fastest-csv 0.0.2-java → 0.0.4-java

data/ext/csv_parser/CsvParser.java

@@ -0,0 +1,89 @@
+//
+// Copyright (c) Maarten Oelering, BrightCode BV
+//
+
+package org.brightcode;
+
+import java.util.ArrayList;
+import java.util.List;
+
+public class CsvParser {
+
+    private static int DEF_ARRAY_LEN = 32;
+
+    private static int UNQUOTED = 0;
+    private static int IN_QUOTED = 1;
+    private static int QUOTE_IN_QUOTED = 2;
+    
+    public static List parseLine(String line) {
+        int length = line.length();
+        if (length == 0)
+            return null;
+        
+        int state = UNQUOTED;
+        StringBuilder value = new StringBuilder(length);   // field value, no longer than line
+        List<String> array = new ArrayList<String>(DEF_ARRAY_LEN);
+        
+        for (int i = 0; i < length; i++) {
+            char c = line.charAt(i);
+            switch (c) {
+                case ',':
+                    if (state == UNQUOTED) {
+                        if (value.length() == 0) {
+                            array.add(null);
+                        } 
+                        else {
+                            array.add(value.toString());
+                            value.setLength(0);
+                        }
+                    }
+                    else if (state == IN_QUOTED) {
+                        value.append(c);
+                    }
+                    else if (state == 2) {
+                        array.add(value.toString());
+                        value.setLength(0);
+                        state = UNQUOTED;
+                    }
+                    break;
+                case '"':
+                    if (state == UNQUOTED) {
+                        state = IN_QUOTED;
+                    }
+                    else if (state == IN_QUOTED) {
+                        state = QUOTE_IN_QUOTED;
+                    }
+                    else if (state == QUOTE_IN_QUOTED) {
+                        value.append(c);   // escaped quote
+                        state = IN_QUOTED;
+                    }
+                    break;
+                case '\r':
+                case '\n':
+                    if (state == IN_QUOTED) {
+                        value.append(c);
+                    }
+                    else {
+                        i = length;  // only parse first line if multiline
+                    }
+                    break;
+                default:
+                    value.append(c);
+                    break;
+            }
+        }
+        if (state == UNQUOTED) {
+            if (value.length() == 0) {
+                array.add(null);
+            } 
+            else {
+                array.add(value.toString());
+                value.setLength(0);
+            }
+        }
+        else if (state == QUOTE_IN_QUOTED) {
+            array.add(value.toString());
+        }
+        return array;
+    }
+}

data/ext/csv_parser/CsvParserService.java

@@ -19,6 +19,8 @@ public class CsvParserService implements BasicLibraryService {
 
     private Ruby runtime;
 
+    private static int DEF_ARRAY_LEN = 32;
+
     private static int UNQUOTED = 0;
     private static int IN_QUOTED = 1;
     private static int QUOTE_IN_QUOTED = 2;
@@ -29,6 +31,7 @@ public class CsvParserService implements BasicLibraryService {
         this.runtime = runtime;
 
         RubyModule mCsvParser = runtime.defineModule("CsvParser");
+        // TODO: CallbackFactory#getSingletonMethod is deprecated
         CallbackFactory callbackFactory = runtime.callbackFactory(CsvParserService.class);
         mCsvParser.defineModuleFunction("parse_line", 
             callbackFactory.getSingletonMethod("parseLine", RubyString.class));
@@ -45,7 +48,7 @@ public class CsvParserService implements BasicLibraryService {
         
         int state = UNQUOTED;
         StringBuilder value = new StringBuilder(length);   // field value, no longer than line
-        RubyArray array = RubyArray.newArray(runtime, 36);
+        RubyArray array = RubyArray.newArray(runtime, DEF_ARRAY_LEN);
         
         for (int i = 0; i < length; i++) {
             char c = seq.charAt(i);

data/ext/csv_parser/parser.c

@@ -9,16 +9,12 @@
   #include "ruby/io.h"
 #endif
 
-/* default allocated size is 16 */
 #define DEF_ARRAY_LEN 32
 
 #define UNQUOTED 0
 #define IN_QUOTED 1
 #define QUOTE_IN_QUOTED 2
 
-/*
-static VALUE cFastestCSV;
-*/
 static VALUE mCsvParser;
 
 static VALUE parse_line(VALUE self, VALUE str)
@@ -32,7 +28,7 @@ static VALUE parse_line(VALUE self, VALUE str)
     if (len == 0)
         return Qnil;
     
