Class: OverdriveMetadata

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

Defined Under Namespace

Classes: EAudioBook, EBook, ERecord

Constant Summary collapse

VERSION = 
'1.0.2.4'
OD_ORG = 
'OverDrive, Inc.'
OD_URL = 
'http://www.overdrive.com'
ACCESS = 
'Mode of access: World Wide Web.'
URL_MSG = 
'Click to download this resource.'
DISCLAIM = 
'Record generated from Overdrive metadata spreadsheet.'
READ_ERR = 
'Error reading spreadsheet! Ensure file is closed, verfiy location and save as XLS'
HEADERS = 
{
  :oclc      => 19,
  :date      => 12,
  :time      => 21,
  :isbn      => 1,
  :author    => 4,
  :title     => 2,
  :place     => 11,
  :publisher => 3,
  :requires  => 10,
  :format    => 9,
  :filesize  => 8,
  :reader    => 14,
  :title_src => 13,
  :summary   => 15,
  :subjects  => 5,
  :download  => 7,
  :excerpt   => 16,
  :cover     => 17,
  :thumb     => 18,
}

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(metadata_file, agency, header = true, ebook_regex = nil) ⇒ OverdriveMetadata

Returns a new instance of OverdriveMetadata.



38
39
40
41
42
43
44
45
46
47
48
49
50
 
# File 'lib/overdrive_metadata.rb', line 38

def initialize(metadata_file, agency, header = true, ebook_regex = nil)
  begin
    @metadata = Spreadsheet.open(metadata_file).worksheet 0
  rescue Exception => ex
      raise READ_ERR
  end
  @agency      = agency
  @ebook_regex = ebook_regex.nil? ? '(ebook|epub|kindle|pdf)' : ebook_regex
  @records     = []
  @count       = 0
  @header      = header
  @content_rec = {}
end

Instance Attribute Details

#count ⇒ Object (readonly)

Returns the value of attribute count.



7
8
9
 
# File 'lib/overdrive_metadata.rb', line 7

def count
  @count
end

#records ⇒ Object (readonly)

Returns the value of attribute records.



7
8
9
 
# File 'lib/overdrive_metadata.rb', line 7

def records
  @records
end

Instance Method Details

#clean_string(input_str) ⇒ Object 



186
187
188
 
# File 'lib/overdrive_metadata.rb', line 186

def clean_string(input_str)
  return input_str.gsub(/&lt;.*&gt;/, '').gsub(/&amp;/, '&').gsub(/&quot;/, '"').gsub(/&apos;/, "'").gsub(/&#160;/, '').gsub(/&#235;/, 'e').gsub(/<\/?[^>]*>/, '').gsub(/\s{2}+/, ' ').strip rescue ''
end

#create_record(data) ⇒ Object 



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
 
# File 'lib/overdrive_metadata.rb', line 74

def create_record(data)
  @count += 1
  field = package_data(data)

  if @content_rec.has_key? field[:download]
    record  = @content_rec[field[:download]]
    format  = MARC::DataField.new('500', ' ', ' ', ['a', "#{field[:format]} (file size: #{field[:filesize]} MB)."])
    record.fields.insert(record.fields.index { |f| f.tag == '500' }, format)

    unless field[:excerpt].empty?
      excerpt = MARC::DataField.new('856', '4', '0', ['u', field[:excerpt]], ['y', "Excerpt (#{field[:format]})."])
      record.fields.insert(record.fields.index { |f| f.tag == '856' }, excerpt)
    end
    return nil
  end

  r = field[:format].match(/#{@ebook_regex}/i) ? EBook.new : EAudioBook.new
  r.make_control_field('001', field[:oclc])
  r.make_006
  r.make_007
  r.make_fixed_field(field[:year], field[:month], field[:day])
  r.make_data_field('020', ' ', ' ', {'a' => field[:isbn] + ' ' + r.isbn}) unless field[:isbn].empty?
  r.make_data_field('037', ' ', ' ', {'b' => OD_ORG, 'n' => OD_URL})
  r.make_data_field('040', ' ', ' ', {'a' => @agency, 'c' => @agency})
  r.make_data_field('100', '1', ' ', {'a' => normalize_author(field[:author])})
  r.make_title(field[:title], field[:author])
  r.make_publication(field[:place], field[:publisher], field[:year])
  r.make_physical(field[:hours], field[:minutes]) # ebooks will ignore
  r.make_data_field('306', ' ', ' ', {'a' => field[:hours] + field[:minutes] + field[:seconds]})
  r.make_data_field('538', ' ', ' ', {'a' => ACCESS})
  r.make_data_field('538', ' ', ' ', {'a' => 'Requires ' + field[:requires] + '.'})  
  r.make_data_field('500', ' ', ' ', {'a' => "#{field[:format]} (file size: #{field[:filesize]} MB)."})
  r.make_data_field('511', '0', ' ', {'a' => "Read by #{field[:reader]}."}) unless field[:reader].empty? 
  r.make_data_field('520', ' ', ' ', {'a' => field[:summary]}) unless field[:summary].match(/^#+$/)
  r.make_data_field('500', ' ', ' ', {'a' => "Title from: #{field[:title_src]}."})
  
