changed to fasterCSV for ruby < 1.9

added simple check for uri and integer

README

@@ -6,7 +6,9 @@
 # Date: 15.09.2011
 
 ## FILES ##
-csv2rdf.rb                            -- main ruby script to convert NORMARC file to RDF
+csv2rdf.rb        -- main ruby script to convert NORMARC file to RDF
+example.csv       -- example table
+example.rdf       -- example output
 
 ## FEATURES ##
 takes csv and uses column headers as properties
@@ -23,7 +25,7 @@ rdf-n3.rb (for n3 output)
 ## UBUNTU INSTALL ##
 (for rdf-xml support)
 sudo apt-get install libxml2-dev libxslt1-dev
-gem install rdf rdf-rdfxml rdf-n3
+gem install rdf rdf-rdfxml rdf-n3 (csv if ruby < 1.9)
 
 ## USAGE ##
 csv2rdf.rb  -i input_file.csv -o output_file -b base_uri -t rdf_type [-r recordlimit]
@@ -35,4 +37,4 @@ csv2rdf.rb  -i input_file.csv -o output_file -b base_uri -t rdf_type [-r recordl
 
 ## EXAMPLE ##
 
-ruby csv2rdf.rb -i input.csv -b http://example.com/bookreviews/ -t http://purl.org/stuff/rev#Review -o output.rdf
+ruby csv2rdf.rb -i example.csv -b http://example.com/bookreviews/ -t http://purl.org/stuff/rev#Review -o example.rdf

csv2rdf.rb

@@ -1,7 +1,13 @@
 #!/usr/bin/env ruby
 
 require 'rubygems'
-require 'csv'
+if RUBY_VERSION < "1.9" 
+  require "rubygems" 
+  require "faster_csv" 
+  CSV = FCSV 
+else 
+  require "csv" 
+end 
 require 'rdf'
 require 'rdf/rdfxml'
 require 'rdf/n3'
@@ -49,8 +55,6 @@ def initialize(record)
   def construct_uri
     @uri = RDF::URI.intern($base_uri)
     id = "#{@record}"
-
-    #id = "#{@record[headers[0]]}"
     id.gsub!(/[^\w\s\-ÆØÅæøå]/,"")
     id.gsub!(/\s/,"_")
     @uri += id
@@ -74,24 +78,27 @@ def write_record
 
 count = 0
 
-csv = CSV.read($input_file)
-headers = csv.shift.map {|i| i.to_s }
-string_data = csv.map {|row| row.map {|cell| cell.to_s } }
-array_of_hashes = string_data.map {|row| Hash[*headers.zip(row).flatten] }
+csv = CSV.read($input_file, {:headers => true, :encoding => 'UTF-8'})
 # start writer handle
 RDF::Writer.open($output_file) do | writer |
 @@writer = writer
 
-  array_of_hashes.each do | record |
+  csv.each do | record |
     count += 1
     if $recordlimit then break if count > $recordlimit end
       # take the content of the first column to make record id
-      rdfrecord = RDFModeler.new(record["#{headers[0]}"])
+      rdfrecord = RDFModeler.new(record[0])
       rdfrecord.set_type(RDF::URI($rdf_type))
       record.each do |k,v| 
-        unless v.empty?
+        unless v.nil?
           v.strip_leading_and_trailing_punct
-          rdfrecord.assert(RDF::URI(k), RDF::Literal("#{v}"))
+          if v =~ /^http/
+            rdfrecord.assert(RDF::URI(k), RDF::URI("#{v}"))
+          elsif v =~ /^[\d]+$/
+            rdfrecord.assert(RDF::URI(k), RDF::Literal("#{v}", :datatype => RDF::XSD.integer))
+          else
+            rdfrecord.assert(RDF::URI(k), v)
+          end
         end
       end