "publisher": {
-        "name": "Μπίμπης Στερέωμα",
+        "text": "Μπίμπης Στερέωμα",
         "b_id": "244"
       },
-      "publication_year": null,
-      "pages": "269",
       "isbn": null,
       "isbn_13": null,
-      "status": "Κυκλοφορεί",
-      "price": "6,85",
       "award": [
 
       ],
@@ -279,7 +363,25 @@
           "b_id": "179"
         }
       ],
-      "b_id": "46856"
+      "b_id": "46856",
+      "publication": {
+        "year": null,
+        "version": null,
+        "place": "Θεσσαλονίκη"
+      },
+      "format": "Βιβλίο",
+      "original_language": null,
+      "original_title": null,
+      "price": "6,85",
+      "availability": "Κυκλοφορεί",
+      "last_update": null,
+      "series": {
+      },
+      "physical_description": {
+        "pages": "269",
+        "size": "21x14",
+        "cover_type": "Μαλακό εξώφυλλο"
+      }
     },
     {
       "title": "Η επανάσταση κατά της νέας τάξης",
@@ -294,15 +396,11 @@
       "contributors": {
       },
       "publisher": {
-        "name": "Μπίμπης Στερέωμα",
+        "text": "Μπίμπης Στερέωμα",
         "b_id": "244"
       },
-      "publication_year": null,
-      "pages": "71",
       "isbn": null,
       "isbn_13": null,
-      "status": "Κυκλοφορεί",
-      "price": "3,73",
       "award": [
 
       ],
@@ -314,7 +412,25 @@
           "b_id": "2336"
         }
       ],
-      "b_id": "46763"
+      "b_id": "46763",
+      "publication": {
+        "year": null,
+        "version": null,
+        "place": "Θεσσαλονίκη"
+      },
+      "format": "Βιβλίο",
+      "original_language": null,
+      "original_title": null,
+      "price": "3,73",
+      "availability": "Κυκλοφορεί",
+      "last_update": null,
+      "series": {
+      },
+      "physical_description": {
+        "pages": "71",
+        "size": "21x14",
+        "cover_type": "Μαλακό εξώφυλλο"
+      }
     },
     {
       "title": "Επιχείρησις: Μαργαριτάρια",
@@ -329,15 +445,11 @@
       "contributors": {
       },
       "publisher": {
-        "name": "Δωδώνη Εκδοτική ΕΠΕ",
+        "text": "Δωδώνη Εκδοτική ΕΠΕ",
         "b_id": "1"
       },
-      "publication_year": null,
-      "pages": "188",
       "isbn": "960-248-541-8",
       "isbn_13": "978-960-248-541-5",
-      "status": "Κυκλοφορεί",
-      "price": "10,60",
       "award": [
 
       ],
@@ -349,7 +461,25 @@
           "b_id": "1309"
         }
       ],
-      "b_id": "33301"
+      "b_id": "33301",
+      "publication": {
+        "year": null,
+        "version": null,
+        "place": "Αθήνα"
+      },
+      "format": "Βιβλίο",
+      "original_language": null,
+      "original_title": null,
+      "price": "10,60",
+      "availability": "Κυκλοφορεί",
+      "last_update": null,
+      "series": {
+      },
+      "physical_description": {
+        "pages": "188",
+        "size": "20x13",
+        "cover_type": "Μαλακό εξώφυλλο"
+      }
     }
   ]
-}
+}

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: bookshark
 version: !ruby/object:Gem::Version
-  version: 1.0.0.beta.5
+  version: 1.0.0.pre.1
 platform: ruby
 authors:
 - Dimitris Klisiaris
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-04-18 00:00:00.000000000 Z
+date: 2015-05-03 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: nokogiri
@@ -36,14 +36,14 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '3.1'
+        version: '4.0'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '3.1'
+        version: '4.0'
 - !ruby/object:Gem::Dependency
   name: json
   requirement: !ruby/object:Gem::Requirement
@@ -137,6 +137,7 @@ files:
 - lib/bookshark/crawlers/publisher_crawler.rb
 - lib/bookshark/extractors/author_extractor.rb
 - lib/bookshark/extractors/base.rb
+- lib/bookshark/extractors/bibliographical_book_extractor.rb
 - lib/bookshark/extractors/book_extractor.rb
 - lib/bookshark/extractors/category_extractor.rb
 - lib/bookshark/extractors/publisher_extractor.rb