patinfo2csv 1.0.0 → 1.0.1

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (8) hide show
  1. data/History.txt +6 -0
  2. data/README.txt +3 -2
  3. data/bin/patinfo2csv +10 -10
  4. data/lib/patinfo2csv/cli.rb +16 -15
  5. data/lib/patinfo2csv/converter.rb +17 -13
  6. data/lib/patinfo2csv.rb +103 -78
  7. data/lib/version.rb +1 -1
  8. metadata +7 -7
data/History.txt CHANGED
@@ -1,3 +1,9 @@
1
+ === 1.0.1 / 23.03.2012
2
+
3
+ * Added 3-bit character patterns
4
+ * updated as pattern
5
+ * added pattern for unexpected string
6
+
1
7
  === 1.0.0 / 22.03.2012
2
8
 
3
9
  * Initial commit for patinfo2csv
data/README.txt CHANGED
@@ -13,13 +13,14 @@ only 4 chapters(de, text) in patinfo.yaml are extracted.
13
13
  * usage
14
14
  * composition
15
15
 
16
- You can specify the pharmacode, see `patinfo2csv --help`
16
+ You can specify the eancode, see `patinfo2csv --help`
17
17
 
18
18
  == Usage
19
19
 
20
20
  $ patinfo2csv --help
21
- Usage: patinfo2csv <input patinfo.yaml> <output patinfo.csv> <input pharmacode.txt>
21
+ Usage: patinfo2csv <input patinfo.yaml> <output patinfo.csv> <input eancode.txt> [-l (de|fr)]
22
22
  -v, --version Show version
23
+ -l, --lang Language option (de|fr) default de
23
24
  -h, --help Print this help message
24
25
 
25
26
  == Requirements
data/bin/patinfo2csv CHANGED
@@ -11,7 +11,7 @@ LANGUAGE = [
11
11
 
12
12
  argv = ARGV.dup
13
13
  opts = Slop.parse! argv, :help => true do
14
- banner "Usage: patinfo2csv <input patinfo.yaml> <output patinfo.csv> <input pharmacode.txt> [-l (de|fr)]"
14
+ banner "Usage: patinfo2csv <input patinfo.yaml> <output patinfo.csv> <input eancode.txt> [-l (de|fr)]"
15
15
  on :v, :version, 'Show version'
16
16
  on :l, :lang, 'Language option (de|fr) default de', :argument => true
17
17
  end
@@ -23,6 +23,15 @@ else
23
23
  lang = "de"
24
24
  end
25
25
 
26
+ root = Pathname.new(__FILE__).realpath.parent.parent
27
+ $:.unshift root.join('lib') if $0 == __FILE__
28
+
29
+ if opts.version? then
30
+ require 'version'
31
+ puts "patinfo2csv " + Patinfo2csv::VERSION + "\n"
32
+ exit
33
+ end
34
+
26
35
  if argv.empty? \
27
36
  or !argv.length.eql? 3 \
28
37
  or !File.exist?(argv[0]) \
@@ -32,14 +41,5 @@ if argv.empty? \
32
41
  exit
33
42
  end
34
43
 
