bamfcsv 0.0.2 → 0.1.0

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    bamfcsv (0.0.1)
+    bamfcsv (0.0.2)
 
 GEM
   remote: http://rubygems.org/

data/ext/bamfcsv/bamfcsv_ext.c

@@ -1,7 +1,4 @@
 #include <stdlib.h>
-#include <ruby/ruby.h>
-#include <fcntl.h>
-#include <sys/mman.h>
 #include "bamfcsv_ext.h"
 
 struct s_Row *alloc_row() {
@@ -71,10 +68,14 @@ VALUE build_matrix_from_pointer_tree(struct s_Row *first_row, int num_rows) {
     rb_ary_store(matrix,i,row);
     for (j = 0; j < cur_row->cell_count; j++) {
       if (*(cur_cell->start) == '"' 
-          && *((cur_cell->start)+((cur_cell->len-1)*sizeof(char))) == '"')
-        new_string = rb_str_new(cur_cell->start+sizeof(char), cur_cell->len-(sizeof(char)*2));
-      else
-        new_string = rb_str_new(cur_cell->start, cur_cell->len);
+          && *((cur_cell->start)+(cur_cell->len-1)) == '"')
+        new_string = rb_str_new(cur_cell->start+1, cur_cell->len-2);
+      else {
+        if (cur_cell->len)
+          new_string = rb_str_new(cur_cell->start, cur_cell->len);
+        else
+          new_string = Qnil; /* Empty, unquoted cells are nil, for default ruby CSV compatibility */
+      }
       if (cur_cell->has_quotes) {
         rb_funcall(new_string, gsub, 2, dquote, quote);
       }
@@ -87,17 +88,19 @@ VALUE build_matrix_from_pointer_tree(struct s_Row *first_row, int num_rows) {
   return matrix;
 }
 
-void finalize_cell(struct s_Cell *cell, char *cur) {
-  if (*(cur-sizeof(char)) == '\r') 
-    cell->len = cur-(cell->start)-sizeof(char);
+void finalize_cell(struct s_Cell *cell, char *cur, int quote_count) {
+  if (*(cur-1) == '\r') 
+    cell->len = cur-(cell->start)-1;
   else
     cell->len = cur-(cell->start);
+
+  if (quote_count) cell->has_quotes = 1;
 }
 
 VALUE build_matrix(char *buf, int bufsize) {
   int str_start = 0;
   int num_rows = 1;
-  int in_quote = 0;
+  int quote_count = 0, quotes_matched = 1;
 
   struct s_Row *first_row = alloc_row();
   struct s_Row *cur_row = first_row;
@@ -112,33 +115,43 @@ VALUE build_matrix(char *buf, int bufsize) {
   for (cur = buf; cur < buf+bufsize; cur++) {
 
     if (*cur == '"') {
-      if (in_quote)
-        if (*(cur+1) != ',')
-          cur_cell->has_quotes = 1;
-      in_quote = !in_quote;
+      if (0 == quote_count && cur_cell->start != cur) /* Quotes begin past opening of cell */
+        rb_raise(BAMFCSV_MalformedCSVError_class, "Illegal quoting on line %d, cell %d: Quoted cell must open with '\"'", num_rows, cur_row->cell_count+1);
+      else
+        ++quote_count;
     }
 
-    if (!in_quote) {
+    quotes_matched = !(quote_count & 1); /* count is even */
+
+    if (quotes_matched) { 
 
       if (*cur == ',') {
         
-        finalize_cell(cur_cell,cur);
+        if (quote_count && *(cur-1) != '"')
+          rb_raise(BAMFCSV_MalformedCSVError_class, "Unclosed quoted field on line %d, cell %d.", num_rows, cur_row->cell_count+1);
+
+        finalize_cell(cur_cell,cur,quote_count);
         cur_cell->next_cell = alloc_cell();
         cur_cell = cur_cell->next_cell;
-        cur_cell->start = cur+sizeof(char);
+        cur_cell->start = cur+1;
         cur_row->cell_count += 1;
+        quote_count = 0;
 
       }
       
       if (*cur == '\n') {
         
-        finalize_cell(cur_cell,cur);
+        if (quote_count && !(*(cur-1) == '"' || *(cur-1) == '\r' && *(cur-2) == '"'))
+            rb_raise(BAMFCSV_MalformedCSVError_class, "Unclosed quoted field on line %d, cell %d: EOL", num_rows, cur_row->cell_count+1);
+
+        finalize_cell(cur_cell,cur,quote_count);
         cur_row->cell_count += 1;
         cur_row->next_row = alloc_row();
         cur_row = cur_row -> next_row;
         cur_row->first_cell = alloc_cell();
         cur_cell = cur_row->first_cell;
-        cur_cell->start = cur+sizeof(char);
+        cur_cell->start = cur+1;
+        quote_count = 0;
         
         num_rows++;
 
