ooxl 0.0.1.4.24 → 0.0.1.5.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 8eb9194a8f21b6a1ff814c7307581b87039174af
-  data.tar.gz: 19327c4f196fa7be33de042dd02a5637855d9326
+  metadata.gz: 72289ba7eb6dba7ef0e3feec1b329f59d5ed9547
+  data.tar.gz: 568bfdd6c97d43698e665eac15f0e453fac151e3
 SHA512:
-  metadata.gz: 2fa656c463f7278a8ba232e890eca33137a65b6065a0936b90a4017c3efaab77c2e4b9921fcc780b80912b55b2dcfd332a43a0ebd4bab2fea8bc38e256c8385f
-  data.tar.gz: e8ef1b7a8c5a9705690ba0b512f550faa2c785a87f796a0b85f7fce407559ddcdbacb6695345ad58503f1c3daf0206133b687c6e05ad9f26820d07517d451d29
+  metadata.gz: 4a8c5d321345da254a3c397dccc77d090435408f895c7b16b2877d6d24f59721c9cb1caffd4e39337e8ebec02495f15e3c0133f0e0faa6d58cb5f602ee419951
+  data.tar.gz: e4e7aca211e24fb842b959b431850671191aed6292f7f5ef7cd8c31267564dde3ab467c9aca19351f2f3012c1d24e7b9cee6a8a231510b4073b1b7b077ef480e

data/lib/ooxl/ooxl.rb

@@ -33,9 +33,11 @@ class OOXL
   end
 
   def sheet(sheet_name)
-    sheet_index = @workbook.sheets.index { |sheet| sheet[:name] == sheet_name}
-    raise "No #{sheet_name} in workbook." if sheet_index.nil?
-    sheet = @sheets.fetch((sheet_index+1).to_s)
+    sheet_meta = @workbook.sheets.find { |sheet| sheet[:name] == sheet_name }
+    raise "No #{sheet_name} in workbook." if sheet_meta.nil?
+
+    sheet_index = sheet_meta[:sheet_id]
+    sheet = @sheets.fetch(sheet_index)
 
     # shared variables
     sheet.name = sheet_name
@@ -73,8 +75,7 @@ class OOXL
   end
 
   def fetch_comments(sheet_index)
-    final_sheet_index = sheet_index+1
-    relationship = @relationships[final_sheet_index.to_s]
+    relationship = @relationships[sheet_index]
     @comments[relationship.comment_id] if relationship.present?
   end
 

data/lib/ooxl/util.rb

@@ -1,12 +1,13 @@
-class OOXL
+class OOXL
   module Util
-    COLUMN_LETTERS = ('A'..'ZZZZ').to_a
-    def letter_equivalent(index)
-      COLUMN_LETTERS.fetch(index)
+    COLUMN_LETTERS = [nil] + ('A'..'ZZZZ').to_a
+
+    def letter_index(col_letter)
+      column_letter_to_number(col_letter) - 1
     end
 
-    def letter_index(letter)
-      COLUMN_LETTERS.index { |c_letter| c_letter == letter}
+    def letter_equivalent(col_index)
+      column_number_to_letter(col_index + 1)
     end
 
     def to_column_letter(reference)
@@ -14,13 +15,17 @@ class OOXL
     end
 
     def uniform_reference(ref)
-      ref.to_s[/[A-Z]/] ? letter_index(ref) + 1 : ref
+      ref.to_s[/[A-Z]/] ? column_letter_to_number(ref) : ref
     end
 
     def node_value_extractor(node)
       node.try(:value)
     end
 
+    def column_number_to_letter(index)
+      COLUMN_LETTERS.fetch(index)
+    end
+
     def column_letter_to_number(column_letter)
       pow = column_letter.length - 1
       result = 0

data/lib/ooxl/version.rb

@@ -1,3 +1,3 @@
 class OOXL
-  VERSION = "0.0.1.4.24"
+  VERSION = "0.0.1.5.3"
 end

