Module: Oddb2xml

Defined in:
lib/oddb2xml.rb,
lib/oddb2xml/cli.rb,
lib/oddb2xml/calc.rb,
lib/oddb2xml/util.rb,
lib/oddb2xml/builder.rb,
lib/oddb2xml/compare.rb,
lib/oddb2xml/options.rb,
lib/oddb2xml/version.rb,
lib/oddb2xml/extractor.rb,
lib/oddb2xml/compressor.rb,
lib/oddb2xml/downloader.rb,
lib/oddb2xml/semantic_check.rb,
lib/oddb2xml/chapter_70_hack.rb

Defined Under Namespace

Modules: DownloadMethod, Options, TxtExtractorMethods Classes: BagXmlDownloader, BagXmlExtractor, Builder, Calc, Chapter70xtractor, Cli, CompareV5, Compressor, Downloader, EphaDownloader, EphaExtractor, Extractor, GalenicForm, GalenicGroup, LppvDownloader, LppvExtractor, MedregbmDownloader, MedregbmExtractor, MigelDownloader, MigelExtractor, RefdataDownloader, RefdataExtractor, SemanticCheck, SemanticCheckXML, StammXML, SwissmedicDownloader, SwissmedicExtractor, SwissmedicInfoDownloader, SwissmedicInfoExtractor, ZurroseDownloader, ZurroseExtractor

Constant Summary collapse

FAKE_GTIN_START = 
'999999'
WorkDir = 
Dir.pwd
Downloads = 
"#{Dir.pwd}/downloads"
COLUMNS_JULY_2015 =

please keep this constant in sync between (GEM) swissmedic-diff/lib/swissmedic-diff.rb and (GEM) oddb2xml/lib/oddb2xml/extractor.rb

{
  :iksnr => /Zulassungs-Nummer/i,                  # column-nr: 0
  :seqnr => /Dosisstärke-nummer/i,
  :name_base => /Präparatebezeichnung/i,
  :company => /Zulassungsinhaberin/i,
  :production_science => /Heilmittelcode/i,
  :index_therapeuticus => /IT-Nummer/i,            # column-nr: 5
  :atc_class => /ATC-Code/i,
  :registration_date => /Erstzulassungs-datum./i,
  :sequence_date => /Zul.datum Dosisstärke/i,
  :expiry_date => /Gültigkeitsdauer der Zulassung/i,
  :ikscd => /Packungscode/i,                 # column-nr: 10
  :size => /Packungsgrösse/i,
  :unit => /Einheit/i,
  :ikscat => /Abgabekategorie Packung/i,
  :ikscat_seq => /Abgabekategorie Dosisstärke/i,
  :ikscat_preparation => /Abgabekategorie Präparat/i, # column-nr: 15
  :substances => /Wirkstoff/i,
  :composition => /Zusammensetzung/i,
  :indication_registration => /Anwendungsgebiet Präparat/i,
  :indication_sequence => /Anwendungsgebiet Dosisstärke/i,
  :gen_production => /Gentechnisch hergestellte Wirkstoffe/i, # column-nr 20
  :insulin_category => /Kategorie bei Insulinen/i,
  # swissmedi corrected in february 2018 the typo  betäubunsmittel to  betäubungsmittel-
    :drug_index       => /Verz. bei betäubun.*smittel-haltigen Präparaten/i,
}
XML_OPTIONS = 
{
'xmlns:xsd'         => 'http://www.w3.org/2001/XMLSchema',
'xmlns:xsi'         => 'http://www.w3.org/2001/XMLSchema-instance',
'xmlns'             => 'http://wiki.oddb.org/wiki.php?pagename=Swissmedic.Datendeklaration',
'CREATION_DATETIME' => Time.new.strftime('%FT%T%z'),
'PROD_DATE'         => Time.new.strftime('%FT%T%z'),
'VALID_DATE'        => Time.new.strftime('%FT%T%z'),
'GENERATED_BY'      => "oddb2xml #{VERSION}"
}
VERSION = 
"2.5.9"
@@prodno_to_ean13 =

Needed for ensuring consistency for the Artikelstamm

{}
@@no8_to_ean13 = 
{}
@@ean13_to_prodno = 
{}
@@ean13_to_no8 = 
{}

Class Method Summary collapse

Class Method Details

.add_epha_changes_for_ATC(iksnr, atc_code, force_run: false) ⇒ Object 



61
62
63
64
65
66
67
68
69
70
71
72
73
 
# File 'lib/oddb2xml/util.rb', line 61

def Oddb2xml.add_epha_changes_for_ATC(iksnr, atc_code, force_run: false)
  @atc_csv_content  =  {} if force_run
  if @atc_csv_content.size == 0
    open(@atc_csv_origin).readlines.each{
      |line|
        items = line.split(',')
        @atc_csv_content[[items[0], items[1]]] = items[2]
    }

  end
  new_value = @atc_csv_content[[iksnr.to_s, atc_code]]
  new_value ? new_value : atc_code
end

