Fluentd parser plugin pattern not matched with data - ruby

Hi I'm writing a custom plugin for fluentD, I need to use this plugin to parse one field from syslog5424 format, the ruby code works just fine when you run it without fluentd but not when it runs as a plugin
error
#<Fluent::Plugin::Parser::ParserError: pattern not matched with data '[kubernetes#47450 app="api-mma" pod-template-hash="9849d88d6"
namespace_name="qa" object_name="api-mma-9849d88d6-59mzd"
container_name="api-mma"
vm_id="91cdf0e0-b1fb-45bc-8a3f-c85b027af505"]'>
The custom plugin:
require 'fluent/plugin/parser'
require 'json'
module Fluent::Plugin
class TKGIMetadataParser < Parser
# Register this parser as 'tkgi_metadata'
Fluent::Plugin.register_parser('tkgi_metadata', self)
# `delimiter` is configurable with ' ' as default
config_param :delimiter, :string, default: ' '
def configure(conf)
super
if #delimiter.length != 1
raise ConfigError, "delimiter must be a single character. #{#delimiter} is not."
end
end
def parse(text)
left_braket_pos = text.index('['.freeze)
right_braket_pos = text.index(']'.freeze)
data = text.slice(left_braket_pos + 1, right_braket_pos - 1)
source, key_values = data.split(' ', 2)
record = {}
#key_values.gsub!(/[\"]/,'')
key_values.split(' ').each do |kv|
k, v = kv.split('=', 2)
record[k] = v
end
record.merge!(source: source)
yield record.to_json
end
end
end
fluentd.conf
<filter k8s>
#type parser
key_name syslog5424_sd
reserve_data true
reserve_time true
<parse>
#type tkgi_metadata
</parse>
</filter>
sample log:
[kubernetes#47450 app="api-mma" pod-template-hash="9849d88d6" namespace_name="qa" object_name="api-mma-9849d88d6-59mzd" container_name="api-mma" vm_id="91cdf0e0-b1fb-45bc-8a3f-c85b027af505"]

You need to yield time and a record as a hash:
def parse(text)
# ...
yield convert_values(parse_time(record), record)
end
Example: https://github.com/fluent/fluentd/blob/master/lib/fluent/plugin/parser_ltsv.rb#L46-L47

I discovered that you need to pass two parameters two yield in order to this work, what mentioned in #Panic answer defiantly gave me a headstart but I end up doing
def parse(text)
# ...
yield nil, record
end

Related

How to "observe" a stream in Ruby's CSV module?