data/lib/ooxl/xl_objects/row.rb

@@ -10,20 +10,18 @@ class OOXL
     end
 
     def [](id)
-      cell = if id.is_a?(String)
-        cells.find { |row| row.id == id}
+      if id.is_a?(String)
+        cell(id)
       else
-        cells[id]
+        cells[id] || BlankCell.new(id)
       end
-      (cell.present?) ? cell : BlankCell.new(id)
     end
 
     def cells
       if @options[:padded_cells]
         unless @cells.blank?
           'A'.upto(@cells.last.column).map do |column_letter|
-            cell = @cells.find { |cell| cell.column == column_letter}
-            (cell.blank?) ? BlankCell.new("#{column_letter}#{id}") : cell
+            cell(column_letter)
           end
         end
       else
@@ -32,8 +30,14 @@ class OOXL
     end
 
     def cell(cell_id)
-      cell_final_id = cell_id[/[A-Z]{1,}\d+/] ? cell_id : "#{cell_id}#{id}"
-      cells.find { |cell| cell.id == cell_final_id}
+      cell_final_id = cell_id[/[A-Z]+\d+/] ? cell_id : "#{cell_id}#{id}"
+      cell_id_map[cell_final_id]
+    end
+
+    def cell_id_map
+      @cell_id_map ||= cells.each_with_object(Hash.new { |_, k| BlankCell.new(k) }) do |cell, result|
+        result[cell.id] = cell
+      end
     end
 
     def each
@@ -41,12 +45,16 @@ class OOXL
     end
 
     def self.load_from_node(row_node, shared_strings, styles, options)
-      new(id: row_node.attributes["r"].try(:value),
+      new(id: extract_id(row_node),
           spans: row_node.attributes["spans"].try(:value),
           height: row_node.attributes["ht"].try(:value),
           cells: row_node.xpath('c').map {  |cell_node| OOXL::Cell.load_from_node(cell_node, shared_strings, styles)},
           options: options )
     end
+
+    def self.extract_id(row_node)
+      row_node.attributes["r"].try(:value)
+    end
   end
 end
 

data/lib/ooxl/xl_objects/row_cache.rb

@@ -0,0 +1,128 @@
+class OOXL
+  class RowCache
+    include Enumerable
+
+    attr_accessor :styles
+
+    # built on-demand -- use rows instead
+    attr_reader :row_cache
+
+    # built on-demand -- use fetch_row_by_id instead
+    attr_reader :row_id_map
+
+    delegate :size, to: :row_nodes
+
+    def initialize(sheet_xml, shared_strings, options = {})
+      @shared_strings = shared_strings
+      @sheet_xml = sheet_xml
+      @options = options
+      @row_cache = []
+      @row_id_map = {}
+    end
+
+    def [](id)
+      fetch_row_by_id(id)
+    end
+
+    alias_method :row, :[]
+
+    def each(&block)
+      if @options[:padded_rows]
+        padded_rows(&block)
+      else
+        rows(&block)
+      end
+    end
+
+    def rows(&block)
+      # track yield count to know if caller broke out of loop
+      rows_yielded = 0
+      row_cache.each do |r|
+        yield r if block_given?
+        rows_yielded += 1
+      end
+
+      if !all_rows_loaded? && rows_yielded == row_cache.count
+        parse_more_rows(&block)
+      end
+
+      row_cache
+    end
+
+    def row_range(start_index, end_index)
+      return enum_for(:row_range, start_index, end_index) unless block_given?
+
+      rows do |row|
+        row_id = row.id.to_i
+        next if row_id < start_index
+        break if row_id > end_index
+
+        yield row
+      end
+    end
+
+    def max_row_index
+      return 0 if row_nodes.empty?
+
+      if all_rows_loaded?
+        row_cache.last.id.to_i
+      else
+        Row.extract_id(row_nodes.last).to_i
+      end
+    end
+
+    private
+
+    def parse_more_rows
+      row_nodes.drop(row_cache.count).each do |row_node|
+        row = parse_row(row_node)
+        row_cache << row
+        row_id_map[row.id] = row
+        yield row if block_given?
+      end
+    end
+
+    def all_rows_loaded?
+      row_cache.count == row_nodes.count
+    end
+
+    def fetch_row_by_id(row_id)
+      row_id = row_id.to_s
+      return row_id_map[row_id] if all_rows_loaded? || row_id_map.key?(row_id)
+
+      parse_more_rows do |row|
+        return row if row.id == row_id
+      end
+
+      nil
+    end
+
+    def padded_rows
+      real_rows_yielded = 0
+      yielded_rows = []
+      (1..Float::INFINITY).each do |row_index|
+        row = row(row_index)
+        if row.blank?
+          row = Row.new(id: row_index.to_s, cells: [])
+        else
+          real_rows_yielded += 1
+        end
+
+        yielded_rows << row
+        yield row if block_given?
+
+        break if real_rows_yielded == row_cache.count && all_rows_loaded?
+      end
+      yielded_rows
+    end
+
+    def row_nodes
+      @row_nodes ||= @sheet_xml.xpath('//sheetData/row')
+    end
+
+    def parse_row(row_node)
+      Row.load_from_node(row_node, @shared_strings, @styles, @options)
+    end
+  end
+end
+

