Fortune quadrophobiac

## index.html
<body>
  <div id="container" class="container">
      walla walla glub glub
  </div>
</body>


## discussing OO principles with ODI engineers
@pikesley: @pezholio  I'm aware this may not arise in the time remaining on the internship period, but if you do decide to write a gem for something-something-ODI could you alert us?

Sam [3:25 PM]
I can show you build-a-gem-101 in like an hour

Sam [3:25 PM]
Up to and including putting it through travis and publishing to rubygems

stephenfortune [3:25 PM]
I'd love to see the travis part, but I'm interested in being a fly on the wall to the rationale of creation which was discussed earlier

## gist:f5e9f8a22de3f6dde48e
To clarify if non-blocking fashion means making the CSV parsing CPU bound rather than IO Bound, and to check what refactoring is desired (I had previously moved the parse_csv methods to a separate file, not sure if this is required or not, and if it is required it will require rewriting a lot of tests)

stephenfortune [2:13 PM]
I think that the existing code already has the CSV parsing as CPU bound but it would be helpful to clarify

James Smith [2:14 PM]
in my mind it’s more that no matter how the parsing is bound (CPU or IO), you can’t tell what it’s doing halfway through - and that’s what we need to be able to do.

James Smith [2:14 PM]
if that makes sense. And if it is in fact correct. Might not be

## csvlint_validation_exception_spec.rb
  context "error matching tests" do

    it "checks for blank rows" do
      # stream = StringIO.new('"","",') # this returns blank_rows
      # stream = StringIO.new('"","",\r\n"","",\r\n') # this returns :whitespace
      # stream = StringIO.new('"a",“b“,"c"""') # should return stray quotes but doesn't - there isn't a stray quote test in current code base
      # stream = StringIO.new("\"a,\"b\",\"c\"\n") # returns unclosed quotes
      validator = Csvlint::Validator.new(stream,"header" => false) # dialect options to avoid one liners being read as header

      expect(validator.valid?).to eql(false)

## validate.rb
if type == :stray_quote && !wrapper.line.match(csv.row_sep)
  # above evaluates to true && false
  # however it is false only by placing bang against wrapper.line.match(csv.row_sep) which evals to either
  # nil or MatchData - seems a strange way to order flow
  # basically it want's whatever is matched against the row_sep (generated from the CSV.new of data inputted) to not match
  # because if it matches then bang generates a false and only a stray_quote is logged
 build_errors(:line_breaks, :structure)

 # the tests that exist pass \r rather than \r\n when testing for line breaks

## enumerator_pseudocode.rb
# ref - http://blog.honeybadger.io/using-lazy-enumerators-to-work-with-large-files-in-ruby/
# The Lazy Enumerator class contains a number of built in collection and mapping methods
# e.g. ```map, flat_map, select, reject, grep, zip, take, take_while, drop, drop_while, cycle```
# the basic way that lazy evaluation works with the above methods is like so
enum.lazy.select {|variable| (conditional_eval_of_variable) } # returns an array where the eval returns true
# in pseudo code fashion I would imagine it would proceed like this
CSVlintValidator.select { |validation| output(@ErrorCollector) if !validation.valid? } # returns object or array for the rows which raise messages
# validation.valid? could be rewritten in Csvlint::ErrorCollector to be errors.empty? && warnings.empty?
# ~ && info_messages.empty? or the output could only happen when an error arises, informing the user that there is already
# one error detected and do they wish to continue evaluating in 'lazy' fashion

## gist:32fc608b5e8a8c23b0d8
git clone https://github.com/theodi/chef-csvlint.git
cd chef-csvlint/
rbenv local 2.1.6
gem install bundler -v 1.10.5
bundle exec kitchen converge

to sanity check your local env
which ruby # should return /Users/$USERNAME/.rbenv/shims/ruby
which gem # should return /Users/$USERNAME/.rbenv/shims/gem
ruby -v # returns ruby version

## datapackage.js
var exportdata = function() {
  var window = BrowserWindow.getFocusedWindow();

  datapackage = new BrowserWindow({width: 450, height: 600, 'always-on-top': true});
  datapackage.loadUrl('file://' + __dirname + '/../comma-chameleon/views/datapackage.html');

  datapackage.on('closed', function() {
    datapackage = null;
  });

## hot.js
 var hot = new Handsontable(container, {
   colHeaders: true,
-  rowHeaders: true,
+  rowHeaders: function(index) {
+    if (index === 0){
+      return "Header";
+    } else {
+      return index;
+    }
+  },

## validation.js
// line 54-86, feature-prebaked
// inclusion of printErrs is because type returns as undefined
function displayValidationMessages(validation) {
  console.log(JSON.stringify(validation,null, 2));
  var $messagePanel = $('#message-panel');
  $messagePanel.html("<h4>Validation results <img src='" + validation.badges.png  +"' /></h4>");
  var resultsTemplate = _.template('<p><%= validation.errors.length %> errors, <%= validation.warnings.length %> warnings and <%= validation.info.length %> info messages:</p>')
  $messagePanel.append(resultsTemplate({'validation': validation}));
//TODO the errorText function below is truncated and isn't doing what it should be doing
  var printErrs = validation.errors[0];
	<body>
	<div id="container" class="container">
	walla walla glub glub
	</div>
	</body>
	@pikesley: @pezholio I'm aware this may not arise in the time remaining on the internship period, but if you do decide to write a gem for something-something-ODI could you alert us?