I am writing a class that takes a CSV files, transforms it, and then writes the new data out.
module Transformer
class Base
def initialize(file)
#file = file
end
def original_data(&block)
opts = { headers: true }
CSV.open(file, 'rb', opts, &block)
end
def transformer
# complex manipulations here like modifying columns, picking only certain
# columns to put into new_data, etc but simplified to `+10` to keep
# example concise
-> { |row| new_data << row['some_header'] + 10 }
end
def transformed_data
self.original_data(self.transformer)
end
def write_new_data
CSV.open('new_file.csv', 'wb', opts) do |new_data|
transformed_data
end
end
end
end
What I'd like to be able to do is:
Look at the transformed data without writing it out (so I can test that it transforms the data correctly, and I don't need to write it to file right away: maybe I want to do more manipulation before writing it out)
Don't slurp all the file at once, so it works no matter the size of the original data
Have this as a base class with an empty transformer so that instances only need to implement their own transformers but the behavior for reading and writing is given by the base class.
But obviously the above doesn't work because I don't really have a reference to new_data in transformer.
How could I achieve this elegantly?
I can recommend one of two approaches, depending on your needs and personal taste.
I have intentionally distilled the code to just its bare minimum (without your wrapping class), for clarity.
1. Simple read-modify-write loop
Since you do not want to slurp the file, use CSV::Foreach. For example, for a quick debugging session, do:
CSV.foreach "source.csv", headers: true do |row|
row["name"] = row["name"].upcase
row["new column"] = "new value"
p row
end
And if you wish to write to file during that same iteration:
require 'csv'
csv_options = { headers: true }
# Open the target file for writing
CSV.open("target.csv", "wb") do |target|
# Add a header
target << %w[new header column names]
# Iterate over the source CSV rows
CSV.foreach "source.csv", **csv_options do |row|
# Mutate and add columns
row["name"] = row["name"].upcase
row["new column"] = "new value"
# Push the new row to the target file
target << row
end
end
2. Using CSV::Converters
There is a built in functionality that might be helpful - CSV::Converters - (see the :converters definition in the CSV::New documentation)
require 'csv'
# Register a converter in the options hash
csv_options = { headers: true, converters: [:stripper] }
# Define a converter
CSV::Converters[:stripper] = lambda do |value, field|
value ? value.to_s.strip : value
end
CSV.open("target.csv", "wb") do |target|
# same as above
CSV.foreach "source.csv", **csv_options do |row|
# same as above - input data will already be converted
# you can do additional things here if needed
end
end
3. Separate input and output from your converter classes
Based on your comment, and since you want to minimize I/O and iterations, perhaps extracting the read/write operations from the responsibility of the transformers might be of interest. Something like this.
require 'csv'
class NameCapitalizer
def self.call(row)
row["name"] = row["name"].upcase
end
end
class EmailRemover
def self.call(row)
row.delete 'email'
end
end
csv_options = { headers: true }
converters = [NameCapitalizer, EmailRemover]
CSV.open("target.csv", "wb") do |target|
CSV.foreach "source.csv", **csv_options do |row|
converters.each { |c| c.call row }
target << row
end
end
Note that the above code still does not handle the header, in case it was changed. You will probably have to reserve the last row (after all transformations) and prepend its #headers to the output CSV.
There are probably plenty other ways to do it, but the CSV class in Ruby does not have the cleanest interface, so I try to keep code that deals with it as simple as I can.

How to read multiple XML files then output to multiple CSV files with the same XML filenames