data/lib/ooxl/xl_objects/sheet.rb

@@ -1,10 +1,14 @@
 require_relative 'sheet/data_validation'
+require_relative 'row_cache'
+
 class OOXL
   class Sheet
     include OOXL::Util
     include Enumerable
-    attr_reader :columns, :data_validations, :shared_strings
-    attr_accessor :comments, :styles, :defined_names, :name
+
+    attr_reader :columns, :data_validations, :shared_strings, :styles
+    attr_accessor :comments, :defined_names, :name
+    delegate :[], :each, :rows, :row, to: :@row_cache
 
     def initialize(xml, shared_strings, options={})
       @xml = Nokogiri.XML(xml).remove_namespaces!
@@ -12,8 +16,8 @@ class OOXL
       @comments = {}
       @defined_names = {}
       @styles = []
-      @loaded_cache = {}
       @options = options
+      @row_cache = RowCache.new(@xml, @shared_strings, options)
     end
 
     def code_name
@@ -41,70 +45,35 @@ class OOXL
       end
     end
 
-    def [](id)
-      if id.is_a?(String)
-        rows.find { |row| row.id == id}
-      else
-        rows[id]
-      end
-    end
 
-    def row(index, stream: false)
-      if @loaded_cache[:rows] || !stream
-        rows.find { |row| row.id == index.to_s}
-      else
-        found_row = nil
-        rows do |row|
-          if row.id == index.to_s
-            found_row = row
-            break
-          end
-        end
-        found_row
-      end
+    # DEPRECATED: stream is no longer separate
+    def stream_row(index)
+      row(index)
     end
 
     # test mode
     def cells_by_column(column_letter)
-      columns = []
-      rows.each do |row|
-        columns << row.cells.find { |cell| to_column_letter(cell.id) == column_letter}
+      rows.map do |row|
+        row.cells.find { |cell| to_column_letter(cell.id) == column_letter}
       end
-      columns
     end
 
-    def cell(cell_id, stream: false)
-      column_letter, row_index = cell_id.partition(/\d+/)
-      current_row = row(row_index, stream: stream)
-      current_row.cell(column_letter) unless current_row.nil?
+    def last_column(row_index=1)
+      @last_column ||= {}
+      @last_column[row_index] ||= begin
+        cells = row(row_index).try(:cells) 
+        cells.last.column if cells.present?
+      end
     end
 
-    def formula(cell_id, stream: false)
-      cell(cell_id, stream: stream).try(:formula)
+    def cell(cell_id)
+      column_letter, row_index = cell_id.partition(/\d+/)
+      current_row = row(row_index)
+      current_row.cell(column_letter) unless current_row.nil?
     end
 
