roo 1.13.2 → 2.0.0beta1

data/lib/roo/excelx/comments.rb

@@ -0,0 +1,24 @@
+require 'roo/excelx/extractor'
+
+module Roo
+  class Excelx::Comments < Excelx::Extractor
+
+    def comments
+      @comments ||= extract_comments
+    end
+
+    private
+
+    def extract_comments
+      if doc_exists?
+        Hash[doc.xpath("//comments/commentList/comment").map do |comment|
+          value = (comment.at_xpath('./text/r/t') || comment.at_xpath('./text/t')).text
+          [::Roo::Utils.ref_to_key(comment.attributes['ref'].to_s), value]
+        end]
+      else
+        {}
+      end
+    end
+
+  end
+end

data/lib/roo/excelx/extractor.rb

@@ -0,0 +1,20 @@
+module Roo
+  class Excelx::Extractor
+    def initialize(path)
+      @path = path
+    end
+
+    private
+
+    def doc
+      @doc ||=
+        if doc_exists?
+          ::Roo::Utils.load_xml(@path).remove_namespaces!
+        end
+    end
+
+    def doc_exists?
+      @path && File.exist?(@path)
+    end
+  end
+end

data/lib/roo/excelx/relationships.rb

@@ -0,0 +1,26 @@
+require 'roo/excelx/extractor'
+
+module Roo
+  class Excelx::Relationships < Excelx::Extractor
+    def [](index)
+      to_a[index]
+    end
+
+    def to_a
+      @relationships ||= extract_relationships
+    end
+
+    private
+
+    def extract_relationships
+      if doc_exists?
+        Hash[doc.xpath("/Relationships/Relationship").map do |rel|
+          [rel.attribute('Id').text, rel]
+        end]
+      else
+        []
+      end
+    end
+
+  end
+end

data/lib/roo/excelx/shared_strings.rb

@@ -0,0 +1,40 @@
+require 'roo/excelx/extractor'
+
+module Roo
+  class Excelx::SharedStrings < Excelx::Extractor
+    def [](index)
+      to_a[index]
+    end
+
+    def to_a
+      @array ||= extract_shared_strings
+    end
+
+    private
+
+    def extract_shared_strings
+      if doc_exists?
+        # read the shared strings xml document
+        doc.xpath("/sst/si").map do |si|
+          shared_string = ''
+          si.children.each do |elem|
+            case elem.name
+              when 'r'
+                elem.children.each do |r_elem|
+                  if r_elem.name == 't'
+                    shared_string << r_elem.content
+                  end
+                end
+              when 't'
+                shared_string = elem.content
+            end
+          end
+          shared_string
+        end
+      else
+        []
+      end
+    end
+
+  end
+end