@@ -148,8 +161,16 @@ VALUE build_matrix(char *buf, int bufsize) {
 
   }
 
-  if (cur_row->cell_count == 0) {
+  if (!quotes_matched) /* Reached EOF without matching quotes */
+    rb_raise(BAMFCSV_MalformedCSVError_class, "Illegal quoting on line %d, cell %d: File ends without closing '\"'", num_rows, cur_row->cell_count+1);
+  else if (quote_count && *cur != '"')
+    rb_raise(BAMFCSV_MalformedCSVError_class, "Unclosed quoted field on line %d, cell %d: EOF", num_rows, cur_row->cell_count+1);
+
+  if (cur_row->cell_count == 0) { /* Ended with newline */
     num_rows--;
+  } else { /* No newline before EOF */
+    finalize_cell(cur_cell, cur, quote_count);
+    cur_row->cell_count++;
   }
 
   matrix = build_matrix_from_pointer_tree(first_row, num_rows);
@@ -160,32 +181,17 @@ VALUE build_matrix(char *buf, int bufsize) {
 
 }
 
-VALUE mm_parse(const char *file) {
-
-  char *mmapped_csv;
-  int filesize, csv;
-
-  csv = open(file, O_RDONLY);
-  filesize = lseek(csv, 0, SEEK_END);
-  mmapped_csv = (char*) mmap(0, filesize, PROT_READ, MAP_SHARED, csv, 0);
-
-  VALUE matrix = build_matrix(mmapped_csv,filesize);
-
-  munmap(mmapped_csv, filesize);
-  close(csv);
-
-  return matrix;
-}
-
-VALUE read_path(VALUE self, VALUE file) {
+VALUE parse_string(VALUE self, VALUE string) {
 
-  return mm_parse(RSTRING_PTR(file));
+  return build_matrix(RSTRING_PTR(string), NUM2INT(rb_str_length(string)));
 
 }
 
 void Init_bamfcsv() {
 
-  VALUE module = rb_define_module("BAMFCSV");
-  rb_define_module_function(module, "read_path", read_path, 1);
+  BAMFCSV_module = rb_define_module("BAMFCSV");
+  VALUE bamfcsv_singleton_class = rb_singleton_class(BAMFCSV_module);
+  rb_define_private_method(bamfcsv_singleton_class, "__parse_string", parse_string, 1);
 
+  BAMFCSV_MalformedCSVError_class = rb_define_class_under(BAMFCSV_module, "MalformedCSVError", rb_eRuntimeError);
 }

data/ext/bamfcsv/bamfcsv_ext.h

@@ -1,6 +1,11 @@
 #ifndef _BAMFCSV_EXT_H
 #define _BAMFCSV_EXT_H
 
+#include <ruby/ruby.h>
+
+VALUE BAMFCSV_module;
+VALUE BAMFCSV_MalformedCSVError_class;
+
 struct s_Row {
   struct s_Cell *first_cell;
   struct s_Row *next_row;

data/lib/bamfcsv/version.rb

@@ -1,3 +1,3 @@
 module BAMFCSV
-  VERSION = "0.0.2"
+  VERSION = "0.1.0"
 end

data/lib/bamfcsv.rb

@@ -3,11 +3,11 @@ require 'bamfcsv/bamfcsv'
 module BAMFCSV
 
   def self.read(thing_to_read)
-    if  String === thing_to_read
-      raise Errno::ENOENT.new("#{thing_to_read} does not exist") unless File.exist? thing_to_read
-      raise Errno::EISDIR.new("#{thing_to_read} is a directory") if File.directory? thing_to_read
-      read_path(thing_to_read)
-    end
+    __parse_string(File.read(thing_to_read))
+  end
+
+  def self.parse(csv_str)
+    __parse_string(csv_str)
   end
 
 end

data/spec/lib/bamfcsv_spec.rb

@@ -1,10 +1,14 @@
 require 'spec_helper'
 
 describe BAMFCSV do
-  it "has a parse method" do
+  it "has a read method" do
     BAMFCSV.should respond_to(:read)
   end
 
