csv-utils 0.3.0 → 0.3.5

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 98303ab9b2df05bc501c1c66b66a62be5ade9d79ab38a5b8bda8eb52d91b26cc
-  data.tar.gz: 8adfd2144220de2cc4f23136ee4eb7314a3c16eeac68be87e1dc19b1ac7dc350
+  metadata.gz: fc87e9b60715e648c97c91c1bbe562eb72a6fd964782abf3f1468588c053537b
+  data.tar.gz: c368dda5d1b829a6dff9ac934e42b716c8c704d3a4605f8c4b195d1a10bcc94e
 SHA512:
-  metadata.gz: a2a2b2067a9ca06920b171230a122eba479c1f91af3919e2965eaec6d073fff34d544221a92cffaa1b9546078960aee0c9b9031e7b652368e975cff9b196214c
-  data.tar.gz: '0786cfb3e75771ccb68bfa0e2cba42994c7c04a5c8be14432ae6467425536e7dfd4a4ef33403ae5bd129eafd871a07077610c000a78762052e0b055192c0cc16'
+  metadata.gz: f8ead05ff26fcb2540403bcb47e5eacad33fcacf58f65ff48442406d0556f6dc5a23164258303a779a718ebc6bd8d232f6bad93d20d49275da05988bb0cc39c1
+  data.tar.gz: f5c9d16c15fc3af24947cb1aeb122962f26483cfba282a6676b1a47c8ecc9504952785898a0c94c771520b60fc3fbc84f3453c76ba9ebd896f0e19e546b23f69

data/bin/csv-explorer

@@ -0,0 +1,12 @@
+#!/usr/bin/env ruby
+
+require 'csv-utils'
+require 'irb'
+
+CSV_FILE = ARGV.shift
+
+def csv
+  @csv ||= CSVUtils::CSVIterator.new(CSV_FILE)
+end
+
+IRB.start

data/bin/csv-validator

@@ -0,0 +1,77 @@
+#!/usr/bin/env ruby
+
+require 'csv'
+begin
+  require 'rchardet'
+rescue LoadError
+  $stderr.puts 'gem install rchardet'
+  exit 1
+end
+
+def utf8?(str)
+  str
+    .force_encoding('utf-8')
+    .valid_encoding?
+end
+
+def convert_to_utf8(str, current_encoding)
+  str.force_encoding(current_encoding)
+  return nil unless str.valid_encoding?
+
+  str.encode('utf-8')
+end
+
+def detect_encoding(col)
+  CharDet.detect(col)['encoding']
+end
+
+def strip_bom!(col)
+  col.sub!("\xEF\xBB\xBF".force_encoding('ASCII-8BIT'), '')
+end
+
+csv = CSV.open(ARGV[0], 'rb')
+id_column_name = ARGV[1]
+
+headers = csv.shift
+strip_bom!(headers[0])
+
+id_column_num = nil
+if id_column_name
+  unless headers.include?(id_column_name)
+    $stderr.puts("header #{id_column_name} not found in current set of headers")
+    exit 1
+  end
+
+  id_column_num = headers.index(id_column_name)
+end
+
+out = nil
+if id_column_num
+  out = CSV.open('utf8-correctsion.csv', 'wb')
+  out << [id_column_name, 'Row', 'Col', 'Header', 'Value']
+end
+
+csv_lineno = 1
+
+while (row = csv.shift)
+  csv_lineno += 1
+
+  unless row.size == headers.size
+    $stderr.puts "row(#{csv_lineno}): invalid number of columns, expected #{headers.size} got #{row.size}"
+  end
+
+  row.each_with_index do |col, idx|
+    next if utf8?(col)
+
+    $stderr.puts "row(#{csv_lineno}),col(#{idx + 1}) #{headers[idx]}: none UTF-8 characters found in \"#{col}\""
+    if (col_utf8_encoded = convert_to_utf8(col, detect_encoding(col)))
+      puts "row(#{csv_lineno}),col(#{idx + 1}) #{headers[idx]}: converted to UTF-8 from #{detect_encoding(col)} \"#{col_utf8_encoded}\""
+      out << [row[id_column_num], csv_lineno, (idx + 1), headers[idx], col_utf8_encoded]
+    else
+      $stderr.puts "row(#{csv_lineno}),col(#{idx + 1}) #{headers[idx]}: unknown character encoding"
+    end
+  end
+end
+
+csv.close
+out.close if out