data/lib/roo/excelx/sheet_doc.rb

@@ -0,0 +1,202 @@
+require 'roo/excelx/extractor'
+
+module Roo
+  class Excelx::SheetDoc < Excelx::Extractor
+    def initialize(path, relationships, styles, shared_strings, workbook, options = {})
+      super(path)
+      @options = options
+      @relationships = relationships
+      @styles = styles
+      @shared_strings = shared_strings
+      @workbook = workbook
+    end
+
+    def cells(relationships)
+      @cells ||= extract_cells(relationships)
+    end
+
+    def hyperlinks(relationships)
+      @hyperlinks ||= extract_hyperlinks(relationships)
+    end
+
+    # Get the dimensions for the sheet.
+    # This is the upper bound of cells that might
+    # be parsed. (the document may be sparse so cell count is only upper bound)
+    def dimensions
+      @dimensions ||= extract_dimensions
+    end
+
+    # Yield each row xml element to caller
+    def each_row_streaming(&block)
+      Roo::Utils.each_element(@path, 'row', &block)
+    end
+
+    # Yield each cell as Excelx::Cell to caller for given
+    # row xml
+    def each_cell(row_xml)
+      return [] unless row_xml
+      row_xml.children.each do |cell_element|
+        key = ::Roo::Utils.ref_to_key(cell_element['r'])
+        yield cell_from_xml(cell_element, hyperlinks(@relationships)[key])
+      end
+    end
+
+    private
+
+    def cell_from_xml(cell_xml, hyperlink)
+      # This is error prone, to_i will silently turn a nil into a 0
+      # and it works by coincidence that Format[0] is general
+      style = cell_xml['s'].to_i   # should be here
+      # c: <c r="A5" s="2">
+      # <v>22606</v>
+      # </c>, format: , tmp_type: float
+      value_type =
+        case cell_xml['t']
+        when 's'
+          :shared
+        when 'b'
+          :boolean
+        # 2011-02-25 BEGIN
+        when 'str'
+          :string
+        # 2011-02-25 END
+        # 2011-09-15 BEGIN
+        when 'inlineStr'
+          :inlinestr
+        # 2011-09-15 END
+        else
+          format = @styles.style_format(style)
+          Excelx::Format.to_type(format)
+        end
+      formula = nil
+      row, column = ::Roo::Utils.split_coordinate(cell_xml['r'])
+      cell_xml.children.each do |cell|
+        case cell.name
+        when 'is'
+          cell.children.each do |inline_str|
+            if inline_str.name == 't'
+              return Excelx::Cell.new(inline_str.content,:string,formula,:string,inline_str.content,style, hyperlink, @workbook.base_date, Excelx::Cell::Coordinate.new(row, column))
+            end
+          end
+        when 'f'
+          formula = cell.content
+        when 'v'
+          if [:time, :datetime].include?(value_type) && cell.content.to_f >= 1.0
+            value_type =
+              if (cell.content.to_f - cell.content.to_f.floor).abs > 0.000001
+                :datetime
+              else
+                :date
+              end
+          end
+          excelx_type = [:numeric_or_formula,format.to_s]
+          value =
+            case value_type
+            when :shared
+              value_type = :string
+              excelx_type = :string
+              @shared_strings[cell.content.to_i]
+            when :boolean
+              (cell.content.to_i == 1 ? 'TRUE' : 'FALSE')
+            when :date, :time, :datetime
+              cell.content
+            when :formula
+              cell.content.to_f
+            when :string
+              excelx_type = :string
+              cell.content
+            else
+              value_type = :float
+              cell.content
+            end
+          return Excelx::Cell.new(value,value_type,formula,excelx_type,cell.content,style, hyperlink, @workbook.base_date, Excelx::Cell::Coordinate.new(row, column))
+        end
+      end
+      Excelx::Cell.new(nil, nil, nil, nil, nil, nil, nil, nil, Excelx::Cell::Coordinate.new(row, column))
+    end
+
+    def extract_hyperlinks(relationships)
+      Hash[doc.xpath("/worksheet/hyperlinks/hyperlink").map do |hyperlink|
+        if hyperlink.attribute('id') && relationship = relationships[hyperlink.attribute('id').text]
+          [::Roo::Utils.ref_to_key(hyperlink.attributes['ref'].to_s), relationship.attribute('Target').text]
+        end
+      end.compact]
+    end
+
+    def expand_merged_ranges(cells)
+      # Extract merged ranges from xml
+      merges = {}
+      doc.xpath("/worksheet/mergeCells/mergeCell").each do |mergecell_xml|
+        tl, br = mergecell_xml['ref'].split(/:/).map {|ref| ::Roo::Utils.ref_to_key(ref)}
+        for row in tl[0]..br[0] do
+          for col in tl[1]..br[1] do
+            next if row == tl[0] && col == tl[1]
+            merges[[row,col]] = tl
+          end
+        end
+      end
+      # Duplicate value into all cells in merged range
+      merges.each do |dst, src|
+        cells[dst] = cells[src]
+      end
+    end
+
+    def extract_cells(relationships)
+      extracted_cells = Hash[doc.xpath("/worksheet/sheetData/row/c").map do |cell_xml|
+        key = ::Roo::Utils.ref_to_key(cell_xml['r'])
+        [key, cell_from_xml(cell_xml, hyperlinks(relationships)[key])]
+      end]
+      if @options[:expand_merged_ranges]
+        expand_merged_ranges(extracted_cells)
+      end
+      extracted_cells
+    end
+
+    def extract_dimensions
+      Roo::Utils.each_element(@path, 'dimension') do |dimension|
+        return dimension.attributes["ref"].value
+      end
+    end
+
+=begin
+Datei xl/comments1.xml
+  <?xml version="1.0" encoding="UTF-8" standalone="yes" ?>
+  <comments xmlns="http://schemas.openxmlformats.org/spreadsheetml/2006/main">
+    <authors>
+      <author />
+    </authors>
+    <commentList>
+      <comment ref="B4" authorId="0">
+        <text>
+          <r>
+            <rPr>
+              <sz val="10" />
+              <rFont val="Arial" />
+              <family val="2" />
+            </rPr>
+            <t>Kommentar fuer B4</t>
+          </r>
+        </text>
+      </comment>
+      <comment ref="B5" authorId="0">
+        <text>
+          <r>
+            <rPr>
+            <sz val="10" />
+            <rFont val="Arial" />
+            <family val="2" />
+          </rPr>
+          <t>Kommentar fuer B5</t>
+        </r>
+      </text>
+    </comment>
+  </commentList>
+  </comments>
+=end
+=begin
+    if @comments_doc[self.sheets.index(sheet)]
+      read_comments(sheet)
+    end
+=end
+  end
+end

data/lib/roo/excelx/styles.rb

@@ -0,0 +1,62 @@
+require 'roo/font'
+require 'roo/excelx/extractor'
+
+module Roo
+  class Excelx::Styles < Excelx::Extractor
+    # convert internal excelx attribute to a format
+    def style_format(style)
+      id = num_fmt_ids[style.to_i]
+      num_fmts[id] || Excelx::Format::STANDARD_FORMATS[id.to_i]
+    end
+
+    def definitions
+      @definitions ||= extract_definitions
+    end
+
+    private
+
+    def num_fmt_ids
+      @num_fmt_ids ||= extract_num_fmt_ids
+    end
+
+    def num_fmts
+      @num_fmts ||= extract_num_fmts
+    end
+
+    def fonts
+     @fonts ||= extract_fonts
+    end
+
+    def extract_definitions
+      doc.xpath("//cellXfs").flat_map do |xfs|
+        xfs.children.map do |xf|
+          fonts[xf['fontId'].to_i]
+        end
+      end
+    end
+
+    def extract_fonts
+      doc.xpath("//fonts/font").map do |font_el|
+        Font.new.tap do |font|
+          font.bold = !font_el.xpath('./b').empty?
+          font.italic = !font_el.xpath('./i').empty?
+          font.underline = !font_el.xpath('./u').empty?
+        end
+      end
+    end
+
+    def extract_num_fmt_ids
+      doc.xpath("//cellXfs").flat_map do |xfs|
+        xfs.children.map do |xf|
+          xf['numFmtId']
+        end
+      end
+    end
+
+    def extract_num_fmts
+      Hash[doc.xpath("//numFmt").map do |num_fmt|
+        [num_fmt['numFmtId'], num_fmt['formatCode']]
+      end]
+    end
+  end
+end

data/lib/roo/excelx/workbook.rb

@@ -0,0 +1,59 @@
+require 'roo/excelx/extractor'
+
+module Roo
+  class Excelx::Workbook < Excelx::Extractor
+    class Label
+      attr_reader :sheet, :row, :col, :name
+
+      def initialize(name, sheet, row, col)
+        @name = name
+        @sheet = sheet
+        @row = row.to_i
+        @col = ::Roo::Utils.letter_to_number(col)
+      end
+
+      def key
+        [@row, @col]
+      end
+    end
+
+    def initialize(path)
+      super
+      if !doc_exists?
+        raise ArgumentError, 'missing required workbook file'
+      end
+    end
+
+    def sheets
+      doc.xpath("//sheet")
+    end
+
+    # aka labels
+    def defined_names
+      Hash[doc.xpath("//definedName").map do |defined_name|
+        # "Sheet1!$C$5"
+        sheet, coordinates = defined_name.text.split('!$', 2)
+        col,row = coordinates.split('$')
+        name = defined_name['name']
+        [name, Label.new(name, sheet,row,col)]
+      end]
+    end
+
+    def base_date
+      @base_date ||=
+        begin
+          # Default to 1900 (minus one day due to excel quirk) but use 1904 if
+          # it's set in the Workbook's workbookPr
+          # http://msdn.microsoft.com/en-us/library/ff530155(v=office.12).aspx
+          result = Date.new(1899,12,30) # default
+          doc.css("workbookPr[date1904]").each do |workbookPr|
+            if workbookPr["date1904"] =~ /true|1/i
+              result = Date.new(1904,01,01)
+              break
+            end
+          end
+          result
+        end
+    end
+  end
+end

data/lib/roo/font.rb

@@ -0,0 +1,17 @@
+module Roo
+  class Font
+    attr_accessor :bold, :italic, :underline
+
+    def bold?
+      @bold
+    end
+
+    def italic?
+      @italic
+    end
+
+    def underline?
+      @underline
+    end
+  end
+end