	Sam [3:25 PM]
	I can show you build-a-gem-101 in like an hour

	Sam [3:25 PM]
	Up to and including putting it through travis and publishing to rubygems

	stephenfortune [3:25 PM]
	I'd love to see the travis part, but I'm interested in being a fly on the wall to the rationale of creation which was discussed earlier
	To clarify if non-blocking fashion means making the CSV parsing CPU bound rather than IO Bound, and to check what refactoring is desired (I had previously moved the parse_csv methods to a separate file, not sure if this is required or not, and if it is required it will require rewriting a lot of tests)

	stephenfortune [2:13 PM]
	I think that the existing code already has the CSV parsing as CPU bound but it would be helpful to clarify

	James Smith [2:14 PM]
	in my mind it’s more that no matter how the parsing is bound (CPU or IO), you can’t tell what it’s doing halfway through - and that’s what we need to be able to do.

	James Smith [2:14 PM]
	if that makes sense. And if it is in fact correct. Might not be
	context "error matching tests" do

	it "checks for blank rows" do
	# stream = StringIO.new('"","",') # this returns blank_rows
	# stream = StringIO.new('"","",\r\n"","",\r\n') # this returns :whitespace
	# stream = StringIO.new('"a",“b“,"c"""') # should return stray quotes but doesn't - there isn't a stray quote test in current code base
	# stream = StringIO.new("\"a,\"b\",\"c\"\n") # returns unclosed quotes
	validator = Csvlint::Validator.new(stream,"header" => false) # dialect options to avoid one liners being read as header

	expect(validator.valid?).to eql(false)
	if type == :stray_quote && !wrapper.line.match(csv.row_sep)
	# above evaluates to true && false
	# however it is false only by placing bang against wrapper.line.match(csv.row_sep) which evals to either
	# nil or MatchData - seems a strange way to order flow
	# basically it want's whatever is matched against the row_sep (generated from the CSV.new of data inputted) to not match
	# because if it matches then bang generates a false and only a stray_quote is logged
	build_errors(:line_breaks, :structure)

	# the tests that exist pass \r rather than \r\n when testing for line breaks
	# ref - http://blog.honeybadger.io/using-lazy-enumerators-to-work-with-large-files-in-ruby/
	# The Lazy Enumerator class contains a number of built in collection and mapping methods
	# e.g. ```map, flat_map, select, reject, grep, zip, take, take_while, drop, drop_while, cycle```
	# the basic way that lazy evaluation works with the above methods is like so
	enum.lazy.select {\|variable\| (conditional_eval_of_variable) } # returns an array where the eval returns true
	# in pseudo code fashion I would imagine it would proceed like this
	CSVlintValidator.select { \|validation\| output(@ErrorCollector) if !validation.valid? } # returns object or array for the rows which raise messages
	# validation.valid? could be rewritten in Csvlint::ErrorCollector to be errors.empty? && warnings.empty?
	# ~ && info_messages.empty? or the output could only happen when an error arises, informing the user that there is already
	# one error detected and do they wish to continue evaluating in 'lazy' fashion
	git clone https://github.com/theodi/chef-csvlint.git
	cd chef-csvlint/
	rbenv local 2.1.6
	gem install bundler -v 1.10.5
	bundle exec kitchen converge

	to sanity check your local env
	which ruby # should return /Users/$USERNAME/.rbenv/shims/ruby
	which gem # should return /Users/$USERNAME/.rbenv/shims/gem
	ruby -v # returns ruby version
	var exportdata = function() {
	var window = BrowserWindow.getFocusedWindow();

	datapackage = new BrowserWindow({width: 450, height: 600, 'always-on-top': true});
	datapackage.loadUrl('file://' + __dirname + '/../comma-chameleon/views/datapackage.html');

	datapackage.on('closed', function() {
	datapackage = null;
	});
	var hot = new Handsontable(container, {
	colHeaders: true,
	- rowHeaders: true,
	+ rowHeaders: function(index) {
	+ if (index === 0){
	+ return "Header";
	+ } else {
	+ return index;
	+ }
	+ },
	// line 54-86, feature-prebaked
	// inclusion of printErrs is because type returns as undefined
	function displayValidationMessages(validation) {
	console.log(JSON.stringify(validation,null, 2));
	var $messagePanel = $('#message-panel');
	$messagePanel.html("<h4>Validation results <img src='" + validation.badges.png +"' /></h4>");
	var resultsTemplate = _.template('<p><%= validation.errors.length %> errors, <%= validation.warnings.length %> warnings and <%= validation.info.length %> info messages:</p>')
	$messagePanel.append(resultsTemplate({'validation': validation}));
	//TODO the errorText function below is truncated and isn't doing what it should be doing
	var printErrs = validation.errors[0];