35
- root = Pathname.new(__FILE__).realpath.parent.parent
36
- $:.unshift root.join('lib') if $0 == __FILE__
37
-
38
- if opts.version? then
39
- require 'patinfo2csv/version'
40
- puts "patinfo2csv " + Patinfo2csv::VERSION + "\n"
41
- exit
42
- end
43
-
44
44
  require 'patinfo2csv'
45
45
  Patinfo2csv::CLI.run(argv[0], argv[2], argv[1], lang)
data/lib/patinfo2csv/cli.rb CHANGED
@@ -7,22 +7,23 @@ require 'patinfo2csv'
7
7
  module Patinfo2csv
8
8
  class CLI
9
9
  class << self
10
- # Param:: patinfo.yaml
11
- # Param:: pharmacode.txt
12
- # Param:: patinfo.csv
13
- def run(patinfo, pharmacode, output, lang)
10
+ # Param:: patinfo_yaml patinfo.yaml
11
+ # Param:: code_txt eancode.txt
12
+ # Param:: output_file patinfo.csv
13
+ # Param:: lang (de|fr)
14
+ def run(patinfo_yaml, code_txt, output_file, lang)
14
15
  cv = Patinfo2csv::Converter.new()
15
- cv.lang = lang
16
- cv.patinfo = self.load_yaml(patinfo)
17
- cv.pharmacode = self.parse_txt(pharmacode)
16
+ cv.lang = lang
17
+ cv.patinfos = self.load_yaml(patinfo_yaml)
18
+ cv.codes = self.parse_txt(code_txt)
18
19
  rows = cv.to_csv
19
- self.output_rows(rows, output)
20
- report(rows, cv.pharmacode, lang)
20
+ self.output_rows(rows, output_file)
21
+ report(rows, cv.codes, lang)
21
22
  end
22
- def load_yaml(patinfo)
23
+ def load_yaml(patinfo_yaml)
23
24
  # before yaml loading, replace escaped chars
24
25
  file = ''
25
- File.open(patinfo, "r:ascii:utf-8") { |f|
26
+ File.open(patinfo_yaml, "r:ascii:utf-8") { |f|
26
27
  while line = f.gets
27
28
  ESCAPED_STR_CODE_MAP.each do |code, char|
28
29
  line.gsub!(code, char)
@@ -33,18 +34,18 @@ module Patinfo2csv
33
34
  fh = StringIO.new file
34
35
  YAML.load_documents(fh) # utf-8
35
36
  end
36
- def parse_txt(pharmacode)
37
+ def parse_txt(code_txt)
37
38
  codes = []
38
- File.open(pharmacode, "r:utf-8") do |input|
39
+ File.open(code_txt, "r:utf-8") do |input|
39
40
  while line=input.gets
40
41
  codes << line.gsub(/[^\d]/, '').chomp
41
42
  end
42
43
  end
43
44
  codes
44
45
  end
45
- def output_rows(rows, output)
46
+ def output_rows(rows, output_file)
46
47
  return unless rows
47
- open(output, "w:utf-8") do |out|
48
+ open(output_file, "w:utf-8") do |out|
48
49
  rows.each do |row|
49
50
  out.print row, "\n"
50
51
  end
data/lib/patinfo2csv/converter.rb CHANGED
@@ -3,7 +3,7 @@
3
3
 
4
4
  module Patinfo2csv
5
5
  class Converter
6
- attr_accessor :patinfo, :pharmacode, :lang
6
+ attr_accessor :patinfos, :codes, :lang
7
7
  attr_reader :rows
8
8
  CHAPTERS = [
9
9
  :effects,
@@ -14,32 +14,36 @@ module Patinfo2csv
14
14
  ]
15
15
  DELIMITER = ';'
16
16
  def initialize
17
- @lang = "de"
18
- @patinfo = []
19
- @pharmacode = []
20
- @rows = []
17
+ @lang = "de"
18
+ @patinfos = []
19
+ @codes = []
20
+ @rows = []
21
21
  end
22
22
  def to_csv
23
23
  data = []
24
- @patinfo.each_with_index do |row, index|
24
+ @patinfos.each_with_index do |row, index|
25
25
  target = []
26
26
  next if row['article_codes'].nil?
27
- row['article_codes'].each do |code|
28
- if @pharmacode.include?(code[:article_pcode])
29
- target << code[:article_pcode]
30
- @pharmacode.delete(code[:article_pcode])
27
+ row['article_codes'].each do |article|
28
+ if @codes.include?(article[:article_ean13]) # EAN
29
+ target << article[:article_ean13]
30
+ @codes.delete(article[:article_ean13])
31
+ #elsif @code.include?(code[:article_pcode]) # Pharmacode
32
+ # target << code[:article_pcode]
33
+ # @code.delete(code[:article_pcode])
31
34
  end
32
35
  end
33
36
  next if target.empty?
34
37
  chapters = extract_chapters(row)
35
38
  next if chapters.empty?
36
- target.each do |pcode|
37
- @rows << [%Q!"#{pcode}"!, chapters].flatten.join(DELIMITER)
39
+ target.each do |code|
40
+ @rows << [%Q!"#{code}"!, chapters].flatten.join(DELIMITER)
38
41
  end
39
42
  end
40
43
  unless @rows.empty? #header
41
44
  @rows.unshift([
42
- "Pharmacode",
45
+ "EAN",
46
+ #"Pharmacode",
43
47
  "Effects",
44
48
  "Contra Indication",
45
49
  "Pregnancy",
data/lib/patinfo2csv.rb CHANGED
@@ -1,7 +1,7 @@
1
1
  #!/usr/bin/env ruby
2
2
  # encoding: utf-8
3
3
 