  if r.is_a? EAudioBook
    r.make_data_field('500', ' ', ' ', {'a' => 'Unabridged.'})
    r.make_data_field('500', ' ', ' ', {'a' => "Duration: #{field[:hours]} hr., #{field[:minutes]} min."})
  end
  
  field[:subjects].each { |s| r.make_data_field('655', ' ', '7', {'a' => clean_string(s).strip + '.', '2' => 'local'}) }
  r.make_data_field('655', ' ', '7', {'a' => r.subject, '2' => 'local'})
  r.make_data_field('700', '1', ' ', {'a' => normalize_author(field[:reader])})
  r.make_data_field('856', '4', '0', {'u' => field[:excerpt], 'y' => "Excerpt (#{field[:format]})."})
  r.make_data_field('856', '4', '0', {'u' => field[:download], 'y' => URL_MSG})
  
  if @agency == 'JTH'
    if r.is_a? EAudioBook
      r.make_data_field('856', '4', '2', {'u' => field[:cover], 'y' => "<img class=\"scl_mwthumb\" src=\"#{field[:thumb]}\" alt=\"Artwork for this title - #{field[:title].gsub(/[^A-Za-z ]/, '')}\" />"})
    end
    r.make_data_field('907', ' ', ' ', {'a' => 'ER'})
  end
  
  r.make_data_field('991', ' ', ' ', {'a' => DISCLAIM})

  @content_rec[field[:download]] = r.record
  return r.record
end

#kb_to_mb(size) ⇒ Object 



190
191
192
 
# File 'lib/overdrive_metadata.rb', line 190

def kb_to_mb(size)
  return (size.to_f / 1024 + 1).to_i.to_s
end

#make_id(id_string) ⇒ Object 



172
173
174
 
# File 'lib/overdrive_metadata.rb', line 172

def make_id(id_string)
  return id_string[-9..-1].gsub(/\W/, '')
end

#map ⇒ Object 



52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
 
# File 'lib/overdrive_metadata.rb', line 52

def map
  @metadata.each do |row|
    if @header
      @header = false
      next
    end

    begin
      @records << create_record(row)
    rescue Exception => ex
      puts "Row #{@count.to_s}\t#{ex.message}"
      puts "#{ex.backtrace[0]}"
      puts
      next
    end

  end

  @records.compact!
  @records
end

#normalize_author(author) ⇒ Object 



176
177
178
179
180
181
182
183
184
 
# File 'lib/overdrive_metadata.rb', line 176

def normalize_author(author)
  return author if author.empty?
  author   = author.split(',')[0]
  names    = author.split ' '
  surname  = names.last + ', '
  fullname = surname + names[0 .. names.length - 2].join(' ')
  fullname += '.' unless fullname[-1] == '.'
  return fullname
end

#package_data(data) ⇒ Object 



134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
 
# File 'lib/overdrive_metadata.rb', line 134

def package_data(data)
  values = {}
  values[:isbn]             = data[HEADERS[:isbn]]
  values[:date]             = data[HEADERS[:date]]
  values[:place]            = data[HEADERS[:place]]
  values[:publisher]        = data[HEADERS[:publisher]]
  values[:month]            = ''
  values[:day]              = ''
  if values[:date].match(/\d{1,2}\/\d{1,2}\/\d{2,4}/)
    month, day, year        = values[:date].split '/'
    values[:month]          = month
    values[:day]            = day
    values[:year]           = year.size == 4 ? year : "20#{year}"
  end
  values[:year]             = values[:date].match(/\d{4}/).to_s unless year # fall-back
  values[:time]             = data[HEADERS[:time]]
  hr, mn, sc                = values[:time].split ':' rescue ''
  values[:hours]            = hr ? hr : ''
  values[:minutes]          = mn ? mn : ''
  values[:seconds]          = sc ? sc : ''
  values[:author]           = clean_string data[HEADERS[:author]]
  values[:title]            = clean_string data[HEADERS[:title]]
  values[:title_src]        = data[HEADERS[:title_src]]
  values[:reader]           = clean_string data[HEADERS[:reader]]
  values[:requires]         = data[HEADERS[:requires]]
  values[:format]           = data[HEADERS[:format]]
  values[:filesize]         = kb_to_mb(data[HEADERS[:filesize]])
  values[:summary]          = clean_string data[HEADERS[:summary]]
  values[:subjects]         = data[HEADERS[:subjects]].split(',') rescue []
  values[:download]         = data[HEADERS[:download]]
  values[:excerpt]          = data[HEADERS[:excerpt]]
  values[:thumb]            = data[HEADERS[:thumb]]
  values[:cover]            = data[HEADERS[:cover]]
  values[:oclc]             = data[HEADERS[:oclc]].to_s.empty? ? 'ovr' + make_id(values[:download]) : 'ocn' + data[HEADERS[:oclc]]
  values.each { |k, v| values[k] = '' if v.nil? }
  return values
end