I am trying to parse multiple XML files then output them into CSV files to list out the proper rows and columns.
I was able to do so by processing one file at a time by defining the filename, and specifically output them into a defined output file name:
File.open('H:/output/xmloutput.csv','w')
I would like to write into multiple files and make their name the same as the XML filenames without hard coding it. I tried doing it multiple ways but have had no luck so far.
Sample XML:
<?xml version="1.0" encoding="UTF-8"?>
<record:root>
<record:Dataload_Request>
<record:name>Bob Chuck</record:name>
<record:Address_Data>
<record:Street_Address>123 Main St</record:Street_Address>
<record:Postal_Code>12345</record:Postal_Code>
</record:Address_Data>
<record:Age>45</record:Age>
</record:Dataload_Request>
</record:root>
Here is what I've tried:
require 'nokogiri'
require 'set'
files = ''
input_folder = "H:/input"
output_folder = "H:/output"
if input_folder[input_folder.length-1,1] == '/'
input_folder = input_folder[0,input_folder.length-1]
end
if output_folder[output_folder.length-1,1] != '/'
output_folder = output_folder + '/'
end
files = Dir[input_folder + '/*.xml'].sort_by{ |f| File.mtime(f)}
file = File.read(input_folder + '/' + files)
doc = Nokogiri::XML(file)
record = {} # hashes
keys = Set.new
records = [] # array
csv = ""
doc.traverse do |node|
value = node.text.gsub(/\n +/, '')
if node.name != "text" # skip these nodes: if class isnt text then skip
if value.length > 0 # skip empty nodes
key = node.name.gsub(/wd:/,'').to_sym
if key == :Dataload_Request && !record.empty?
records << record
record = {}
elsif key[/^root$|^document$/]
# neglect these keys
else
key = node.name.gsub(/wd:/,'').to_sym
# in case our value is html instead of text
record[key] = Nokogiri::HTML.parse(value).text
# add to our key set only if not already in the set
keys << key
end
end
end
end
# build our csv
File.open('H:/output/.*csv', 'w') do |file|
file.puts %Q{"#{keys.to_a.join('","')}"}
records.each do |record|
keys.each do |key|
file.write %Q{"#{record[key]}",}
end
file.write "\n"
end
print ''
print 'output files ready!'
print ''
end
I have been getting 'read memory': no implicit conversion of Array into String (TypeError) and other errors.
Here's a quick peer-review of your code, something like you'd get in a corporate environment...
Instead of writing:
input_folder = "H:/input"
input_folder[input_folder.length-1,1] == '/' # => false
Consider doing it using the -1 offset from the end of the string to access the character:
input_folder[-1] # => "t"
That simplifies your logic making it more readable because it's lacking unnecessary visual noise:
input_folder[-1] == '/' # => false
See [] and []= in the String documentation.
This looks like a bug to me:
files = Dir[input_folder + '/*.xml'].sort_by{ |f| File.mtime(f)}
file = File.read(input_folder + '/' + files)
files is an array of filenames. input_folder + '/' + files is appending an array to a string:
foo = ['1', '2'] # => ["1", "2"]
'/parent/' + foo # =>
# ~> -:9:in `+': no implicit conversion of Array into String (TypeError)
# ~> from -:9:in `<main>'
How you want to deal with that is left as an exercise for the programmer.
doc.traverse do |node|
is icky because it sidesteps the power of Nokogiri being able to search for a particular tag using accessors. Very rarely do we need to iterate over a document tag by tag, usually only when we're peeking at its structure and layout. traverse is slower so use it as a very last resort.
length is nice but isn't needed when checking whether a string has content:
value = 'foo'
value.length > 0 # => true
value > '' # => true
value = ''
value.length > 0 # => false
value > '' # => false
Programmers coming from Java like to use the accessors but I like being lazy, probably because of my C and Perl backgrounds.
Be careful with sub and gsub as they don't do what you're thinking they do. Both expect a regular expression, but will take a string which they do a escape on before beginning their scan.
You're passing in a regular expression, which is OK in this case, but it could cause unexpected problems if you don't remember all the rules for pattern matching and that gsub scans until the end of the string:
foo = 'wd:barwd:' # => "wd:barwd:"
key = foo.gsub(/wd:/,'') # => "bar"
In general I recommend people think a couple times before using regular expressions. I've seen some gaping holes opened up in logic written by fairly advanced programmers because they didn't know what the engine was going to do. They're wonderfully powerful, but need to be used surgically, not as a universal solution.
The same thing happens with a string, because gsub doesn't know when to quit:
key = foo.gsub('wd:','') # => "bar"
So, if you're looking to change just the first instance use sub:
key = foo.sub('wd:','') # => "barwd:"
I'd do it a little differently though.
foo = 'wd:bar'
I can check to see what the first three characters are:
foo[0,3] # => "wd:"
Or I can replace them with something else using string indexing:
foo[0,3] = ''
foo # => "bar"
There's more but I think that's enough for now.
You should use Ruby's CSV class. Also, you don't need to do any string matching or regex stuff. Use Nokogiri to target elements. If you know the node names in the XML will be consistent it should be pretty simple. I'm not exactly sure if this is the output you want, but this should get you in the right direction:
require 'nokogiri'
require 'csv'
def xml_to_csv(filename)
xml_str = File.read(filename)
xml_str.gsub!('record:','') # remove the record: namespace
doc = Nokogiri::XML xml_str
csv_filename = filename.gsub('.xml', '.csv')
CSV.open(csv_filename, 'wb' ) do |row|
row << ['name', 'street_address', 'postal_code', 'age']
row << [
doc.xpath('//name').text,
doc.xpath('//Street_Address').text,
doc.xpath('//Postal_Code').text,
doc.xpath('//Age').text,
]
end
end
# iterate over all xml files
Dir.glob('*.xml').each { |filename| xml_to_csv(filename) }

Type Checking does not work with Ruby and Liquid

I am currently making my blog with Jekyll. Jekyll custom plugin uses Ruby and Jekyll is using Liquid. I am currently getting input via custom liquid tag and processing it there. I would like to check whether the string contains integers or not. So I have the following code. I realised that the input is not of the type String but rather Jekyll::Token type. So I changed the inputs to string but I cannot detect whether string contains an integer. Here is my code:
module Jekyll
class TypecheckTag < Liquid::Tag
def is_int(word)
return word.count("0-3000") > 0
end
def initialize(tag_name, word, tokens)
super
#word = word.to_s
end
def render(context)
if /\A\d+\z/.match(#word)
#result = 'int'
else
#result = 'string'
end
end
end
end
Liquid::Template.register_tag('typecheck', Jekyll::TypecheckTag)
Unfortunately, it always returns 'string' say even if we have a string "16" for example.
The following code work fine
def render(input_str)
if /\A[-+]?\d+\z/.match(input_str.strip)
#result = 'int'
else
#result = 'string'
end
end
Sample input with output
p render('16') # "int"
p render('16a') # "string"
p render('16 ') # "int"
p render('asd') #"string"
p render('') #"string"
I have used strip because of render('16 ') # "int" If you don't want ignore that. Hope it will help you.

Compare REXML elements for name/attribute equality in RSpec

Is there a matcher for comparing REXML elements for logical equality in RSpec? I tried writing a custom matcher that converts them to formatted strings, but it fails if the attribute order is different. (As noted in the XML spec, the order of attributes should not be significant.)
I could grind through writing a custom matcher that compares the name, namespace, child nodes, attributes, etc., etc., but this seems time-consuming and error-prone, and if someone else has already done it I'd rather not reinvent the wheel.
I ended up using the equivalent-xml gem and writing an RSpec custom matcher to convert the REXML to Nokogiri, compare with equivalent-xml, and pretty-print the result if needed.
The test assertion is pretty simple:
expect(actual).to be_xml(expected)
or
expect(actual).to be_xml(expected, path)
if you want to display the file path or some sort of identifier (e.g. if you're comparing a lot of documents).
The match code is a little fancier than it needs to be because it handles REXML, Nokogiri, and strings.
module XMLMatchUtils
def self.to_nokogiri(xml)
return nil unless xml
case xml
when Nokogiri::XML::Element
xml
when Nokogiri::XML::Document
xml.root
when String
to_nokogiri(Nokogiri::XML(xml, &:noblanks))
when REXML::Element
to_nokogiri(xml.to_s)
else
raise "be_xml() expected XML, got #{xml.class}"
end
end
def self.to_pretty(nokogiri)
return nil unless nokogiri
out = StringIO.new
save_options = Nokogiri::XML::Node::SaveOptions::FORMAT | Nokogiri::XML::Node::SaveOptions::NO_DECLARATION
nokogiri.write_xml_to(out, encoding: 'UTF-8', indent: 2, save_with: save_options)
out.string
end
def self.equivalent?(expected, actual, filename = nil)
expected_xml = to_nokogiri(expected) || raise("expected value #{expected || 'nil'} does not appear to be XML#{" in #{filename}" if filename}")
actual_xml = to_nokogiri(actual)
EquivalentXml.equivalent?(expected_xml, actual_xml, element_order: false, normalize_whitespace: true)
end
def self.failure_message(expected, actual, filename = nil)
expected_string = to_pretty(to_nokogiri(expected))
actual_string = to_pretty(to_nokogiri(actual)) || actual
# Uncomment this to dump expected/actual to file for manual diffing
#
# now = Time.now.to_i
# FileUtils.mkdir('tmp') unless File.directory?('tmp')
# File.open("tmp/#{now}-expected.xml", 'w') { |f| f.write(expected_string) }
# File.open("tmp/#{now}-actual.xml", 'w') { |f| f.write(actual_string) }
diff = Diffy::Diff.new(expected_string, actual_string).to_s(:text)
"expected XML differs from actual#{" in #{filename}" if filename}:\n#{diff}"
end
def self.to_xml_string(actual)
to_pretty(to_nokogiri(actual))
end
def self.failure_message_when_negated(actual, filename = nil)
"expected not to get XML#{" in #{filename}" if filename}:\n\t#{to_xml_string(actual) || 'nil'}"
end
end
The actual matcher is fairly straightforward:
RSpec::Matchers.define :be_xml do |expected, filename = nil|
match do |actual|
XMLMatchUtils.equivalent?(expected, actual, filename)
end
failure_message do |actual|
XMLMatchUtils.failure_message(expected, actual, filename)
end
failure_message_when_negated do |actual|
XMLMatchUtils.failure_message_when_negated(actual, filename)
end
end

How do I force one field in Ruby's CSV output to be wrapped with double-quotes?

I'm generating some CSV output using Ruby's built-in CSV. Everything works fine, but the customer wants the name field in the output to have wrapping double-quotes so the output looks like the input file. For instance, the input looks something like this:
1,1.1.1.1,"Firstname Lastname",more,fields
2,2.2.2.2,"Firstname Lastname, Jr.",more,fields
CSV's output, which is correct, looks like:
1,1.1.1.1,Firstname Lastname,more,fields
2,2.2.2.2,"Firstname Lastname, Jr.",more,fields
I know CSV is doing the right thing by not double-quoting the third field just because it has embedded blanks, and wrapping the field with double-quotes when it has the embedded comma. What I'd like to do, to help the customer feel warm and fuzzy, is tell CSV to always double-quote the third field.
I tried wrapping the field in double-quotes in my to_a method, which creates a "Firstname Lastname" field being passed to CSV, but CSV laughed at my puny-human attempt and output """Firstname Lastname""". That is the correct thing to do because it's escaping the double-quotes, so that didn't work.
Then I tried setting CSV's :force_quotes => true in the open method, which output double-quotes wrapping all fields as expected, but the customer didn't like that, which I expected also. So, that didn't work either.
I've looked through the Table and Row docs and nothing appeared to give me access to the "generate a String field" method, or a way to set a "for field n always use quoting" flag.
I'm about to dive into the source to see if there's some super-secret tweaks, or if there's a way to monkey-patch CSV and bend it to do my will, but wondered if anyone had some special knowledge or had run into this before.
And, yes, I know I could roll my own CSV output, but I prefer to not reinvent well-tested wheels. And, I'm also aware of FasterCSV; That's now part of Ruby 1.9.2, which I'm using, so explicitly using FasterCSV buys me nothing special. Also, I'm not using Rails and have no intention of rewriting it in Rails, so unless you have a cute way of implementing it using a small subset of Rails, don't bother. I'll downvote any recommendations to use any of those ways just because you didn't bother to read this far.
Well, there's a way to do it but it wasn't as clean as I'd hoped the CSV code could allow.
I had to subclass CSV, then override the CSV::Row.<<= method and add another method forced_quote_fields= to make it possible to define the fields I want to force-quoting on, plus pull two lambdas from other methods. At least it works for what I want:
require 'csv'
class MyCSV < CSV
def <<(row)
# make sure headers have been assigned
if header_row? and [Array, String].include? #use_headers.class
parse_headers # won't read data for Array or String
self << #headers if #write_headers
end
# handle CSV::Row objects and Hashes
row = case row
when self.class::Row then row.fields
when Hash then #headers.map { |header| row[header] }
else row
end
#headers = row if header_row?
#lineno += 1
#do_quote ||= lambda do |field|
field = String(field)
encoded_quote = #quote_char.encode(field.encoding)
encoded_quote +
field.gsub(encoded_quote, encoded_quote * 2) +
encoded_quote
end
#quotable_chars ||= encode_str("\r\n", #col_sep, #quote_char)
#forced_quote_fields ||= []
#my_quote_lambda ||= lambda do |field, index|
if field.nil? # represent +nil+ fields as empty unquoted fields
""
else
field = String(field) # Stringify fields
# represent empty fields as empty quoted fields
if (
field.empty? or
field.count(#quotable_chars).nonzero? or
#forced_quote_fields.include?(index)
)
#do_quote.call(field)
else
field # unquoted field
end
end
end
output = row.map.with_index(&#my_quote_lambda).join(#col_sep) + #row_sep # quote and separate
if (
#io.is_a?(StringIO) and
output.encoding != raw_encoding and
(compatible_encoding = Encoding.compatible?(#io.string, output))
)
#io = StringIO.new(#io.string.force_encoding(compatible_encoding))
#io.seek(0, IO::SEEK_END)
end
#io << output
self # for chaining
end
alias_method :add_row, :<<
alias_method :puts, :<<
def forced_quote_fields=(indexes=[])
#forced_quote_fields = indexes
end
end
That's the code. Calling it:
data = [
%w[1 2 3],
[ 2, 'two too', 3 ],
[ 3, 'two, too', 3 ]
]
quote_fields = [1]
puts "Ruby version: #{ RUBY_VERSION }"
puts "Quoting fields: #{ quote_fields.join(', ') }", "\n"
csv = MyCSV.generate do |_csv|
_csv.forced_quote_fields = quote_fields
data.each do |d|
_csv << d
end
end
puts csv
results in:
# >> Ruby version: 1.9.2
# >> Quoting fields: 1
# >>
# >> 1,"2",3
# >> 2,"two too",3
# >> 3,"two, too",3
This post is old, but I can't believe no one thought of this.
Why not do:
csv = CSV.generate :quote_char => "\0" do |csv|
where \0 is a null character, then just add quotes to each field where they are needed:
csv << [product.upc, "\"" + product.name + "\"" # ...
Then at the end you can do a
csv.gsub!(/\0/, '')
I doubt if this will help the customer feeling warm and fuzzy after all this time, but this seems to work:
require 'csv'
#prepare a lambda which converts field with index 2
quote_col2 = lambda do |field, fieldinfo|
# fieldinfo has a line- ,header- and index-method
if fieldinfo.index == 2 && !field.start_with?('"') then
'"' + field + '"'
else
field
end
end
# specify above lambda as one of the converters
csv = CSV.read("test1.csv", :converters => [quote_col2])
p csv
# => [["aaa", "bbb", "\"ccc\"", "ddd"], ["fff", "ggg", "\"hhh\"", "iii"]]
File.open("test1.txt","w"){|out| csv.each{|line|out.puts line.join(",")}}
CSV has a force_quotes option that will force it to quote all fields (it may not have been there when you posted this originally). I realize this isn't exactly what you were proposing, but it's less monkey patching.
2.1.0 :008 > puts CSV.generate_line [1,'1.1.1.1','Firstname Lastname','more','fields']
1,1.1.1.1,Firstname Lastname,more,fields
2.1.0 :009 > puts CSV.generate_line [1,'1.1.1.1','Firstname Lastname','more','fields'], force_quotes: true
"1","1.1.1.1","Firstname Lastname","more","fields"
The drawback is that the first integer value ends up listed as a string, which changes things when you import into Excel.
It's been a long time, but since the CSV library has been patched, this might help someone if they're now facing this issue:
require 'csv'
# puts CSV::VERSION # this should be 3.1.9+
headers = ['id', 'ip', 'name', 'foo', 'bar']
data = [
[1, '1.1.1.1','Firstname Lastname','more','fields'],
[2, '2.2.2.2','Firstname Lastname, Jr.','more','fields']
]
quoter = Proc.new do |field, field_meta|
# the index starts at zero, that's why the third field would be 2:
field = '"' + field + '"' if field_meta.index == 2 && fields_meta.index > 1
field = '"' + field + '"' if field.is_a?(String) && field.include?(',')
# ^ CSV format needs to escape fields containing comma(s): ,
field
end
file = CSV.generate(headers: true, quote_char: '', write_converters: quoter) do |csv|
csv << headers
data.each { |row| csv << row }
end
puts file
the output would be:
id,ip,name,foo,bar
1,1.1.1.1,"Firstname Lastname",more,fields
2,2.2.2.2,"Firstname Lastname, Jr.",more,fields
It doesn't look like there's any way to do this with the existing CSV implementation short of monkey-patching/rewriting it.
However, assuming you have full control over the source data, you could do this:
Append a custom string including a comma (i.e. one that would never be naturally found in the data) to the end of the field in question for each row; maybe something like "FORCE_COMMAS,".
Generate the CSV output.
Now that you have CSV output with quotes on every row for your field, remove the custom string: csv.gsub!(/FORCE_COMMAS,/, "")