4
- require 'patinfo2csv/version'
4
+ require 'version'
5
5
  require 'patinfo2csv/cli'
6
6
  require 'patinfo2csv/converter'
7
7
 
@@ -10,82 +10,107 @@ module Patinfo2csv
10
10
  # escaped backslash and utf-8 code map
11
11
  # for patinfo.yaml(ASII)
12
12
  ESCAPED_STR_CODE_MAP = {
13
- "\\x24" => "$",
14
- "\\x25" => "%",
15
- "\\x26" => "&",
16
- "\\x2D" => "-",
17
- "\\x5F" => "_",
18
- "\\xC2\\xA0" => " ",
19
- "\\xC2\\xAB" => "«",
20
- "\\xC2\\xBB" => "»",
21
- "\\xC2\\xAE" => "®",
22
- "\\xC2\\xA9" => "©",
23
- "\\xC2\\xB5" => "µ",
24
- "\\xC2\\xBD" => "½",
25
- "\\xC2\\xBC" => "¼",
26
- "\\xC2\\xBE" => "¾",
27
- "\\xC3\\x82" => "Â",
28
- "\\xC3\\x83" => "Ã",
29
- "\\xC3\\x84" => "Ä",
30
- "\\xC3\\x85" => "Å",
31
- "\\xC3\\x86" => "Æ",
32
- "\\xC3\\x87" => "Ç",
33
- "\\xC3\\x88" => "È",
34
- "\\xC3\\x89" => "É",
35
- "\\xC3\\x8A" => "Ê",
36
- "\\xC3\\x8B" => "Ë",
37
- "\\xC3\\x8C" => "Ì",
38
- "\\xC3\\x8D" => "Í",
39
- "\\xC3\\x8E" => "Î",
40
- "\\xC3\\x8F" => "Ï",
41
- "\\xC3\\x90" => "Ð",
42
- "\\xC3\\x91" => "Ñ",
43
- "\\xC3\\x92" => "Ò",
44
- "\\xC3\\x93" => "Ó",
45
- "\\xC3\\x94" => "Ô",
46
- "\\xC3\\x95" => "Õ",
47
- "\\xC3\\x96" => "Ö",
48
- "\\xC3\\x97" => "×",
49
- "\\xC3\\x98" => "Ø",
50
- "\\xC3\\x99" => "Ù",
51
- "\\xC3\\x9A" => "Ú",
52
- "\\xC3\\x9B" => "Û",
53
- "\\xC3\\x9C" => "Ü",
54
- "\\xC3\\x9D" => "Ý",
55
- "\\xC3\\x9E" => "Þ",
56
- "\\xC3\\x9F" => "ß",
57
- "\\xC3\\xA0" => "à",
58
- "\\xC3\\xA1" => "á",
59
- "\\xC3\\xA2" => "â",
60
- "\\xC3\\xA3" => "ã",
61
- "\\xC3\\xA4" => "ä",
62
- "\\xC3\\xA5" => "å",
63
- "\\xC3\\xA6" => "æ",
64
- "\\xC3\\xA7" => "ç",
65
- "\\xC3\\xA8" => "è",
66
- "\\xC3\\xA9" => "é",
67
- "\\xC3\\xAA" => "ê",
68
- "\\xC3\\xAB" => "ë",
69
- "\\xC3\\xAC" => "ì",
70
- "\\xC3\\xAD" => "í",
71
- "\\xC3\\xAE" => "î",
72
- "\\xC3\\xAF" => "ï",
73
- "\\xC3\\xB0" => "ð",
74
- "\\xC3\\xB1" => "ñ",
75
- "\\xC3\\xB2" => "ò",
76
- "\\xC3\\xB3" => "ó",
77
- "\\xC3\\xB4" => "ô",
78
- "\\xC3\\xB5" => "õ",
79
- "\\xC3\\xB6" => "ö",
80
- "\\xC3\\xB7" => "÷",
81
- "\\xC3\\xB8" => "ø",
82
- "\\xC3\\xB9" => "ù",
83
- "\\xC3\\xBA" => "ú",
84
- "\\xC3\\xBB" => "û",
85
- "\\xC3\\xBC" => "ü",
86
- "\\xC3\\xBD" => "ý",
87
- "\\xC3\\xBE" => "þ",
88
- "\\xC3\\xBF" => "ÿ",
89
- "\\xE2\\x80\\x93" => "–",
13
+ /\\x24/ => "$",
14
+ /\\x25/ => "%",
15
+ /\\x26/ => "&",