+  it "has a parse method" do
+    BAMFCSV.should respond_to(:parse)
+  end
+
   describe "#read" do
     it "is a matrix given a filename" do
       BAMFCSV.read("spec/fixtures/test.csv").should be_instance_of Array
@@ -19,7 +23,7 @@ describe BAMFCSV do
     end
 
     it "interprets empty cells correctly" do
-      BAMFCSV.read("spec/fixtures/bamf-comma-comma.csv").should == [["BAMF","","CSV"]]
+      BAMFCSV.read("spec/fixtures/bamf-comma-comma.csv").should == [["BAMF",nil,"CSV"]]
     end
 
     it "escapes cells that are quoted" do
@@ -46,4 +50,70 @@ describe BAMFCSV do
       end.should raise_error Errno::EISDIR
     end
   end
+
+  describe "#parse" do
+    it "correctly parses the last cell even if there is no newline" do
+      BAMFCSV.parse("1,2").should == [["1","2"]]
+    end
+
+    it 'correctly escaptes ""' do
+      BAMFCSV.parse("1,\"\"2\"\"\n").should == [["1", '"2"']]
+    end
+
+    it "parses unquoted empty cells as nil" do
+      BAMFCSV.parse("1,,2").should == [["1",nil,"2"]]
+    end
+
+    it 'parses quoted empty cells as ""' do
+      BAMFCSV.parse("1,\"\",2").should == [["1","","2"]]
+    end
+
+    describe "default CSV module compatibility" do
+      it "adds a nil cell after a trailing comma with no newline" do
+        BAMFCSV.parse("1,2,").should == [["1","2",nil]]
+      end
+
+      it "adds a nil cell after a trailing comma with an ending newline" do
+        BAMFCSV.parse("1,2,\n").should == [["1","2",nil]]
+      end
+
+      describe "when a quoted cell ends a line" do
+        it "does not raise an exception" do
+          expect { BAMFCSV.parse(%Q|1,2,"3,4"\n5,6,7|) }.should_not raise_error
+          expect { BAMFCSV.parse(%Q|1,2,"3,4"\r\n5,6,7|) }.should_not raise_error
+        end
+
+        it "correctly parses a quoted cell at the end of a line" do
+          BAMFCSV.parse(%Q|1,2,"3,4"\n5,6,7|).should == [["1","2","3,4"],["5","6","7"]]
+          BAMFCSV.parse(%Q|1,2,"3,4"\r\n5,6,7|).should == [["1","2","3,4"],["5","6","7"]]
+        end
+      end
+
+      it "raises BAMFCSV::MalformedCSVError when quotes appear in a cell which was not started with quotes" do
+        expect { BAMFCSV.parse(' ""') }.should raise_error(BAMFCSV::MalformedCSVError)
+        expect { BAMFCSV.parse(" \"\"\n") }.should raise_error(BAMFCSV::MalformedCSVError)
+        expect { BAMFCSV.parse(" \"\"\r\n") }.should raise_error(BAMFCSV::MalformedCSVError)
+        expect { BAMFCSV.parse('1, "",3') }.should raise_error(BAMFCSV::MalformedCSVError)
+      end
+
+      it "raises BAMFCSV::MalformedCSVError when a quoted cell is not closed at its end" do
+        expect { BAMFCSV.parse('"') }.should raise_error(BAMFCSV::MalformedCSVError)
+        expect { BAMFCSV.parse('" ""') }.should raise_error(BAMFCSV::MalformedCSVError)
+        expect { BAMFCSV.parse("\"\n") }.should raise_error(BAMFCSV::MalformedCSVError)
+        expect { BAMFCSV.parse("\"\r\n") }.should raise_error(BAMFCSV::MalformedCSVError)
+        expect { BAMFCSV.parse("\" \"\"\n") }.should raise_error(BAMFCSV::MalformedCSVError)
+        expect { BAMFCSV.parse("\" \"\"\r\n") }.should raise_error(BAMFCSV::MalformedCSVError)
+        expect { BAMFCSV.parse('1,"2,3') }.should raise_error(BAMFCSV::MalformedCSVError)
+        expect { BAMFCSV.parse("1,\"2,3\n") }.should raise_error(BAMFCSV::MalformedCSVError)
+        expect { BAMFCSV.parse("1,\"2,3\r\n") }.should raise_error(BAMFCSV::MalformedCSVError)
+      end
+
+      it "raises BAMFCSV::MalformedCSVError when quoted cell is closed before its end" do
+        expect { BAMFCSV.parse('"" ') }.should raise_error(BAMFCSV::MalformedCSVError)
+        expect { BAMFCSV.parse("\"\" \n") }.should raise_error(BAMFCSV::MalformedCSVError)
+        expect { BAMFCSV.parse("\"\" \r\n") }.should raise_error(BAMFCSV::MalformedCSVError)
+        expect { BAMFCSV.parse('1,"" ,2') }.should raise_error(BAMFCSV::MalformedCSVError)
+      end
+    end
+  end
 end