.add_hash(string) ⇒ Object 



156
157
158
159
160
161
162
163
164
165
 
# File 'lib/oddb2xml/util.rb', line 156

def Oddb2xml.add_hash(string)
  doc = Nokogiri::XML.parse(string)
  nr = 0
  doc.root.elements.each do |node|
    nr += 1
    next if node.name.eql?('RESULT')
    node['SHA256'] = Digest::SHA256.hexdigest node.text
  end
  doc.to_xml
end

.calc_checksum(str) ⇒ Object 



10
11
12
13
14
15
16
17
18
19
 
# File 'lib/oddb2xml/util.rb', line 10

def Oddb2xml.calc_checksum(str)
  str = str.strip
  sum = 0
  val =   str.split(//u)
  12.times do |idx|
    fct = ((idx%2)*2)+1
    sum += fct*val[idx].to_i
  end
  ((10-(sum%10))%10).to_s
end

.check_column_indices(sheet) ⇒ Object

please keep this constant in sync between (GEM) swissmedic-diff/lib/swissmedic-diff.rb and (GEM) oddb2xml/lib/oddb2xml/extractor.rb



112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
 
# File 'lib/oddb2xml/util.rb', line 112

def Oddb2xml.check_column_indices(sheet)
  row = sheet[4] # Headers are found at row 4

  error_2015 = nil
  COLUMNS_JULY_2015.each{
    |key, value|
    header_name = row[COLUMNS_JULY_2015.keys.index(key)].value.to_s
    unless value.match(header_name)
      puts "#{__LINE__}: #{key} ->  #{COLUMNS_JULY_2015.keys.index(key)} #{value}\nbut was  #{header_name}" if $VERBOSE
      error_2015 = "Packungen.xlslx_has_unexpected_column_#{COLUMNS_JULY_2015.keys.index(key)}_#{key}_#{value.to_s}_but_was_#{header_name}"
      break
    end
  }
  raise "#{error_2015}" if error_2015
end

.convert_to_8859_1(line) ⇒ Object 



51
52
53
54
55
56
57
58
59
 
# File 'lib/oddb2xml/util.rb', line 51

def Oddb2xml.convert_to_8859_1(line)
  begin
    # We want to ignore lines which are not really UTF-8 encoded
      ausgabe = Oddb2xml.patch_some_utf8(line).encode('ISO-8859-1')
      ausgabe.encode('ISO-8859-1')
  rescue => error
    puts "#{error}: in #{line}"
  end
end

.download_finished(file, remove_file = true) ⇒ Object 



100
101
102
103
104
105
106
107
108
109
 
# File 'lib/oddb2xml/util.rb', line 100

def Oddb2xml.download_finished(file, remove_file = true)
  src  = "#{WorkDir}/#{File.basename(file)}"
  dest = "#{Downloads}/#{File.basename(file)}"
  FileUtils.makedirs(Downloads)
  #return unless File.exists?(file)
  return unless file and File.exists?(file)
  return if File.expand_path(file).eql?(dest)
  FileUtils.cp(src, dest, :verbose => false)
  Oddb2xml.log("download_finished saved as #{dest} #{File.size(dest)} bytes.")
end

.gen_prodno(iksnr, seqnr) ⇒ Object 



7
8
9
 
# File 'lib/oddb2xml/util.rb', line 7

def Oddb2xml.gen_prodno(iksnr, seqnr)
   sprintf('%05d',iksnr) + sprintf('%02d', seqnr)
end

.getEan13forNo8(no8) ⇒ Object 



226
227
228
 
# File 'lib/oddb2xml/util.rb', line 226

def Oddb2xml.getEan13forNo8(no8)
  @@no8_to_ean13[no8] || []
end

.getEan13forProdno(prodno) ⇒ Object 



223
224
225
 
# File 'lib/oddb2xml/util.rb', line 223

def Oddb2xml.getEan13forProdno(prodno)
  @@prodno_to_ean13[prodno] || []
end

.getNo8ForEan13(ean13) ⇒ Object 



232
233
234
 
# File 'lib/oddb2xml/util.rb', line 232

def Oddb2xml.getNo8ForEan13(ean13)
  @@ean13_to_no8[ean13]
end

.getProdnoForEan13(ean13) ⇒ Object 



229
230
231
 
# File 'lib/oddb2xml/util.rb', line 229

def Oddb2xml.getProdnoForEan13(ean13)
  @@ean13_to_prodno[ean13]
end

.html_decode(string) ⇒ Object 



29
30
31
32
33
34
35
36
 
# File 'lib/oddb2xml/util.rb', line 29

def Oddb2xml.html_decode(string)
  german = string
  german = string.force_encoding('ISO-8859-1').encode('UTF-8') if string.encoding.to_s.eql?('ASCII')
  while !german.eql?(HTMLEntities.new.decode(german))
    german = HTMLEntities.new.decode(german)
  end
  Oddb2xml.patch_some_utf8(german).gsub('<br>',"\n")
end

