xlsx,csv: trim whitespace, yield to optional block

Closes #108 and #79
This commit is contained in:
Andy Meneely
2015-10-13 10:37:49 -04:00
parent 521bfac6b2
commit c8e6f9803c
9 changed files with 317 additions and 15 deletions
+19 -4
View File
@@ -1,6 +1,7 @@
require 'roo'
require 'csv'
require 'squib/args/input_file'
require 'squib/args/import'
module Squib
@@ -16,24 +17,30 @@ module Squib
# # | 1 | 2 |
# # | 3 | 4 |
# data = xlsx file: 'data.xlsx', sheet: 0
# {'h1' => [1,3], 'h2' => [2,4]}
# => {'h1' => [1,3], 'h2' => [2,4]}
#
# @option opts file [String] the file to open. Must end in `.xlsx`. Opens relative to the current directory.
# @option opts sheet [Integer] (0) The zero-based index of the sheet from which to read.
# @option opts strip [Boolean] (true) When true, strips leading and trailing whitespace on values and headers
# @option opts qty_header [String] ('qty']) Quantity explosion will be applied to the column this name
# @return [Hash] a hash of arrays based on columns in the spreadsheet
# @api public
def xlsx(opts = {})
input = Args::InputFile.new(file: 'deck.xlsx').load!(opts)
import = Args::Import.new.load!(opts)
s = Roo::Excelx.new(input.file[0])
s.default_sheet = s.sheets[input.sheet[0]]
data = {}
s.first_column.upto(s.last_column) do |col|
header = s.cell(s.first_row,col).to_s
header.strip! if import.strip?
data[header] = []
(s.first_row + 1).upto(s.last_row) do |row|
cell = s.cell(row,col)
# Roo hack for avoiding unnecessary .0's on whole integers (https://github.com/roo-rb/roo/issues/139)
cell = s.excelx_value(row,col) if s.excelx_type(row,col) == [:numeric_or_formula, 'General']
cell.strip! if cell.respond_to?(:strip) && import.strip?
cell = yield(header, cell) if block_given?
data[header] << cell
end#row
end#col
@@ -58,17 +65,25 @@ module Squib
# http://www.ruby-doc.org/stdlib-2.0/libdoc/csv/rdoc/CSV.html
#
# @option opts file [String] the CSV-formatted file to open. Opens relative to the current directory.
# @option opts strip [Boolean] (true) When true, strips leading and trailing whitespace on values and headers
# @option opts qty_header [String] ('qty']) Quantity explosion will be applied to the column this name
# @return [Hash] a hash of arrays based on columns in the table
# @api public
def csv(opts = {})
file = Args::InputFile.new(file: 'deck.csv').load!(opts).file[0]
opts = Squib::SYSTEM_DEFAULTS.merge(opts)
# opts = Squib::InputHelpers.fileify(opts)
import = Args::Import.new.load!(opts)
table = CSV.read(file, headers: true, converters: :numeric)
check_duplicate_csv_headers(table)
hash = Hash.new
table.headers.each do |header|
hash[header.to_s] ||= table[header]
new_header = header.to_s
new_header.strip! if import.strip?
hash[new_header] ||= table[header]
end
if import.strip?
hash.each do |header, col|
col.map! { |str| str.strip! if str.respond_to?(:strip); str }
end
end
return hash
end
+34
View File
@@ -0,0 +1,34 @@
require 'squib/args/arg_loader'
module Squib
# @api private
module Args
class Import
include ArgLoader
def self.parameters
{ strip: true }
end
def self.expanding_parameters
[] # none of them
end
def self.params_with_units
[] # none of them
end
def validate_strip(arg)
raise 'Strip must be true or false' unless arg == true || arg == false
arg
end
def strip?
strip
end
end
end
end