Class: Remi::DataSource::CsvFile

Inherits:
Object
  • Object
show all
Includes:
Remi::DataSource, DataStub
Defined in:
lib/remi/cucumber/data_source.rb,
lib/remi/data_source/csv_file.rb

Instance Attribute Summary collapse

Class Method Summary collapse

Instance Method Summary collapse

Methods included from Remi::DataSource

#feild_symbolizer

Methods included from Remi::DataSubject

#df=

Methods included from DataStub

#empty_stub_df, #stub_df, #stub_row_array, #stub_values

Constructor Details

#initialize(fields: {}, extractor:, csv_options: {}, filename_field: nil, logger: Remi::Settings.logger) ⇒ CsvFile

Returns a new instance of CsvFile.



20
21
22
23
24
25
26
 
# File 'lib/remi/data_source/csv_file.rb', line 20

def initialize(fields: {}, extractor:, csv_options: {}, filename_field: nil, logger: Remi::Settings.logger)
  @fields = fields
  self.extractor = extractor
  @csv_options = self.class.default_csv_options.merge(csv_options)
  @filename_field = filename_field
  @logger = logger
end

Instance Attribute Details

#csv_options ⇒ Object (readonly)

Returns the value of attribute csv_options.



30
31
32
 
# File 'lib/remi/data_source/csv_file.rb', line 30

def csv_options
  @csv_options
end

#extractor ⇒ Object

Returns the value of attribute extractor.



29
30
31
 
# File 'lib/remi/data_source/csv_file.rb', line 29

def extractor
  @extractor
end

#fields ⇒ Object

Returns the value of attribute fields.



28
29
30
 
# File 'lib/remi/data_source/csv_file.rb', line 28

def fields
  @fields
end

Class Method Details

.default_csv_options ⇒ Object 



8
9
10
11
12
13
14
15
16
17
 
# File 'lib/remi/data_source/csv_file.rb', line 8

def self.default_csv_options
  CSV::DEFAULT_OPTIONS.merge({
    headers: true,
    header_converters: Remi::FieldSymbolizers[:standard],
    converters: [],
    col_sep: ',',
    encoding: 'UTF-8',
    quote_char: '"'
  })
end

Instance Method Details

#df ⇒ Object 



96
97
98
 
# File 'lib/remi/data_source/csv_file.rb', line 96

def df
  @dataframe ||= to_dataframe
end

#extract ⇒ Object 



36
37
38
 
# File 'lib/remi/data_source/csv_file.rb', line 36

def extract
  @extracted = Array(@extractor.extract)
end

#extracted ⇒ Object 



40
41
42
 
# File 'lib/remi/data_source/csv_file.rb', line 40

def extracted
  @extracted || extract
end

#field_symbolizer ⇒ Object 



32
33
34
 
# File 'lib/remi/data_source/csv_file.rb', line 32

def field_symbolizer
  self.class.default_csv_options[:header_converters]
end

#first_line ⇒ Object 



63
64
65
66
67
68
 
# File 'lib/remi/data_source/csv_file.rb', line 63

def first_line
  # Readline assumes \n line endings.  Strip out \r if it is a DOS file.
  @first_line ||= File.open(source_filename) do |f|
    f.readline.gsub(/\r/,'')
  end
end

#headers ⇒ Object 



70
71
72
 
# File 'lib/remi/data_source/csv_file.rb', line 70

def headers
  @headers ||= CSV.open(source_filename, 'r', source_csv_options) { |csv| csv.first }.headers
end

#source_filename ⇒ Object

Only going to support single file for now



58
59
60
61
 
# File 'lib/remi/data_source/csv_file.rb', line 58

def source_filename
  raise "Multiple source files detected" if extracted.size > 1
  @source_filename ||= extracted.first
end

#stub_header ⇒ Object 



48
49
50
 
# File 'lib/remi/cucumber/data_source.rb', line 48

def stub_header
  @fields.keys.join(@csv_options[:col_sep])
end

#stub_row_csv ⇒ Object 



52
53
54
 
# File 'lib/remi/cucumber/data_source.rb', line 52

def stub_row_csv
  stub_row_array.join(@csv_options[:col_sep])
end

#stub_tmp_file ⇒ Object 



35
36
37
 
# File 'lib/remi/cucumber/data_source.rb', line 35

def stub_tmp_file
  @stub_tmp_file ||= Tempfile.new('stub_tmp_file.csv').path
end

#to_dataframe ⇒ Object 



78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
 
# File 'lib/remi/data_source/csv_file.rb', line 78

def to_dataframe
  # Assumes that each file has exactly the same structure
  result_df = nil
  extracted.each_with_index do |filename, idx|
    @logger.info "Converting #{filename} to a dataframe"
    csv_df = Daru::DataFrame.from_csv filename, @csv_options

    csv_df[@filename_field] = Daru::Vector.new([filename] * csv_df.size, index: csv_df.index) if @filename_field
    if idx == 0
      result_df = csv_df
    else
      result_df = result_df.concat csv_df
    end
  end

  result_df
end

#valid_headers? ⇒ Boolean

Returns:

  • (Boolean) 


74
75
76
 
# File 'lib/remi/data_source/csv_file.rb', line 74

def valid_headers?
  (fields.keys - headers).empty?
end

#write_stub_tmp_file ⇒ Object 



39
40
41
42
43
44
45
46
 
# File 'lib/remi/cucumber/data_source.rb', line 39

def write_stub_tmp_file
  File.open(stub_tmp_file, "wb") do |file|
    file.puts stub_header
    file.puts stub_row_csv
  end

  stub_tmp_file
end