Class: PseudonymisedFileWrapper

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

Overview

Provide the ability to extract fieldnames and create CSV output from .pseudo files

Instance Method Summary collapse

Constructor Details

#initialize(filename) ⇒ PseudonymisedFileWrapper

Returns a new instance of PseudonymisedFileWrapper.



24
25
26
27
 
# File 'lib/ndr_pseudonymise/pseudonymised_file_wrapper.rb', line 24

def initialize(filename)
  @filename = filename
  @logger = Logger.new(STDOUT)
end

Instance Method Details

#available_fields ⇒ Object 



29
30
31
 
# File 'lib/ndr_pseudonymise/pseudonymised_file_wrapper.rb', line 29

def available_fields
  (@all_fields1 + @all_fields2).sort.uniq
end

#pretty_data ⇒ Object 



81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
 
# File 'lib/ndr_pseudonymise/pseudonymised_file_wrapper.rb', line 81

def pretty_data
  csv_string = CSV.generate do |csv|
    headers = (@all_fields1.map { |name| "mapped:#{name}" } +
    @all_fields2.map { |name| "raw:#{name}" } +
    %w(pseudo_id1 pseudo_id2 key_bundle))
    csv << headers
    @processed_lines.each do |line|
      output_fields = @all_fields1.map { |field| line[:map1][field] } +
      @all_fields2.map { |field| line[:map2][field] }
      output_fields.push(line[:id1], line[:id2], line[:keys])
      csv << output_fields
    end
  end
  csv_string
end

#pretty_write ⇒ Object

Create an excel-readable CSV file, in the same location as the original



63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
 
# File 'lib/ndr_pseudonymise/pseudonymised_file_wrapper.rb', line 63

def pretty_write
  /(?<base_name>.*)\.(?:csv|(?:zip|xlsx?)\.pseudo)/i.match(@filename)
  target_filename = "#{$LAST_MATCH_INFO[:base_name]}_pretty.csv"
  @logger.debug "Writing output to #{target_filename}"
  CSV.open(target_filename, 'w') do |file|
    headers = (@all_fields1.map { |name| "mapped:#{name}" } +
               @all_fields2.map { |name| "raw:#{name}" } +
               %w(pseudo_id1 pseudo_id2 key_bundle))
    file << headers
    @processed_lines.each do |line|
      output_fields = @all_fields1.map { |field| line[:map1][field] } +
                      @all_fields2.map { |field| line[:map2][field] }
      output_fields.push(line[:id1], line[:id2], line[:keys])
      file << output_fields
    end
  end
end

#process ⇒ Object

Read in the source file, accumulating all the field names used in any row



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
59
60
 
# File 'lib/ndr_pseudonymise/pseudonymised_file_wrapper.rb', line 34

def process
  line_counter = 1
  processed_lines = []
  all_fields1 = []
  all_fields2 = []
  CSV.foreach(@filename) do |row|
    if row.size == 1
    # Header; do nothing
    elsif row.size == 7
      cur = { map1: JSON.parse(row[4]),
              map2: JSON.parse(row[6]),
              id1: row[0],
              id2: row[1],
              keys: row[2] }
      processed_lines.push(cur)
      all_fields1.push(*cur[:map1].keys).uniq!
      all_fields2.push(*cur[:map2].keys).uniq!
    else
      @logger.debug"Line #{line_counter} contained unexpected number of fields: #{row.size}"
    end
    line_counter += 1
  end
  @lines = line_counter
  @all_fields1 = all_fields1
  @all_fields2 = all_fields2
  @processed_lines = processed_lines
end