16
+ /\\x2D/ => "-",
17
+ /\\x5F/ => "_",
18
+ /\\xC2\\xA0/ => " ",
19
+ /\\xC2\\xAB/ => "«",
20
+ /\\xC2\\xBB/ => "»",
21
+ /\\xC2\\xAE/ => "®",
22
+ /\\xC2\\xA9/ => "©",
23
+ /\\xC2\\xB5/ => "µ",
24
+ /\\xC2\\xBD/ => "½",
25
+ /\\xC2\\xBC/ => "¼",
26
+ /\\xC2\\xBE/ => "¾",
27
+ /\\xC2\\x[A-z0-9]{2}/ => "",
28
+ /\\xC3\\x82/ => "Â",
29
+ /\\xC3\\x83/ => "Ã",
30
+ /\\xC3\\x84/ => "Ä",
31
+ /\\xC3\\x85/ => "Å",
32
+ /\\xC3\\x86/ => "Æ",
33
+ /\\xC3\\x87/ => "Ç",
34
+ /\\xC3\\x88/ => "È",
35
+ /\\xC3\\x89/ => "É",
36
+ /\\xC3\\x8A/ => "Ê",
37
+ /\\xC3\\x8B/ => "Ë",
38
+ /\\xC3\\x8C/ => "Ì",
39
+ /\\xC3\\x8D/ => "Í",
40
+ /\\xC3\\x8E/ => "Î",
41
+ /\\xC3\\x8F/ => "Ï",
42
+ /\\xC3\\x90/ => "Ð",
43
+ /\\xC3\\x91/ => "Ñ",
44
+ /\\xC3\\x92/ => "Ò",
45
+ /\\xC3\\x93/ => "Ó",
46
+ /\\xC3\\x94/ => "Ô",
47
+ /\\xC3\\x95/ => "Õ",
48
+ /\\xC3\\x96/ => "Ö",
49
+ /\\xC3\\x97/ => "×",
50
+ /\\xC3\\x98/ => "Ø",
51
+ /\\xC3\\x99/ => "Ù",
52
+ /\\xC3\\x9A/ => "Ú",
53
+ /\\xC3\\x9B/ => "Û",
54
+ /\\xC3\\x9C/ => "Ü",
55
+ /\\xC3\\x9D/ => "Ý",
56
+ /\\xC3\\x9E/ => "Þ",
57
+ /\\xC3\\x9F/ => "ß",
58
+ /\\xC3\\xA0/ => "à",
59
+ /\\xC3\\xA1/ => "á",
60
+ /\\xC3\\xA2/ => "â",
61
+ /\\xC3\\xA3/ => "ã",
62
+ /\\xC3\\xA4/ => "ä",
63
+ /\\xC3\\xA5/ => "å",
64
+ /\\xC3\\xA6/ => "æ",
65
+ /\\xC3\\xA7/ => "ç",
66
+ /\\xC3\\xA8/ => "è",
67
+ /\\xC3\\xA9/ => "é",
68
+ /\\xC3\\xAA/ => "ê",
69
+ /\\xC3\\xAB/ => "ë",
70
+ /\\xC3\\xAC/ => "ì",
71
+ /\\xC3\\xAD/ => "í",
72
+ /\\xC3\\xAE/ => "î",
73
+ /\\xC3\\xAF/ => "ï",
74
+ /\\xC3\\xB0/ => "ð",
75
+ /\\xC3\\xB1/ => "ñ",
76
+ /\\xC3\\xB2/ => "ò",
77
+ /\\xC3\\xB3/ => "ó",
78
+ /\\xC3\\xB4/ => "ô",
79
+ /\\xC3\\xB5/ => "õ",
80
+ /\\xC3\\xB6/ => "ö",
81
+ /\\xC3\\xB7/ => "÷",
82
+ /\\xC3\\xB8/ => "ø",
83
+ /\\xC3\\xB9/ => "ù",
84
+ /\\xC3\\xBA/ => "ú",
85
+ /\\xC3\\xBB/ => "û",
86
+ /\\xC3\\xBC/ => "ü",
87
+ /\\xC3\\xBD/ => "ý",
88
+ /\\xC3\\xBE/ => "þ",
89
+ /\\xC3\\xBF/ => "ÿ",
90
+ /\\xC3\\x[A-z0-9]{2}/ => "",
91
+ /\\xE2\\x80\\x90/ => "‐",
92
+ /\\xE2\\x80\\x91/ => "‑",
93
+ /\\xE2\\x80\\x92/ => "‒",
94
+ /\\xE2\\x80\\x93/ => "–",
95
+ /\\xE2\\x80\\x94/ => "—",
96
+ /\\xE2\\x80\\x95/ => "―",
97
+ /\\xE2\\x80\\x98/ => "‘",
98
+ /\\xE2\\x80\\x99/ => "’",
99
+ /\\xE2\\x80\\x9A/ => "‚",
100
+ /\\xE2\\x80\\x9B/ => "‛",
101
+ /\\xE2\\x80\\x9C/ => "“",
102