-    VALUE array = rb_ary_new2(DEF_ARRAY_LEN); 
+    VALUE array = rb_ary_new2(DEF_ARRAY_LEN); /* default allocated size is 16 */
     char value[len];  /* field value, no longer than line */
     int state = 0;
     int index = 0;
@@ -94,10 +90,6 @@ static VALUE parse_line(VALUE self, VALUE str)
 
 void Init_csv_parser()
 {
-    /*
-    cFastestCSV = rb_define_class("FastestCSV", rb_cObject);
-    rb_define_singleton_method(cFastestCSV, "parse_line", parse_line, 1);
-    */
     mCsvParser = rb_define_module("CsvParser");
     rb_define_module_function(mCsvParser, "parse_line", parse_line, 1);
 }

data/lib/fastest-csv/version.rb

@@ -1,3 +1,3 @@
 class FastestCSV
-  VERSION = "0.0.2"
+  VERSION = "0.0.4"
 end

data/lib/fastest_csv.rb

@@ -5,10 +5,15 @@ require 'stringio'
 
 # Fast CSV parser using native code
 class FastestCSV
+  include Enumerable
   
   if RUBY_PLATFORM =~ /java/
-    require 'jruby'
-    org.brightcode.CsvParserService.new.basicLoad(JRuby.runtime)
+    if JRUBY_VERSION =~ /^1\.[0-6]/
+      require 'jruby'
+      org.brightcode.CsvParserService.new.basicLoad(JRuby.runtime)
+    else
+      include_package "org.brightcode"
+    end
   end
 
   # Pass each line of the specified +path+ as array to the provided +block+
@@ -58,7 +63,7 @@ class FastestCSV
   end
   
   def self.parse_line(line)
-    ::CsvParser.parse_line(line)
+    CsvParser.parse_line(line)
   end
 
   # Create new FastestCSV wrapping the specified IO object
@@ -68,8 +73,12 @@ class FastestCSV
   
   # Read from the wrapped IO passing each line as array to the specified block
   def each
-    while row = shift
-      yield row
+    if block_given?
+      while row = shift
+        yield row
+      end
+    else
+      to_enum # return enumerator
     end
   end
   
@@ -80,11 +89,16 @@ class FastestCSV
     table
   end
   alias_method :readlines, :read
+  
+  # Rewind the underlying IO object and reset line counter
+  def rewind
+    @io.rewind
+  end
 
   # Read next line from the wrapped IO and return as array or nil at EOF
   def shift
     if line = @io.gets
-      ::CsvParser.parse_line(line)
+      CsvParser.parse_line(line)
     else
       nil
     end
@@ -105,7 +119,7 @@ end
 class String
   # Equivalent to <tt>FasterCSV::parse_line(self)</tt>
   def parse_csv
-    ::CsvParser.parse_line(self)
+    CsvParser.parse_line(self)
   end
 end
 

data/test/tc_csv_parsing.rb

@@ -13,6 +13,10 @@ require 'fastest_csv'
 # 
 class TestCSVParsing < Test::Unit::TestCase
 
