Class: ODDB::FiParse::MiniFi::Handler

Inherits:
Rpdf2txt::ColumnHandler
  • Object
show all
Defined in:
ext/fiparse/src/minifi.rb

Constant Summary collapse

SECTION_HEADINGS = 
%r{^(Indi[ck]ation
  |Dosierung:?$|Posologie:?$
  |Wirksamkeit|Efficacit.
  |Unerw.nschte Wirkungen|Effets ind.sirables
  |Pharmakokinetik|Pharmacocin.tique
  |Interaktionen|Interactions
  |Kontraindikationen|Contre-?indications
  )
}uix
HYPHEN = 
/-$/u
NAME_PTRN = 
%r{
(?:Wirk?s?toffe?s:
[^\(]*\((?<name>[^,\s®\(]+))(®|\(TM\))
|
(?:(?:(?:(?:Impf|Wirk?)stoffe?s?)
     |(?:Zulassung\s*des\s*ersten[^:]+)):
(?:\s+Informationen\s+zu)?
\s*(?<name>[^,\s®\(]+(,?\s*ad\s*us\.\s*vet\.)?))
|
^(?:Zulassung\s*von\s*[,"„«]{1,2})(?<name>[^,"“»]+)
}ux

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize ⇒ Handler

Returns a new instance of Handler.



34
35
36
37
38
 
# File 'ext/fiparse/src/minifi.rb', line 34

def initialize
  super('', '')
  @minifis = []
  start_page
end

Instance Attribute Details

#minifis ⇒ Object (readonly)

Returns the value of attribute minifis.



12
13
14
 
# File 'ext/fiparse/src/minifi.rb', line 12

def minifis
  @minifis
end

Instance Method Details

#clean! ⇒ Object 



39
40
41
42
43
44
 
# File 'ext/fiparse/src/minifi.rb', line 39

def clean!
  if(@chapter.respond_to?(:clean!))
    @chapter.clean!
    @chapter = nil
  end
end

#column_count ⇒ Object 



95
96
97
 
# File 'ext/fiparse/src/minifi.rb', line 95

def column_count
  2
end

#create_document ⇒ Object 



45
46
47
48
 
# File 'ext/fiparse/src/minifi.rb', line 45

def create_document
  clean!
  @chapter = Text::Chapter.new
end

#new_font(font) ⇒ Object 



49
50
51
52
 
# File 'ext/fiparse/src/minifi.rb', line 49

def new_font(font)
  set_bold(font)
  @font = font
end

#send_column ⇒ Object 



53
54
55
56
 
# File 'ext/fiparse/src/minifi.rb', line 53

def send_column
  super
  set_bold(@font)
end

#send_line_break ⇒ Object 



57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
 
# File 'ext/fiparse/src/minifi.rb', line 57

def send_line_break
  line = (@lines.last || []).join.strip
  case line
  when %r{^Zulassung eines( pflanzlichen)? Arzneimittels mit( einem)? neue[mn] Wirk?stoff}iu,
    %r{^Neuzulassung eines Medikamente?s mit( einem)? neue[mn] Wirkstoff}iu,
    %r{^(Erst)?Zulassung eines neuen (Impf|Wirk)stoffe?s}iu,
    %r{^Zulassung des ersten}iu, %r{^Zulassung von [,"„«]{1,2}}iu
    if(@reached_news)
      @current = {}
      @minifis.push(@current)
      @current[:de] = create_document
    end
  when %r{^Autorisation d.un (phyto)?m.dicament contenant un nouveau}iu,
    %r{^Autorisation d.livr.es? pour (un|de) nouveaux? principes? actifs?}iu,
    %r{^Autorisation d.livr.es? pour (un|de) nouveaux? produits? de vaccins?}iu,
    %r{^Autorisation d.un nouveau (principe actif|vaccin)}iu,
    %r{^Autorisation du premier}iu, %r{^Autorisation de [,"„«]{1,2}}iu,
    %r{^Autorisation d.(livr.e pour )?un m.dicament avec un nouveau}iu
    @current[:fr] = create_document if @current
  when %r{Arzneimittel Nachrichten( / M.dicaments)?$}u,
    %r{^M.dicaments$}u
    @reached_news = true
    #ignore
  when %r{Arzneimittel\s*Statistik\s*/\s*Miscellan.es}u,
    %r{Regulatory News / R.glementation}u, %r{Regulatory News}u,
    %r{Infosplitter(\s*/\s*En\s*vrac)?}u, 
    %r{Richtigstellung}u, %r{Berichtigung}u,
    %r{Medizinprodukte(\s*/\s*Dispositifs m.dicaux)?}u,
    %r{Zulassung eines Tierarzneimittels}u,
    %r{Entlassung aus der Heilmittelkontrolle}u, 
    %r{Mitteilung an die Fachpersonen}u,
    %r{Zum Vertrieb freigegebene Chargen}u,
    %r{^Heparinum natricum ad usum externum}u # dtstcpw for 02/2006
    clean!
    @current = nil
  end
  super
end