+ /\\xE2\\x80\\x9D/ => "”",
103
+ /\\xE2\\x80\\x9E/ => "„",
104
+ /\\xE2\\x80\\x9F/ => "‟",
105
+ /\\xE2\\x80\\xA6/ => "…",
106
+ /\\xE2\\x80\\xB0/ => "‰",
107
+ /\\xE2\\x80\\xB1/ => "‱",
108
+ /\\xE2\\x80\\xB2/ => "′",
109
+ /\\xE2\\x80\\xB3/ => "″",
110
+ /\\xE2\\x80\\xB4/ => "‴",
111
+ /\\xE2\\x80\\xB5/ => "‵",
112
+ /\\xE2\\x80\\xB6/ => "‶",
113
+ /\\xE2\\x80\\xB7/ => "‷",
114
+ /\\xE2\\x[A-z0-9]{2}\\x[A-z0-9]{2}/ => "",
90
115
  }
91
116
  end
data/lib/version.rb CHANGED
@@ -1,3 +1,3 @@
1
1
  module Patinfo2csv
2
- VERSION = "1.0.0"
2
+ VERSION = "1.0.1"
3
3
  end
metadata CHANGED
@@ -1,7 +1,7 @@
1
1
  --- !ruby/object:Gem::Specification
2
2
  name: patinfo2csv
3
3
  version: !ruby/object:Gem::Version
4
- version: 1.0.0
4
+ version: 1.0.1
5
5
  prerelease:
6
6
  platform: ruby
7
7
  authors:
@@ -9,11 +9,11 @@ authors:
9
9
  autorequire:
10
10
  bindir: bin
11
11
  cert_chain: []
12
- date: 2012-03-22 00:00:00.000000000 Z
12
+ date: 2012-03-23 00:00:00.000000000 Z
13
13
  dependencies:
14
14
  - !ruby/object:Gem::Dependency
15
15
  name: rdoc
16
- requirement: &25013040 !ruby/object:Gem::Requirement
16
+ requirement: &10742100 !ruby/object:Gem::Requirement
17
17
  none: false
18
18
  requirements:
19
19
  - - ~>
@@ -21,10 +21,10 @@ dependencies:
21
21
  version: '3.10'
22
22
  type: :development
23
23
  prerelease: false
24
- version_requirements: *25013040
24
+ version_requirements: *10742100
25
25
  - !ruby/object:Gem::Dependency
26
26
  name: hoe
27
- requirement: &25012580 !ruby/object:Gem::Requirement
27
+ requirement: &10741680 !ruby/object:Gem::Requirement
28
28
  none: false
29
29
  requirements:
30
30
  - - ~>
@@ -32,7 +32,7 @@ dependencies:
32
32
  version: '2.13'
33
33
  type: :development
34
34
  prerelease: false
35
- version_requirements: *25012580
35
+ version_requirements: *10741680
36
36
  description: ! 'Patinfo2csv extracts and converts patinfo.yaml to patinfo.csv;
37
37
 
38
38
  only 4 chapters(de, text) in patinfo.yaml are extracted.
@@ -49,7 +49,7 @@ description: ! 'Patinfo2csv extracts and converts patinfo.yaml to patinfo.csv;
49
49
  * composition
50
50
 
51
51
 
52
- You can specify the pharmacode, see `patinfo2csv --help`'
52
+ You can specify the eancode, see `patinfo2csv --help`'
53
53
  email:
54
54
  - yasaka@ywesee.com, zdavatz@ywesee.com
55
55
  executables: