saxlsx 0.2.0 → 0.3.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: f952d45e55462cfa02bcbf4c2dd5aeb4617f6971
-  data.tar.gz: 824471f88aa1b5892f35471688cb1ba0c5d6554e
+  metadata.gz: 5b0cbd91196c90e22a5dc4ede2bf372792cc5e3f
+  data.tar.gz: 96bec50a0662ab5741c54e329c80c073c9c202e2
 SHA512:
-  metadata.gz: eeb2a2adc96d6b43235ef07df32615459e23b9dab884eafaa4df9e816fda25e605788ceb9591abce7e5c27de62de7c13b224330d8dd68f3b6020d30b6b6ac1d1
-  data.tar.gz: 827d651635d5f47eb613559730bfe1f747c4912621142f832437e4eb3c71892448e68427918ad5012fddcda0bb82340ff09ab55c2268ceef8286e02ea3e71f14
+  metadata.gz: e55094b6131ae28e50afb8bed91d91f15b200596f94cb6215f9bea6c5e9303d158f281ec6495d8cbb481528f3893ea614748d3fd3c9130b06555df6d98930e10
+  data.tar.gz: fbd741be1e3b91784bc11502351a4195d31d0459de3765bd7de6acc64cebf4b5580a9eb0fc19bab0ad844f27ab316c84085ff9b4155bf76d52e5cef2f2d05738

data/lib/saxlsx/file_system.rb

@@ -26,8 +26,13 @@ module Saxlsx
       @zip.read('xl/sharedStrings.xml')
     end
 
-    def sheets
-      @zip.glob('xl/worksheets/sheet*.xml').sort.map{ |f| @zip.read(f).match(/<sheetData>.*<\/sheetData>/).to_s }
+    def styles
+      @zip.read('xl/styles.xml')
+    end
+
+    def sheet(i)
+      f = @zip.glob('xl/worksheets/sheet*.xml').sort[i]
+      @zip.read(f)
     end
 
   end

data/lib/saxlsx/rows_collection.rb

@@ -7,14 +7,21 @@ module Saxlsx
       @index = index
       @file_system = file_system
       @shared_strings = shared_strings
+      @sheet = file_system.sheet(index)
     end
 
     def each(&block)
-      RowsCollectionParser.parse @index, @file_system, @shared_strings, &block
+      RowsCollectionParser.parse @index, @sheet, @shared_strings, &block
     end
 
+    def count
+      @count ||= @sheet.match(/<dimension ref="[^:]+:[A-Z]*(\d+)"/)[1].to_i
+    end
+
+    alias :size :count
+
     def [](value)
       to_a[value]
     end
   end
-end
+end

data/lib/saxlsx/rows_collection_parser.rb

@@ -1,12 +1,49 @@
 module Saxlsx
   class RowsCollectionParser < Ox::Sax
+    NUM_FORMATS = {
+      0  => :string,         # General
+      1  => :fixnum,         # 0
+      2  => :float,          # 0.00
+      3  => :fixnum,         # #,##0
+      4  => :float,          # #,##0.00
+      5  => :unsupported,    # $#,##0_);($#,##0)
+      6  => :unsupported,    # $#,##0_);[Red]($#,##0)
+      7  => :unsupported,    # $#,##0.00_);($#,##0.00)
+      8  => :unsupported,    # $#,##0.00_);[Red]($#,##0.00)
+      9  => :percentage,     # 0%
+      10 => :percentage,     # 0.00%
+      11 => :bignum,         # 0.00E+00
+      12 => :unsupported,    # # ?/?
+      13 => :unsupported,    # # ??/??
+      14 => :date,           # mm-dd-yy
+      15 => :date,           # d-mmm-yy
+      16 => :date,           # d-mmm
+      17 => :date,           # mmm-yy
+      18 => :time,           # h:mm AM/PM
+      19 => :time,           # h:mm:ss AM/PM
+      20 => :time,           # h:mm
+      21 => :time,           # h:mm:ss
+      22 => :date_time,      # m/d/yy h:mm
+      37 => :unsupported,    # #,##0 ;(#,##0)
+      38 => :unsupported,    # #,##0 ;[Red](#,##0)
+      39 => :unsupported,    # #,##0.00;(#,##0.00)
+      40 => :unsupported,    # #,##0.00;[Red](#,##0.00)
+      45 => :time,           # mm:ss
+      46 => :time,           # [h]:mm:ss
+      47 => :time,           # mmss.0
+      48 => :bignum,         # ##0.0E+0
+      49 => :unsupported     # @
+    }
 
-    def self.parse(index, file_system, shared_strings, &block)
-      SaxParser.parse self.new(shared_strings, &block), file_system.sheets[index]
+    DATE_SYSTEM_1900 = DateTime.new(1899, 12, 30)
+
+    def self.parse(index, data, workbook, &block)
+      SaxParser.parse self.new(workbook, &block), data
     end
 
-    def initialize(shared_strings, &block)
-      @shared_strings = shared_strings
+    def initialize(workbook, &block)
+      @shared_strings = workbook.shared_strings
+      @number_formats = workbook.number_formats
       @block = block
     end
 
@@ -18,7 +55,10 @@ module Saxlsx
         @next_column = 'A'
       end
 
-      @current_type = nil if name == :c
+      if name == :c
+        @current_type = nil
+        @current_number_format = nil
+      end
     end
 
     def end_element(name)
@@ -30,8 +70,14 @@ module Saxlsx
 
     def attr(name, value)
       if @current_element == :c
-        @current_type = value if name == :t
-        @current_column = value.gsub(/\d/, '') if name == :r
+        case name
+        when :t
+          @current_type = value
+        when :r
+          @current_column = value.gsub(/\d/, '')
+        when :s
+          @current_number_format = detect_format_type(value.to_i)
+        end
       end
     end
 
@@ -50,17 +96,58 @@ module Saxlsx
 
     def value_of(text)
       case @current_type
-        when 's'
-          @shared_strings[text.to_i]
-        when 'b'
-          BooleanParser.parse text
-        when 'n'
+      when 's'
+        @shared_strings[text.to_i]
+      when 'b'
+        BooleanParser.parse text
+      else
+        case @current_number_format
+        when :date
+          DATE_SYSTEM_1900 + text.to_i
+        when :date_time
+          # Round time to seconds
+          date = DATE_SYSTEM_1900 + (text.to_f * 86400).round.fdiv(86400)
+          DateTime.new(date.year, date.month, date.day, date.hour, date.minute, date.second)
+        when :fixnum
+          text.to_i
+        when :float
           text.to_f
+        when :bignum
+          BigDecimal.new(text)
+        when :percentage
+          text.to_f / 100
         else
-          CGI.unescapeHTML(text)
+          if @current_type == 'n'
+            text.to_f
+          elsif text =~ /\A-?\d+(\.\d+(?:e[+-]\d+)?)?\Z/i # Auto convert numbers
+            $1 ? text.to_f : text.to_i
+          else
+            CGI.unescapeHTML(text)
+          end
+        end
       end
+    end
 
+    def detect_format_type(index)
+      format = @number_formats[index]
+      NUM_FORMATS[format] || detect_custom_format_type(format)
     end
 
+    # This is the least deterministic part of reading xlsx files. Due to
+    # custom styles, you can't know for sure when a date is a date other than
+    # looking at its format and gessing. It's not impossible to guess right,
+    # though.
+    #
+    # http://stackoverflow.com/questions/4948998/determining-if-an-xlsx-cell-is-date-formatted-for-excel-2007-spreadsheets
+    def detect_custom_format_type(code)
+      code = code.gsub(/\[[^\]]+\]/, '') # Strip meta - [...]
+      if code =~ /0/
+        :float
+      elsif code =~ /[ymdhis]/i
+        :date_time
+      else
+        :unsupported
+      end
+    end
   end
 end

data/lib/saxlsx/style_collection.rb

@@ -0,0 +1,15 @@
+module Saxlsx
+  class StyleCollection
+
+    include Enumerable
+
+    def initialize(file_system)
+      @file_system = file_system
+    end
+
+    def each(&block)
+      StyleCollectionParser.parse @file_system, &block
+    end
+
+  end
+end

data/lib/saxlsx/style_collection_parser.rb

@@ -0,0 +1,52 @@
+module Saxlsx
+  class StyleCollectionParser < Ox::Sax
+    def self.parse(file_system, &block)
+      SaxParser.parse self.new(&block), file_system.styles
+    end
+
+    def initialize(&block)
+      @block = block
+      @cell_styles = false
+      @custom_num_fmts = {}
+    end
+
+    def start_element(name)
+      case name
+      when :cellXfs
+        @cell_styles = true
+      when :xf
+        @num_fmt_id = nil
+      when :numFmt
+        @num_fmt_id = nil
+        @num_fmt_code = nil
+      end
+    end
+
+    def end_element(name)
+      case name
+      when :cellXfs
+        @cell_styles = false
+      when :xf
+        if @cell_styles
+          custom_num_fmt_code = @custom_num_fmts[@num_fmt_id]
+          if custom_num_fmt_code
+            @block.call custom_num_fmt_code
+          else
+            @block.call @num_fmt_id
+          end
+        end
+      when :numFmt
+        @custom_num_fmts[@num_fmt_id] = @num_fmt_code
+      end
+    end
+
+    def attr(name, value)
+      case name
+      when :numFmtId
+        @num_fmt_id = value.to_i
+      when :formatCode
+        @num_fmt_code = value
+      end
+    end
+  end
+end

data/lib/saxlsx/version.rb

@@ -1,3 +1,3 @@
 module Saxlsx
-  VERSION = '0.2.0'
+  VERSION = '0.3.0'
 end

data/lib/saxlsx/workbook.rb

@@ -19,7 +19,7 @@ module Saxlsx
     end
 
     def sheets(name=nil)
-      @sheets ||= SheetCollection.new(@file_system, shared_strings).to_a
+      @sheets ||= SheetCollection.new(@file_system, self).to_a
       name.nil? ? @sheets : @sheets.detect { |s| s.name == name }
     end
 
@@ -31,6 +31,10 @@ module Saxlsx
       @shared_strings ||= SharedStringCollection.new(@file_system).to_a
     end
 
+    def number_formats
+      @number_formats ||= StyleCollection.new(@file_system).to_a
+    end
+
     def to_csv(path)
       sheets.each { |s| s.to_csv path }
     end

data/spec/sheet_spec.rb

@@ -12,11 +12,11 @@ describe Sheet do
 
   it 'Rows count' do
     Workbook.open filename do |w|
-      w.sheets[0].should have(7).rows
-      w.sheets[1].should have(9).rows
-      w.sheets[2].should have(3).rows
-      w.sheets[3].should have(2).rows
-      w.sheets[4].should have(3).rows
+      w.sheets[0].rows.count.should eq 7
+      w.sheets[1].rows.count.should eq 9
+      w.sheets[2].rows.count.should eq 3
+      w.sheets[3].rows.count.should eq 2
+      w.sheets[4].rows.count.should eq 3
     end
   end
 
@@ -29,8 +29,20 @@ describe Sheet do
   it 'Rows content' do
     Workbook.open filename do |w|
       w.sheets[0].tap do |s|
-        s.rows[0].should eq ['LevenshteinDistance', 0]
-        s.rows[1].should eq ['Case sensitive', false]
+        s.rows[0].should eq [
+          'LevenshteinDistance',
+          3.14,
+          3,
+          DateTime.new(1970, 1, 1, 1, 0, 0),
+          DateTime.new(1970, 1, 1),
+          BigDecimal.new('3.4028236692093801E+38')
+        ]
+        s.rows[1].should eq [
+          'Case sensitive',
+          false,
+          3.0,
+          DateTime.new(1970, 1, 1, 1, 0, 0)
+        ]
         s.rows[2].should eq ['Fields', 'Type', 'URL Mining']
         s.rows[3].should eq ['autor', 'text', false]
         s.rows[4].should eq ['texto', 'text', false]
@@ -68,8 +80,8 @@ describe Sheet do
       w.sheets[0].to_csv tmp_path
 
       csv = File.open(csv_file, 'r') { |f| f.readlines }
-      csv[0].should eq "\"LevenshteinDistance\",\"0.0\"\n"
-      csv[1].should eq "\"Case sensitive\",\"false\"\n"
+      csv[0].should eq %{"LevenshteinDistance","3.14","3","1970-01-01T01:00:00+00:00","1970-01-01T00:00:00+00:00","0.34028236692093801E39"\n}
+      csv[1].should eq %{"Case sensitive","false","3.0","1970-01-01T01:00:00+00:00"\n}
       csv[2].should eq "\"Fields\",\"Type\",\"URL Mining\"\n"
       csv[3].should eq "\"autor\",\"text\",\"false\"\n"
       csv[4].should eq "\"texto\",\"text\",\"false\"\n"

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: saxlsx
 version: !ruby/object:Gem::Version
-  version: 0.2.0
+  version: 0.3.0
 platform: ruby
 authors:
 - Edgars Beigarts
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-09-11 00:00:00.000000000 Z
+date: 2015-03-25 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rubyzip
@@ -119,6 +119,8 @@ files:
 - lib/saxlsx/sheet.rb
 - lib/saxlsx/sheet_collection.rb
 - lib/saxlsx/sheet_collection_parser.rb
+- lib/saxlsx/style_collection.rb
+- lib/saxlsx/style_collection_parser.rb
 - lib/saxlsx/version.rb
 - lib/saxlsx/workbook.rb
 - saxlsx.gemspec