Customer feels warm and fuzzy.
CSV has changed a bit in Ruby 2.1 as mentioned by #jwadsack, however here's an working version of #the-tin-man's MyCSV. Bit modified, you set the forced_quote_fields via options.
MyCSV.generate(forced_quote_fields: [1]) do |_csv|...
The modified code
require 'csv'
class MyCSV < CSV
def <<(row)
# make sure headers have been assigned
if header_row? and [Array, String].include? #use_headers.class
parse_headers # won't read data for Array or String
self << #headers if #write_headers
end
# handle CSV::Row objects and Hashes
row = case row
when self.class::Row then row.fields
when Hash then #headers.map { |header| row[header] }
else row
end
#headers = row if header_row?
#lineno += 1
output = row.map.with_index(&#quote).join(#col_sep) + #row_sep # quote and separate
if #io.is_a?(StringIO) and
output.encoding != (encoding = raw_encoding)
if #force_encoding
output = output.encode(encoding)
elsif (compatible_encoding = Encoding.compatible?(#io.string, output))
#io.set_encoding(compatible_encoding)
#io.seek(0, IO::SEEK_END)
end
end
#io << output
self # for chaining
end
def init_separators(options)
# store the selected separators
#col_sep = options.delete(:col_sep).to_s.encode(#encoding)
#row_sep = options.delete(:row_sep) # encode after resolving :auto
#quote_char = options.delete(:quote_char).to_s.encode(#encoding)
#forced_quote_fields = options.delete(:forced_quote_fields) || []
if #quote_char.length != 1
raise ArgumentError, ":quote_char has to be a single character String"
end
#
# automatically discover row separator when requested
# (not fully encoding safe)
#
if #row_sep == :auto
if [ARGF, STDIN, STDOUT, STDERR].include?(#io) or
(defined?(Zlib) and #io.class == Zlib::GzipWriter)
#row_sep = $INPUT_RECORD_SEPARATOR
else
begin
#
# remember where we were (pos() will raise an exception if #io is pipe
# or not opened for reading)
#
saved_pos = #io.pos
while #row_sep == :auto
#
# if we run out of data, it's probably a single line
# (ensure will set default value)
#
break unless sample = #io.gets(nil, 1024)
# extend sample if we're unsure of the line ending
if sample.end_with? encode_str("\r")
sample << (#io.gets(nil, 1) || "")
end
# try to find a standard separator
if sample =~ encode_re("\r\n?|\n")
#row_sep = $&
break
end
end
# tricky seek() clone to work around GzipReader's lack of seek()
#io.rewind
# reset back to the remembered position
while saved_pos > 1024 # avoid loading a lot of data into memory
#io.read(1024)
saved_pos -= 1024
end
#io.read(saved_pos) if saved_pos.nonzero?
rescue IOError # not opened for reading
# do nothing: ensure will set default
rescue NoMethodError # Zlib::GzipWriter doesn't have some IO methods
# do nothing: ensure will set default
rescue SystemCallError # pipe
# do nothing: ensure will set default
ensure
#
# set default if we failed to detect
# (stream not opened for reading, a pipe, or a single line of data)
#
#row_sep = $INPUT_RECORD_SEPARATOR if #row_sep == :auto
end
end
end
#row_sep = #row_sep.to_s.encode(#encoding)
# establish quoting rules
#force_quotes = options.delete(:force_quotes)
do_quote = lambda do |field|
field = String(field)
encoded_quote = #quote_char.encode(field.encoding)
encoded_quote +
field.gsub(encoded_quote, encoded_quote * 2) +
encoded_quote
end
quotable_chars = encode_str("\r\n", #col_sep, #quote_char)
#quote = if #force_quotes
do_quote
else
lambda do |field, index|
if field.nil? # represent +nil+ fields as empty unquoted fields
""
else
field = String(field) # Stringify fields
# represent empty fields as empty quoted fields
if field.empty? or
field.count(quotable_chars).nonzero? or
#forced_quote_fields.include?(index)
do_quote.call(field)
else
field # unquoted field
end
end
end
end
end
end

Resources