+  if RUBY_PLATFORM =~ /java/
+    include_package "org.brightcode"
+  end
+
   def test_mastering_regex_example
     ex = %Q{Ten Thousand,10000, 2710 ,,"10,000","It's ""10 Grand"", baby",10K}
     assert_equal( [ "Ten Thousand", "10000", " 2710 ", nil, "10,000",

data/test/tc_interface.rb

@@ -116,4 +116,13 @@ class TestFastestCSVInterface < Test::Unit::TestCase
     @expected = [%w{1 2} + ['3' * long_field_length]]
     test_shift
   end
+  
+  def test_enumerable
+    FastestCSV.open(@path) do |csv|
+      assert(csv.include?(["1", "2", "3"]))
+      csv.rewind
+      assert_equal([["1", "2", "3"], ["4", "5"]], csv.to_a)
+    end
+  end
+  
 end

metadata

@@ -1,80 +1,79 @@
---- !ruby/object:Gem::Specification 
+--- !ruby/object:Gem::Specification
 name: fastest-csv
-version: !ruby/object:Gem::Version 
-  prerelease: 
-  version: 0.0.2
+version: !ruby/object:Gem::Version
+  prerelease:
+  version: 0.0.4
 platform: java
-authors: 
-  - Maarten Oelering
-autorequire: 
+authors:
+- Maarten Oelering
+autorequire:
 bindir: bin
 cert_chain: []
-
-date: 2013-01-13 00:00:00 Z
-dependencies: 
-  - !ruby/object:Gem::Dependency 
-    name: rake-compiler
-    prerelease: false
-    requirement: &id001 !ruby/object:Gem::Requirement 
-      none: false
-      requirements: 
-        - - ">="
-          - !ruby/object:Gem::Version 
-            version: "0"
-    type: :development
-    version_requirements: *id001
+date: 2013-08-16 00:00:00.000000000 Z
+dependencies:
+- !ruby/object:Gem::Dependency
+  name: rake-compiler
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+    none: false
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+    none: false
+  prerelease: false
+  type: :development
 description: Fastest standard CSV parser for MRI Ruby and JRuby
-email: 
-  - maarten@brightcode.nl
+email:
+- maarten@brightcode.nl
 executables: []
-
 extensions: []
-
 extra_rdoc_files: []
-
-files: 
-  - .gitignore
-  - Gemfile
-  - LICENSE
-  - README.md
-  - Rakefile
-  - ext/csv_parser/CsvParserService.java
-  - ext/csv_parser/extconf.rb
-  - ext/csv_parser/parser.c
-  - fastest-csv.gemspec
-  - lib/fastest-csv.rb
-  - lib/fastest-csv/version.rb
-  - lib/fastest_csv.rb
-  - test/tc_csv_parsing.rb
-  - test/tc_interface.rb
-  - lib/csv_parser.jar
+files:
+- .gitignore
+- Gemfile
+- LICENSE
+- README.md
+- Rakefile
+- ext/csv_parser/CsvParser.java
+- ext/csv_parser/CsvParserService.java
+- ext/csv_parser/extconf.rb
+- ext/csv_parser/parser.c
+- fastest-csv.gemspec
+- lib/fastest-csv.rb
+- lib/fastest-csv/version.rb
+- lib/fastest_csv.rb
+- test/tc_csv_parsing.rb
+- test/tc_interface.rb
+- lib/csv_parser.jar
 homepage: https://github.com/brightcode/fastest-csv
 licenses: []
-
-post_install_message: 
+post_install_message:
 rdoc_options: []
-
-require_paths: 
-  - lib
-required_ruby_version: !ruby/object:Gem::Requirement 
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - '>='
+    - !ruby/object:Gem::Version
+      version: '0'
   none: false
-  requirements: 
-    - - ">="
-      - !ruby/object:Gem::Version 
-        version: "0"
-required_rubygems_version: !ruby/object:Gem::Requirement 
+required_rubygems_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - '>='
+    - !ruby/object:Gem::Version
+      version: '0'
   none: false
-  requirements: 
-    - - ">="
-      - !ruby/object:Gem::Version 
-        version: "0"
 requirements: []
-
-rubyforge_project: 
+rubyforge_project:
 rubygems_version: 1.8.24
-signing_key: 
+signing_key:
 specification_version: 3
 summary: Fastest standard CSV parser for MRI Ruby and JRuby
-test_files: 
-  - test/tc_csv_parsing.rb
-  - test/tc_interface.rb
+test_files:
+- test/tc_csv_parsing.rb
+- test/tc_interface.rb