Class: ODDB::Import::Pharma24

Inherits:
Importer
  • Object
show all
Defined in:
lib/oddb/import/pharma24.rb

Instance Method Summary collapse

Methods inherited from Importer

#capitalize_all, #company_name, #postprocess, #utf8

Constructor Details

#initialize ⇒ Pharma24

Returns a new instance of Pharma24.



10
11
12
13
14
15
 
# File 'lib/oddb/import/pharma24.rb', line 10

def initialize
  @count = 0
  @created_companies = 0
  @found = 0
  @host = 'http://www.apotheke-online-internet.de'
end

Instance Method Details

#extract_data(page) ⇒ Object 



73
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
 
# File 'lib/oddb/import/pharma24.rb', line 73

def extract_data page
  data = []
  ## this should be page/'table[h2/a]'
  #  -> but Nokogiri apparently can't handle that
  all_tables = interesting_tables page
  duplicates = []
  all_tables.each do |table|
    duplicates.concat interesting_tables(table)
  end
  (all_tables - duplicates).each do |table|
    link, = table/'h2/a'
    if link
      prod = {
        :name => utf8(link.inner_text),
        :url  => link.attribute('href').to_s,
      }
      if price = (table/:strong).first
        prod.store :price_public, price.inner_text.tr(',', '.').to_f
      end
      ## should be (table/'td[text()="Abgabehinweis:"]').first
      #  -> but Nokogiri apparently can't handle that
      if prescription = td_with_text(table, "Abgabehinweis:")
        td, = prescription.xpath('following-sibling::td')
        prod.store :code_prescription,
                   !!/Rezeptpflichtig/.match(td.inner_text)
      end
      ## should be (table/'td[text()="Packungsinhalt:"]').first
      #  -> but Nokogiri apparently can't handle that
      if content = td_with_text(table, "Packungsinhalt:")
        td, = content.xpath('following-sibling::td')
        size_str = td.inner_text
        if match = /\s*(.*)\s+(\S+)\s+(\S+)\s*$/.match(size_str)
          size = utf8 match[1]
          unit = utf8 match[2]
          name = utf8 match[3]
          if size.empty?
            size, unit, name = unit, name, nil
          end
          prod.update :size => size, :unit => unit, :unitname => name
        end
      end
      if company = (table/'a[@class="liste"]').first
        prod.store :company, utf8(company.inner_text)
      end
      data.push prod
    end
  end
  data
end

#get_alphabetical(agent, fst, snd) ⇒ Object 



62
63
64
65
66
67
68
69
70
71
72
 
# File 'lib/oddb/import/pharma24.rb', line 62

def get_alphabetical agent, fst, snd
  url = "#@host/#{fst}#{snd}.html"
  page = agent.get url
  data = extract_data page
  while (link = (page/'//a[@class="pageResults"]').last) \
          && link.inner_text == '[n?chste?>>]'
    page = agent.get link.attributes['href']
    data.concat extract_data(page)
  end
  data
end

#import(agent, packages, opts = {:all => false}) ⇒ Object 



16
17
18
19
20
21
22
23
 
# File 'lib/oddb/import/pharma24.rb', line 16

def import(agent, packages, opts={:all => false})
  agent.max_history = 1
  packages.collect! { |package| package.odba_id }
  while id = packages.shift
    update_package(agent, ODBA.cache.fetch(id), opts)
  end
  report
end

#import_company(data) ⇒ Object 



24
25
26
27
28
29
30
31
32
33
 
# File 'lib/oddb/import/pharma24.rb', line 24

def import_company(data)
  name = company_name(data[:company])
  company = Business::Company.find_by_name(name) 
  if(company.nil?)
    @created_companies += 1
    company = Business::Company.new
    company.name.de = name
  end
  company
end

