reader.rb in dullard-0.2.0

- old
+ new

@@ -1,21 +1,26 @@
 require 'zip/filesystem'
 require 'nokogiri'
 
-module Dullard; end
+module Dullard
+  class Error < StandardError; end
+  OOXMLEpoch = DateTime.new(1899,12,30)
+  SharedStringPath = 'xl/sharedStrings.xml'
+  StylesPath = 'xl/styles.xml'
+end
 
 class Dullard::Workbook
   # Code borrowed from Roo (https://github.com/hmcgowan/roo/blob/master/lib/roo/excelx.rb)
   # Some additional formats added by Paul Hendryx (phendryx@gmail.com) that are common in LibreOffice.
   FORMATS = {
     'general' => :float,
     '0' => :float,
     '0.00' => :float,
     '#,##0' => :float,
     '#,##0.00' => :float,
-    '0%' => :percentage,
-    '0.00%' => :percentage,
+    '0%' => :float,
+    '0.00%' => :float,
     '0.00E+00' => :float,
     '# ?/?' => :float, #??? TODO:
     '# ??/??' => :float, #??? TODO:
     'mm-dd-yy' => :date,
     'd-mmm-yy' => :date,
@@ -76,28 +81,46 @@
     49 => '@',
   }
 
   def initialize(file, user_defined_formats = {})
     @file = file
-    @zipfs = Zip::File.open(@file)
+    begin
+      @zipfs = Zip::File.open(@file)
+    rescue Zip::Error => e
+      raise Dullard::Error, e.message
+    end
     @user_defined_formats = user_defined_formats
     read_styles
   end
 
   def sheets
-    workbook = Nokogiri::XML::Document.parse(@zipfs.file.open("xl/workbook.xml"))
-    @sheets = workbook.css("sheet").each_with_index.map {|n,i| Dullard::Sheet.new(self, n.attr("name"), n.attr("sheetId"), i+1) }
+    begin
+      workbook = Nokogiri::XML::Document.parse(@zipfs.file.open('xl/workbook.xml'))
+    rescue Zip::Error
+      raise Dullard::Error, 'Invalid file, could not open xl/workbook.xml'
+    end
+    @sheets = workbook.css('sheet').each_with_index.map do |n, i|
+      Dullard::Sheet.new(self, n.attr('name'), n.attr('sheetId'), i+1)
+    end
   end
 
   def string_table
-    @string_tabe ||= read_string_table
+    @string_table ||= read_string_table
   end
 
   def read_string_table
-    @string_table = []
+    return [] unless @zipfs.file.exist? Dullard::SharedStringPath
+
+    begin
+      shared_string = @zipfs.file.open(Dullard::SharedStringPath)
+    rescue Zip::Error
+      raise Dullard::Error, 'Invalid file, could not open shared string file.'
+    end
+
     entry = ''
-    Nokogiri::XML::Reader(@zipfs.file.open("xl/sharedStrings.xml")).each do |node|
+    @string_table = []
+    Nokogiri::XML::Reader(shared_string).each do |node|
       if node.name == "si" and node.node_type == Nokogiri::XML::Reader::TYPE_ELEMENT
         entry = ''
       elsif node.name == "si" and node.node_type == Nokogiri::XML::Reader::TYPE_END_ELEMENT
         @string_table << entry
       elsif node.value?
@@ -106,24 +129,33 @@
     end
     @string_table
   end
 
   def read_styles
-    doc = Nokogiri::XML(@zipfs.file.open("xl/styles.xml"))
-    
     @num_formats = {}
     @cell_xfs = []
+    return unless @zipfs.file.exist? Dullard::StylesPath
+
+    begin
+      doc = Nokogiri::XML(@zipfs.file.open(Dullard::StylesPath))
+    rescue Zip::Error
+      raise Dullard::Error, 'Invalid file, could not open styles'
+    end
     
     doc.css('/styleSheet/numFmts/numFmt').each do |numFmt|
-      numFmtId = numFmt.attributes['numFmtId'].value.to_i
-      formatCode = numFmt.attributes['formatCode'].value
-      @num_formats[numFmtId] = formatCode
+      if numFmt.attributes['numFmtId'] && numFmt.attributes['formatCode']
+        numFmtId = numFmt.attributes['numFmtId'].value.to_i
+        formatCode = numFmt.attributes['formatCode'].value
+        @num_formats[numFmtId] = formatCode
+      end
     end
 
     doc.css('/styleSheet/cellXfs/xf').each do |xf|
-      numFmtId = xf.attributes['numFmtId'].value.to_i
-      @cell_xfs << numFmtId
+      if xf.attributes['numFmtId']
+        numFmtId = xf.attributes['numFmtId'].value.to_i
+        @cell_xfs << numFmtId
+      end
     end
   end
 
