CLI for filter

bin/csv-filter

@@ -0,0 +1,234 @@
+#!/usr/bin/env ruby
+
+require 'optparse'
+require 'csv'
+
+# Returns an array of names.
+# \String +s+ contains a comma-separated list of names.
+def parse_names(s)
+  s.split(",").collect(&:to_sym)
+end
+
+options = {}
+
+parser = OptionParser.new
+
+parser.version = CSV::VERSION
+parser.banner = <<-BANNER
+Usage: #{parser.program_name} [options]
+
+  Reads and parses the CSV text content of the standard input per the given input options.
+  From that content, generates CSV text per the given output options and writes
+  that text to the standard output.
+
+  Three options are similar for input and output:
+
+  - Row separator:
+
+    - The input row separator is determined:
+
+      - By option --input-row-sep, if given.
+      - Otherwise, by option --row-sep, if given.
+      - Otherwise, by auto-discovery.
+
+    - The output row separator is determined:
+
+      - By option --output-row-sep if given.
+      - Otherwise, by option --row-sep if given.
+      - Otherwise, as the newline character ("\\n").
+
+  - Column separator:
+
+    - The input column separator is determined:
+
+      - By option --input-col-sep, if given.
+      - Otherwise, by option --col-sep, if given.
+      - Otherwise, as the comma character (",").
+
+    - The output column separator is determined:
+
+      - By option --output-col-sep if given.
+      - Otherwise, by option --col-sep if given.
+      - Otherwise, as the comma character (",").
+
+  - Quote character:
+
+    - The input quote character is determined:
+
+      - By option --input-quote-char, if given.
+      - Otherwise, by option --quote-char, if given.
+      - Otherwise, as the double-quote character ('"').
+
+    - The output quote character is determined:
+
+      - By option --output-quote-char, if given.
+      - Otherwise, by option --quote-char, if given.
+      - Otherwise, as the double-quote character ('"').
+BANNER
+
+parser.separator(nil)
+parser.separator(
+  'Input Options: see https://docs.ruby-lang.org/en/3.3/CSV.html#class-CSV-label-Options+for+Parsing'
+)
+
+parser.separator(nil)
+parser.on('--converters=CONVERTERS',
+     'Input field converters;',
+     'CONVERTERS is comma-separated converter names (leading colon optional).'
+) do |value|
+  options[:converters] = parse_names(value)
+end
+# parser.on('--empty_value STR',
+#      'Value to be output for empty input field.') do |value|
+#   options[:empty_value] = value
+# end
+parser.on('--field-size-limit=N', Integer,
+     'Maximum field size (characters)') do |value|
+  options[:field_size_limit] = value
+end
+parser.on('--headers=ARG',
+     'Treat the first input row as headers;',
+     'ARG is true or comma-separated headers.'
+) do |value|
+  options[:headers] = value == 'true' ? true : parse_names(value)
+end
+parser.on('--header-converters=CONVERTERS',
+     'Input header converters;',
+     'CONVERTERS is comma-separated converter names.'
+) do |value|
+  options[:header_converters] = parse_names(value)
+end
+parser.on('--liberal-parsing',
+     'Attempt to parse non-conformant input.') do
+  options[:liberal_parsing] = :no_argument
+end
+# parser.on('--nil_value STR',
+#      'Value to be substituted for nil input field.') do |value|
+#   options[:nil_value] = value
+# end
+# parser.on('--return_headers',
+#      'Method shift returns header row.') do
+#   options[:return_headers] = :no_argument
+# end
+# parser.on('--skip_blanks',
+#      'Ignore blank lines.') do
+#   options[:skip_blanks] = :no_argument
+# end
+# parser.on('--skip_lines STR',
+#      'Ignore comment lines;',
+#      'STR is converted to a Regexp that identifies comment lines.') do |value|
+#   options[:skip_lines] = Regexp.new(value)
+# end
+# parser.on('--strip',
+#      'Strip whitespace from input fields.') do
+#   options[:strip] = :no_argument
+# end
+# parser.on('--unconverted_fields',
+#      'Define method unconverted_fields for each created CSV::Row object.') do
+#   options[:unconverted_fields] = :no_argument
+# end
+
+parser.separator(nil)
+parser.separator(
+  'Output Options: see https://docs.ruby-lang.org/en/3.3/CSV.html#class-CSV-label-Options+for+Generating'
+)
+
+parser.separator(nil)
+# parser.on('--force_quotes',
+#      'Double-quote each output field.') do
+#   options[:force_quotes] = :no_argument
+# end
+# parser.on('--quote_empty STR',
+#      'Double-quote each empty output field;',
+#      'STR is true or false.'
+# ) do
+#   options[:quote_empty] = :no_argument
+# end
+# parser.on('--write_converters',
+#      'Output field converters;',
+#      'not implemented in CLI.'
+# ) do |value|
+#   raise NotImplementedError.new('--write_converters is not implemented in the CLI.')
+# end
+# parser.on('--write_headers',
+#      'Write headers to output.') do |value|
+#   options[:write_headers] = :no_argument
+# end
+# parser.on('--write_empty_value STR',
+#      'Value to be substituted for empty output field.') do |value|
+#   options[:write_empty_value] = value
+# end
+# parser.on('--write_nil_value STR',
+#      'Value to be substituted for nil output field.') do |value|
+#   options[:write_nil_value] = value
+# end
+
+parser.separator(nil)
+parser.separator(
+  'Options only for CSV.filter: see https://docs.ruby-lang.org/en/3.3/CSV.html#method-c-filter'
+)
+
+parser.on('--input-col-sep=SEPARATOR',
+     'Input column separator string (aliased as --in-col-sep).') do |value|
+  options[:input_col_sep] = value
+end
+parser.on('--input-quote-char=CHAR',
+     'Input quote character (aliased as --in-quote-char.') do |value|
+  options[:input_quote_char] = value
+end
+parser.on('--input-row-sep=SEPARATOR',
+     'Input row separator string (aliased as --in-row-sep).'
+) do |value|
+  options[:input_row_sep] = value
+end
+parser.on('--output-col-sep=SEPARATOR',
+     'Output column separator string (aliased as --out-col-sep).'
+) do |value|
+  options[:output_col_sep] = value
+end
+parser.on('--output-quote-char=CHAR',
+     'Output quote character (aliased as --out-quote-char).') do |value|
+  options[:output_quote_char] = value
+end
+parser.on('--output-row-sep=SEPARATOR',
+     'Output row separator string (aliased as --out-row-sep).'
+) do |value|
+  options[:output_row_sep] = value
+end
+
+parser.separator(nil)
+parser.separator('Input/Output Options')
+
+parser.separator(nil)
+parser.on('-c', '--col-sep=SEPARATOR',
+     'Column separator string.') do |value|
+  options[:col_sep] = value
+end
+parser.on('-q', '--quote-char=CHAR',
+     'Quote character.') do |value|
+  options[:quote_char] = value
+end
+parser.on('-r', '--row-sep=SEPARATOR',
+     'Row separator string.'
+) do |value|
+  options[:row_sep] = value
+end
+
+parser.separator(nil)
+parser.separator('Generic Options')
+parser.separator(nil)
+
+parser.on('-h', '--help', 'Prints this help.') do
+  puts parser
+  exit
+end
+
+parser.on('-v', '--version', 'Prints version.') do
+  puts CSV::VERSION
+  exit
+end
+
+parser.parse!
+
+CSV.filter(**options) do |row|
+end