#send_page ⇒ Object 



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
133
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
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
 
# File 'ext/fiparse/src/minifi.rb', line 98

def send_page
  if(@chapter)
    max_idx = @lines.collect { |line| 
      line.size }.push(1).max - 1
    @lines.each { |line|
      line[max_idx] ||= nil
    }
    target, section = nil
    if(@chapter.empty?)
      target = @chapter.heading 
      section = @chapter.next_section
    else
      section = @chapter.sections.last 
      target = section.paragraphs.last || section.next_paragraph
    end
    previous_was_hyphenated = false
    row = 0
    @lines.transpose.each_with_index { |column, col|
      previous_width = nil
      column.each_with_index { |line, row|
        if(line.is_a?(String))
          line.gsub!(/\s+/, ' ')
          line.strip!
          case line
          when ''
            if(!target.empty? && !previous_was_hyphenated)
              if(section.subheading.empty?)
                section = @chapter.next_section
              end
              target = section.next_paragraph
            end
          when %r{Arzneimittel Nachrichten( / M.dicaments)?}u,
            %r{^\s*/?\s*M.dicaments$}u, %r{^Actualit.s$}u,
            # misplaced header in 05/2003:416
            %r{^Arzneimittel Statistik$}u
            #ignore
          when %r{Swissmedic\s*Journal\s+[_\d]+}u, %r{^\d+$}u
            target = section.next_paragraph
          else
            bold = @bold_lines[col][row]
            title = [@chapter.heading, 
              section.subheading].any? { |ttl| ttl.equal?(target) }
            if(bold && !previous_was_hyphenated \
               && (!title || (target == @chapter.heading \
                              && SECTION_HEADINGS.match(line))))
              section = @chapter.next_section
              target = section.subheading
            elsif(!bold && title)
              target << "\n"
              target = section.next_paragraph
            elsif(/^[o·-]\s/u.match(line))
              target = section.next_paragraph
            elsif(!title && previous_width \
                  && previous_width < (line.length * 0.65))
              target = section.next_paragraph
            end
            previous_width = line.length
            if(line.match(HYPHEN))
              previous_was_hyphenated = true
              if(!((nxt = column.at(row.next)) \
                   && /^[A-Z]/u.match(nxt.to_s)))
                target << line.gsub(HYPHEN, '')
              else
                target << line
              end
            else
              previous_was_hyphenated = false
              target << line << ' '
            end
            # Ensure all bold text at the start of a page counts
            # towards the heading - this solution works post 12/2005
            if bold && target == @chapter.heading \
              && (other = @lines[row][1])
              target << other
              other.replace ''
            end
          end
        elsif(@chapter.heading == target && NAME_PTRN.match(target))
          target = section.next_paragraph
        end
      }
    }
    if(match = NAME_PTRN.match(@chapter.heading))
      @current[:name] = match['name']
    end
  end
  start_page
  send_line_break
end

#set_bold(font) ⇒ Object 



187
188
189
190
191
192
193
194
195
196
197
198
 
# File 'ext/fiparse/src/minifi.rb', line 187

def set_bold(font)
  line = @columns[@current_column].to_s
  bold = @bold_lines[@current_column]
  prev = (@lines[-2] || [])[@current_column]
  if(font && font.bold? \
     && (bold[@lines.size - 1] || line.empty? \
         || SECTION_HEADINGS.match(line)))
    bold[@lines.size - 1] = true
  elsif(!(font && font.symbol?))
    bold[@lines.size - 1] = false
  end
end

#start_page ⇒ Object 



199
200
201
202
 
# File 'ext/fiparse/src/minifi.rb', line 199

def start_page
  @bold_lines = Hash.new { |me, key| me[key] = Hash.new }
  @lines.clear
end