   
   # Code borrowed from Roo (https://github.com/hmcgowan/roo/blob/master/lib/roo/excelx.rb)
@@ -166,79 +198,94 @@
   def initialize(workbook, name, id, index)
     @workbook = workbook
     @name = name
     @id = id
     @index = index
-    @file = @workbook.zipfs.file.open(path) if @workbook.zipfs.file.exist?(path)
+    begin
+      @file = @workbook.zipfs.file.open(path) if @workbook.zipfs.file.exist?(path)
+    rescue Zip::Error => e
+      raise Dullard::Error, "Couldn't open sheet #{index}: #{e.message}"
+    end
   end
 
   def string_lookup(i)
-    @workbook.string_table[i]
+    @workbook.string_table[i] || (raise Dullard::Error, 'File invalid, invalid string table.')
   end
 
   def rows
     Enumerator.new(row_count) do |y|
       next unless @file
       @file.rewind
       shared = false
       row = nil
+      cell_map = nil # Map of column letter to cell value for a row
       column = nil
       cell_type = nil
       Nokogiri::XML::Reader(@file).each do |node|
         case node.node_type
         when Nokogiri::XML::Reader::TYPE_ELEMENT
           case node.name
           when "row"
-            row = []
-            column = 0
+            cell_map = {}
             next
-          when "c"
-            if node.attributes['t'] != 's' && node.attributes['t'] != 'b'
+          when 'c'
+            node_type = node.attributes['t']
+            cell_index = node.attributes['r']
+            if !cell_index
+              raise Dullard::Error, 'Invalid spreadsheet XML.'
+            end
+
+            if node_type != 's' && node_type != 'b'
               cell_format_index = node.attributes['s'].to_i
               cell_type = @workbook.format2type(@workbook.attribute2format(cell_format_index))
             end
 
-            rcolumn = node.attributes["r"]
-            if rcolumn
-              rcolumn.delete!("0-9")
-              while column < self.class.column_names.size and rcolumn != self.class.column_names[column]
-                row << nil
-                column += 1
-              end
-            end
-            shared = (node.attribute("t") == "s")
-            column += 1
+            column = cell_index.delete('0-9')
+            shared = (node_type == 's')
             next
           end
         when Nokogiri::XML::Reader::TYPE_END_ELEMENT
-          if node.name == "row"
-            y << row
-            next
+          if node.name == 'row'
+            y << process_row(cell_map)
           end
+          next
         end
-        value = node.value
 
-        if value
+        if node.value
+          value = (shared ? string_lookup(value.to_i) : value)
           case cell_type
             when :datetime
             when :time
             when :date
-              value = (DateTime.new(1899,12,30) + value.to_f)
-            when :percentage # ? TODO
+              value = (Dullard::OOXMLEpoch + node.value.to_f)
             when :float
-              value = value.to_f
+              value = node.value.to_f
             else
               # leave as string
           end
           cell_type = nil
-
-          row << (shared ? string_lookup(value.to_i) : value)
+          cell_map[column] = value
         end
       end
     end
   end
 
+  def process_row(cell_map)
+    max = cell_map.keys.map {|c| self.class.column_name_to_index c }.max
+    row = []
+    self.class.column_names[0..max].each do |col|
+      if self.class.column_name_to_index(col) > max
+        break
+      else
+        row << cell_map[col]
+      end
+    end
+    row
+  end
+
+
+
   # Returns A to ZZZ.
   def self.column_names
     if @column_names
       @column_names
     else
@@ -250,22 +297,32 @@
       z = y.map(&proc).flatten
       @column_names = x + y + z
     end
   end
 
+  def self.column_name_to_index(name)
+    if not @column_names_to_indices
+      @column_names_to_indices = {}
+      self.column_names.each_with_index do |name, i|
+        @column_names_to_indices[name] = i
+      end
+    end
+    @column_names_to_indices[name]
+  end
+
   def row_count
     if defined? @row_count
       @row_count
     elsif @file
       @file.rewind
       Nokogiri::XML::Reader(@file).each do |node|
         if node.node_type == Nokogiri::XML::Reader::TYPE_ELEMENT
           case node.name
-          when "dimension"
+          when 'dimension'
             if ref = node.attributes["ref"]
               break @row_count = ref.scan(/\d+$/).first.to_i
             end
-          when "sheetData"
+          when 'sheetData'
             break @row_count = nil
           end
         end
       end
     end