metadata

@@ -1,60 +1,63 @@
---- !ruby/object:Gem::Specification
+--- !ruby/object:Gem::Specification 
 name: bamfcsv
-version: !ruby/object:Gem::Version
-  version: 0.0.2
+version: !ruby/object:Gem::Version 
   prerelease: 
+  version: 0.1.0
 platform: ruby
-authors:
+authors: 
 - Jon Distad
 - Alex Redington
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2011-04-01 00:00:00.000000000 +00:00
+
+date: 2011-04-03 00:00:00 -04:00
 default_executable: 
-dependencies:
-- !ruby/object:Gem::Dependency
+dependencies: 
+- !ruby/object:Gem::Dependency 
   name: rspec
-  requirement: &22500120 !ruby/object:Gem::Requirement
+  prerelease: false
+  requirement: &id001 !ruby/object:Gem::Requirement 
     none: false
-    requirements:
+    requirements: 
     - - ~>
-      - !ruby/object:Gem::Version
+      - !ruby/object:Gem::Version 
         version: 2.5.0
   type: :development
-  prerelease: false
-  version_requirements: *22500120
-- !ruby/object:Gem::Dependency
+  version_requirements: *id001
+- !ruby/object:Gem::Dependency 
   name: fuubar
-  requirement: &22499620 !ruby/object:Gem::Requirement
+  prerelease: false
+  requirement: &id002 !ruby/object:Gem::Requirement 
     none: false
-    requirements:
+    requirements: 
     - - ~>
-      - !ruby/object:Gem::Version
+      - !ruby/object:Gem::Version 
         version: 0.0.2
   type: :development
-  prerelease: false
-  version_requirements: *22499620
-- !ruby/object:Gem::Dependency
+  version_requirements: *id002
+- !ruby/object:Gem::Dependency 
   name: rake-compiler
-  requirement: &22499160 !ruby/object:Gem::Requirement
+  prerelease: false
+  requirement: &id003 !ruby/object:Gem::Requirement 
     none: false
-    requirements:
+    requirements: 
     - - ~>
-      - !ruby/object:Gem::Version
+      - !ruby/object:Gem::Version 
         version: 0.7.1
   type: :development
-  prerelease: false
-  version_requirements: *22499160
+  version_requirements: *id003
 description: BAMFCSV parses csv like a BAMF. BAMF!!
-email:
+email: 
 - jon@thinkrelevance.com
 - lovemachine@thinkrelevance.com
 executables: []
-extensions:
+
+extensions: 
 - ext/bamfcsv/extconf.rb
 extra_rdoc_files: []
-files:
+
+files: 
 - .gitignore
 - .rspec
 - Gemfile
@@ -81,27 +84,39 @@ files:
 has_rdoc: true
 homepage: https://github.com/jondistad/bamfcsv
 licenses: []
+
 post_install_message: 
 rdoc_options: []
-require_paths:
+
+require_paths: 
 - lib
 - ext
-required_ruby_version: !ruby/object:Gem::Requirement
+required_ruby_version: !ruby/object:Gem::Requirement 
   none: false
-  requirements:
-  - - ! '>='
-    - !ruby/object:Gem::Version
-      version: '0'
-required_rubygems_version: !ruby/object:Gem::Requirement
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      version: "0"
+required_rubygems_version: !ruby/object:Gem::Requirement 
   none: false
-  requirements:
-  - - ! '>='
-    - !ruby/object:Gem::Version
-      version: '0'
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      version: "0"
 requirements: []
+
 rubyforge_project: bamfcsv
-rubygems_version: 1.5.2
+rubygems_version: 1.6.2
 signing_key: 
 specification_version: 3
 summary: BAMF!!! Your csv is parsed.
-test_files: []
+test_files: 
+- spec/fixtures/bamf-comma-comma.csv
+- spec/fixtures/double-quotes.csv
+- spec/fixtures/empty.csv
+- spec/fixtures/escapes.csv
+- spec/fixtures/one-column.csv
+- spec/fixtures/terminated-with-cr.csv
+- spec/fixtures/test.csv
+- spec/lib/bamfcsv_spec.rb
+- spec/spec_helper.rb