.log(msg) ⇒ Object 



75
76
77
78
79
80
 
# File 'lib/oddb2xml/util.rb', line 75

def Oddb2xml.log(msg)
  return unless @options[:log]
  # TODO:: require 'pry'; binding.pry if msg.size > 1000
  $stdout.puts "#{Time.now.strftime("%Y-%m-%d %H:%M:%S")}: #{msg[0..250]}"
  $stdout.flush
end

.log_timestamp(msg) ⇒ Object 



5
6
7
8
9
10
 
# File 'lib/oddb2xml/compare.rb', line 5

def self.log_timestamp(msg)
  full_msg = "#{Time.now.strftime("%H:%M:%S")}: #{msg}"
  puts full_msg
  STDOUT.flush
  full_msg
end

.patch_some_utf8(line) ⇒ Object 



38
39
40
41
42
43
44
45
46
47
48
49
 
# File 'lib/oddb2xml/util.rb', line 38

def Oddb2xml.patch_some_utf8(line)
  begin
    line = line.encode('utf-8')
  rescue => error
  end
  begin
    line.gsub("\u0089", "‰").gsub("\u0092", '’').gsub("\u0096", '-').gsub("\u2013",'-').gsub("\u201D", '"').chomp
  rescue => error
    puts "#{error}: in #{line}"
    line
  end
end

.save_options(options) ⇒ Object 



82
83
84
 
# File 'lib/oddb2xml/util.rb', line 82

def Oddb2xml.save_options(options)
  @options = options
end

.setEan13forNo8(no8, ean13) ⇒ Object 



212
213
214
215
216
217
218
219
220
221
222
 
# File 'lib/oddb2xml/util.rb', line 212

def Oddb2xml.setEan13forNo8(no8, ean13)
  if ean13.to_i == 7680006660045  ||  ean13.to_i == 7680006660014
    puts "setEan13forNo8 #{no8} ean13 #{ean13}"
  end
  if @@no8_to_ean13[no8].nil?
    @@no8_to_ean13[no8] = ean13
    @@ean13_to_no8[ean13] = no8
  elsif !@@no8_to_ean13[no8].eql?(ean13)
    puts "@@no8_to_ean13[no8] #{@@no8_to_ean13[no8]} not overridden by #{ean13}"
  end
end

.setEan13forProdno(prodno, ean13) ⇒ Object 



204
205
206
207
208
209
210
211
 
# File 'lib/oddb2xml/util.rb', line 204

def Oddb2xml.setEan13forProdno(prodno, ean13)
  if ean13.to_i == 7680006660045  ||  ean13.to_i == 7680006660014
    puts "setEan13forProdno #{prodno} ean13 #{ean13}"
  end
  @@prodno_to_ean13[prodno] ||= []
  @@prodno_to_ean13[prodno] << ean13
  @@ean13_to_prodno[ean13] = prodno
end

.skip_download(file) ⇒ Object 



90
91
92
93
94
95
96
97
98
 
# File 'lib/oddb2xml/util.rb', line 90

def Oddb2xml.skip_download(file)
  return false if defined?(VCR)
  dest = "#{Downloads}/#{File.basename(file)}"
  if File.exists?(dest)
    FileUtils.cp(dest, file, :verbose => false, :preserve => true) unless File.expand_path(file).eql?(dest)
    return true
  end
  false
end

.skip_download? ⇒ Boolean

Returns:

  • (Boolean) 


86
87
88
 
# File 'lib/oddb2xml/util.rb', line 86

def Oddb2xml.skip_download?
  @options[:skip_download]
end

.validate_via_xsd(xsd_file, xml_file) ⇒ Object 



183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
 
# File 'lib/oddb2xml/util.rb', line 183

def Oddb2xml.validate_via_xsd(xsd_file, xml_file)
  xsd =open(xsd_file).read
  xsd_rtikelstamm_xml = Nokogiri::XML::Schema(xsd)
  doc = Nokogiri::XML(File.read(xml_file))
  xsd_rtikelstamm_xml.validate(doc).each do
    |error|
      if error.message
        puts "Failed validating #{xml_file} with #{File.size(xml_file)} bytes using XSD from #{xsd_file}"
        puts "CMD: xmllint --noout --schema #{xsd_file} #{xml_file}"
      end
      msg = "expected #{error.message} to be nil\nfor #{xml_file}"
      puts msg
      expect(error.message).to be_nil, msg
  end
end

.verify_sha256(file) ⇒ Object 



167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
 
# File 'lib/oddb2xml/util.rb', line 167

def Oddb2xml.verify_sha256(file)
  f = File.open(file)
  doc = Nokogiri::XML(f)
  nr = 0
  doc.root.elements.each do |node|
    nr += 1
    next if node.name.eql?('RESULT')
    sha256 = Digest::SHA256.hexdigest node.text
    unless node['SHA256'].eql?(sha256)
      puts "Verifiying #{node['SHA256']} != expectd #{sha256} against node #{node.text} failed"
      exit (3)
    end
  end
  return true
end