Class: FnbPdfToCsv

Inherits:
Object
  • Object
show all
Defined in:
lib/fnb_pdf_to_csv.rb,
lib/fnb_pdf_to_csv/version.rb

Constant Summary collapse

AMOUNT = 
'\(?[0-9][0-9,]*\.[0-9]{2}\)?\s?(Cr)?'
DATE = 
'\d{2} \w{3}'
VERSION = 
"0.0.3"

Instance Attribute Summary collapse

Class Method Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(file) ⇒ FnbPdfToCsv

Returns a new instance of FnbPdfToCsv.



11
12
13
14
 
# File 'lib/fnb_pdf_to_csv.rb', line 11

def initialize file
  @reader = ::PDF::Reader.new file
  @lines = []
end

Instance Attribute Details

#lines ⇒ Object (readonly)

Returns the value of attribute lines.



6
7
8
 
# File 'lib/fnb_pdf_to_csv.rb', line 6

def lines
  @lines
end

Class Method Details

.parse(file) ⇒ Object 



16
17
18
19
20
 
# File 'lib/fnb_pdf_to_csv.rb', line 16

def self.parse file
  parser = self.new(file)
  parser.parse
  parser
end

Instance Method Details

#clean_amount(amount) ⇒ Object 



75
76
77
78
79
80
81
82
83
 
# File 'lib/fnb_pdf_to_csv.rb', line 75

def clean_amount(amount)
  return amount if amount.nil?
  return 0 - amount[1..-2].to_f if amount[0] == '(' and amount[-1] == ')'
  if amount[-2..-1] == 'Cr'
    return amount[0..-3].tr(',', '').to_f
  else
    return 0 - amount.tr(',', '').to_f
  end
end

#clean_date(date) ⇒ Object 



70
71
72
73
 
# File 'lib/fnb_pdf_to_csv.rb', line 70

def clean_date(date)
  day, month = date.split(/\s/)
  Time.new(Time.new.year, month, day.to_i).strftime("%Y-%m-%d")
end

#clean_line(line) ⇒ Object 



85
86
87
88
89
90
91
92
93
 
# File 'lib/fnb_pdf_to_csv.rb', line 85

def clean_line(line)
  sline = line.dup
  sline[0] = clean_date sline[0]
  sline[4] = clean_amount sline[4]
  sline[5] = clean_amount sline[5]
  sline[6] = clean_amount sline[6]

  sline
end

#mangle_line!(arr) ⇒ Object 



95
96
97
98
99
100
101
102
103
104
105
106
107
 
# File 'lib/fnb_pdf_to_csv.rb', line 95

def mangle_line! arr
  arr.delete_at 0
  arr.map! { |elm| elm.strip unless elm.nil? } # Cleanup

  arr.delete_at 3
  arr.delete_at 4
  arr.delete_at 5
  arr[1] = arr[1].split(/\s{2,}/) # We get the three descriptions as one string
  arr.insert(2, arr[1][1])        # So split them up and add them back
  arr.insert(3, arr[1][2])
  arr[1] = arr[1][0]
  arr
end

#output(file, separator = ',') ⇒ Object 



26
27
28
29
30
31
32
33
 
# File 'lib/fnb_pdf_to_csv.rb', line 26

def output file, separator = ','
  f = File.new file, 'w'
  f.write [
    'Date','Description1','Description2','Description3','Amount','Balance','Accrued Charges'
  ].to_csv(col_sep: separator)

  lines.each { |line| f.write clean_line(line).to_csv(col_sep: separator) }
end

#parse ⇒ Object 



22
23
24
 
# File 'lib/fnb_pdf_to_csv.rb', line 22

def parse
  @reader.pages.each { |page| parse_page page }
end

#parse_line(line) ⇒ Object 



64
65
66
67
68
 
# File 'lib/fnb_pdf_to_csv.rb', line 64

def parse_line line
  line.match(/^\s*(#{DATE})(.*?)(#{AMOUNT})\s+(#{AMOUNT})(\s+#{AMOUNT})?$/) do |m|
    @lines.push mangle_line!(m.to_a)
  end
end

#parse_page(page) ⇒ Object 



60
61
62
 
# File 'lib/fnb_pdf_to_csv.rb', line 60

def parse_page page
  page.text.each_line { |line| parse_line line }
end

#statement(file) ⇒ Object 



35
36
37
38
39
40
41
42
43
44
 
# File 'lib/fnb_pdf_to_csv.rb', line 35

def statement file
  f = File.new file, 'w'
  f.write "5,'Number','Date','Description1','Description2','Description3','Amount','Balance','Accrued Charges'\n"

  count = 1
  lines.each do |line|
    f.write statement_line(line, count).join(',') + "\n"
    count = count + 1
  end
end

#statement_line(line, count) ⇒ Object 



46
47
48
49
50
51
52
53
54
55
56
57
58
 
# File 'lib/fnb_pdf_to_csv.rb', line 46

def statement_line line, count
  sline = line.dup
  sline.insert(0, 5)
  sline.insert(1, count)
  sline[2] = "'#{sline[2]}'"
  sline[3] = '"' + sline[3] + '"' unless (sline[3].nil? or sline[3] == '')
  sline[4] = '"' + sline[4] + '"' unless (sline[4].nil? or sline[4] == '')
  sline[5] = '"' + sline[5] + '"' unless (sline[5].nil? or sline[5] == '')
  sline[6] = clean_amount(sline[6])
  sline[7] = clean_amount(sline[7])

  sline
end