#import_size(data, package) ⇒ Object 



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/oddb/import/pharma24.rb', line 34

def import_size(data, package)
  part = package.parts.first || package.add_part(Drugs::Part.new)
  dose, size, multi = data[:size].to_s.split(/x/i, 3).reverse.compact
  unit = data[:unit].to_s
  if(unit != 'St')
    part.quantity = Drugs::Dose.new(dose, unit)
  elsif(multi.nil?)
    multi = size
    size = dose
  end
  multi = multi.to_i
  size = size.to_i
  part.multi = (multi > 0) ? multi : nil
  part.size = (size > 0) ? size : nil
  if(unitname = data[:unitname])
    unit = Drugs::Unit.find_by_name(unitname)
    unless(unit)
      unit = Drugs::Unit.new  
      unit.name.de = unitname
      unit.save
    end
    part.unit = unit
  end
  part.save
end

#interesting_tables(node) ⇒ Object 



59
60
61
 
# File 'lib/oddb/import/pharma24.rb', line 59

def interesting_tables node
  (node/'table').find_all do |inner_node| !(inner_node/'h2/a').empty? end.to_a
end

#report ⇒ Object 



122
123
124
125
126
127
128
129
 
# File 'lib/oddb/import/pharma24.rb', line 122

def report
  lines = [
    sprintf("Checked %5i Packages", @count),
    sprintf("Updated %5i Packages", @found),
    sprintf("Created %5i Companies", @created_companies),
  ]
  lines
end

#search(agent, term) ⇒ Object 



130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
 
# File 'lib/oddb/import/pharma24.rb', line 130

def search agent, term
  url = "#@host/advanced_search_result.php?keywords=#{term}"
  page = agent.get url
  extract_data page
rescue Zlib::GzipFile::Error => err
  retries ||= 3
  if retries > 0
    retries -= 1
    retry
  else
    err.message << " after 3 retries - url: #{url}"
    raise err
  end
rescue StandardError => err
  err.message << " url: #{url}"
  raise err
end

#td_with_text(table, text) ⇒ Object 



147
148
149
150
151
152
 
# File 'lib/oddb/import/pharma24.rb', line 147

def td_with_text table, text
  nodes = (table/'td').find_all do |node|
    node.text.strip == text
  end
  nodes.first
end

#update_package(agent, package, opts = {}) ⇒ Object 



153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
 
# File 'lib/oddb/import/pharma24.rb', line 153

def update_package agent, package, opts={}
  price = package.price(:public)
  resale = [ :pharma24,
             :csv_product_infos ].include?(package.data_origin(:price_public))
  needs_update = opts[:all] || price.nil? || resale
  if needs_update && (code = package.code(:cid, 'DE'))
    @count += 1
    data, = search agent, code.value
    if data
      @found += 1
      package.name.de = u(data[:name])
      presc = data[:code_prescription]
      if(code = package.code(:prescription))
        if(code.value != presc)
          code.value = presc
        end
      else
        package.add_code Util::Code.new(:prescription, presc, 'DE')
      end
      amount = data[:price_public]
      if(amount > 0)
        update_price package, :public, amount
        if presc
          update_price package, :exfactory, package._price_exfactory
        end
      end
      import_size data, package
      package.save
      if((product = package.product) && product.company.nil?)
        product.company = import_company(data)
        product.save
      end
    end
  end
end

#update_price(package, type, amount) ⇒ Object 



188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
 
# File 'lib/oddb/import/pharma24.rb', line 188

def update_price package, type, amount
  dotype = :"price_#{type}"
  # if this price has been edited manually we won't overwrite
  unless((data_origin = package.data_origin(dotype)) \
     && data_origin.to_s.include?('@'))
    either = false
    if(price = package.price(type, 'DE'))
      if(price != amount)
        price.amount = amount
        either = true
      end
    else
      price = Util::Money.new(amount, type, 'DE')
      package.add_price(price)
      either = true
    end
    if either
      package.data_origins.store dotype, :pharma24
    end
  end
end