data/csv-utils.gemspec

@@ -2,7 +2,7 @@
 
 Gem::Specification.new do |s|
   s.name        = 'csv-utils'
-  s.version     = '0.3.0'
+  s.version     = '0.3.5'
   s.licenses    = ['MIT']
   s.summary     = 'CSV Utils'
   s.description = 'Tools for debugging malformed CSV files'

data/lib/csv-utils.rb

@@ -3,6 +3,7 @@ require 'csv'
 # Collection of tools for working with CSV files.
 module CSVUtils
   autoload :CSVExtender, 'csv_utils/csv_extender'
+  autoload :CSVIterator, 'csv_utils/csv_iterator'
   autoload :CSVOptions, 'csv_utils/csv_options'
   autoload :CSVReport, 'csv_utils/csv_report'
   autoload :CSVRow, 'csv_utils/csv_row'

data/lib/csv_utils/csv_iterator.rb

@@ -0,0 +1,47 @@
+# Search a CSV given a series of steps
+class CSVUtils::CSVIterator
+  include Enumerable
+
+  class RowWrapper < Hash
+    attr_accessor :lineno
+
+    def self.create(headers, row, lineno)
+      row_wrapper = RowWrapper[headers.zip(row)]
+      row_wrapper.lineno = lineno
+      row_wrapper
+    end
+
+    def to_pretty_s
+      reject { |_, v| v.strip.empty? }
+        .each_with_index
+        .map { |(k, v), idx| sprintf('  %-3d %s: %s', idx+1, k, v) }
+        .join("\n") + "\n"
+    end
+  end
+
+  def initialize(src_csv, csv_options = {})
+    @src_csv = CSVUtils::CSVWrapper.new(src_csv, 'rb', csv_options)
+  end
+
+  def each(headers = nil)
+    @src_csv.rewind
+
+    lineno = 0
+    unless headers
+      headers = @src_csv.shift
+      strip_bom!(headers[0])
+      lineno += 1
+    end
+
+    while (row = @src_csv.shift)
+      lineno += 1
+      yield RowWrapper.create(headers, row, lineno)
+    end
+  end
+
+  private
+
+  def strip_bom!(col)
+    col.sub!("\xEF\xBB\xBF".force_encoding('ASCII-8BIT'), '')
+  end
+end

data/lib/csv_utils/csv_wrapper.rb

@@ -35,6 +35,10 @@ class CSVUtils::CSVWrapper
     csv.shift
   end
 
+  def rewind
+    csv.rewind
+  end
+
   def close
     csv.close if close_when_done?
   end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: csv-utils
 version: !ruby/object:Gem::Version
-  version: 0.3.0
+  version: 0.3.5
 platform: ruby
 authors:
 - Doug Youch
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2020-07-19 00:00:00.000000000 Z
+date: 2020-07-28 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: inheritance-helper
@@ -28,8 +28,10 @@ description: Tools for debugging malformed CSV files
 email: dougyouch@gmail.com
 executables:
 - csv-change-eol
+- csv-explorer
 - csv-find-error
 - csv-readline
+- csv-validator
 extensions: []
 extra_rdoc_files: []
 files:
@@ -41,11 +43,14 @@ files:
 - LICENSE
 - README.md
 - bin/csv-change-eol
+- bin/csv-explorer
 - bin/csv-find-error
 - bin/csv-readline
+- bin/csv-validator
 - csv-utils.gemspec
 - lib/csv-utils.rb
 - lib/csv_utils/csv_extender.rb
+- lib/csv_utils/csv_iterator.rb
 - lib/csv_utils/csv_options.rb
 - lib/csv_utils/csv_report.rb
 - lib/csv_utils/csv_row.rb