-    def rows
-      @rows ||= begin
-        all_rows = @xml.xpath('//sheetData/row').map do |row_node|
-          row = Row.load_from_node(row_node, @shared_strings, @styles, @options)
-          yield row if block_given?
-          row
-        end
-        @loaded_cache[:rows] = true
-        all_rows
-      end
-    end
-
-    def each
-      if @options[:padded_rows]
-        last_row_index = rows.last.id.to_i
-        (1.upto(last_row_index)).each do |row_index|
-          row = row(row_index)
-          yield (row.blank?) ? Row.new(id: "#{row_index}", cells: []) : row
-        end
-      else
-        rows  { |row| yield row }
-      end
+    def formula(cell_id)
+      cell(cell_id).try(:formula)
     end
 
     def font(cell_reference)
@@ -115,7 +84,6 @@ class OOXL
       cell(cell_reference).try(:fill)
     end
 
-
     def data_validations
       @data_validations ||= begin
 
@@ -134,6 +102,11 @@ class OOXL
       end
     end
 
+    def styles=(styles)
+      @styles = styles
+      @row_cache.styles = styles
+    end
+
     # a shortcut for:
     # formula =  data_validation('A1').formula
     # ooxl.named_range(formula)
@@ -143,7 +116,7 @@ class OOXL
         if data_validation.formula[/[\s\$\,\:]/]
           (data_validation.formula[/\$/].present?) ? "#{name}!#{data_validation.formula}" : data_validation.formula
         else
-          @defined_names.fetch(data_validation.formula)
+          @defined_names[data_validation.formula]
         end
       end
     end
@@ -155,57 +128,62 @@ class OOXL
       # cell_range values separated by comma
       if cell_range.include?(":")
         cell_letters = cell_range.gsub(/[\d]/, '').split(':')
-        start_index, end_index = cell_range[/[A-Z]{1,}\d+/] ? cell_range.gsub(/[^\d:]/, '').split(':').map(&:to_i) : [1, rows.size]
-        # This will allow values from this pattern
-        # 'SheetName!A1:C3'
-        # The number after the cell letter will be the index
-        # 1 => start_index
-        # 3 => end_index
-        # Expected output would be: [['value', 'value', 'value'], ['value', 'value', 'value'], ['value', 'value', 'value']]
+        start_index, end_index = cell_range[/[A-Z]{1,}\d+/] ? cell_range.gsub(/[^\d:]/, '').split(':').map(&:to_i) : [1, @row_cache.max_row_index]
         if cell_letters.uniq.size > 1
-          start_index.upto(end_index).map do  |row_index|
-            (letter_index(cell_letters.first)..letter_index(cell_letters.last)).map do |cell_index|
-                row = fetch_row_by_id(row_index.to_s)
-                next if row.blank?
-
-                cell_letter = letter_equivalent(cell_index)
-                row["#{cell_letter}#{row_index}"].value
-            end
-          end
+          list_values_from_rectangle(cell_letters, start_index, end_index)
         else
-          cell_letter = cell_letters.uniq.first
-          (start_index..end_index).to_a.map do |row_index|
-            row = fetch_row_by_id(row_index.to_s)
-            next if row.blank?
-            row["#{cell_letter}#{row_index}"].value
-          end
+          list_values_from_column(cell_letters.uniq.first, start_index, end_index)
         end
       else
         # when only one value: B2
-        row_index = cell_range.gsub(/[^\d:]/, '').split(':').map(&:to_i).first
-        row = fetch_row_by_id(row_index.to_s)
-        return if row.blank?
-        [row[cell_range].value]
+        list_values_from_cell(cell_range)
       end
     end
     alias_method :list_values_from_formula, :list_values_from_cell_range
 
-    def self.load_from_stream(xml_stream, shared_strings)
-      self.new(Nokogiri.XML(xml_stream).remove_namespaces!, shared_strings)
+    # This will allow values from this pattern
+    # 'SheetName!A1:C3'
+    # The number after the cell letter will be the index
+    # 1 => start_index
+    # 3 => end_index
+    # Expected output would be: [['value', 'value', 'value'], ['value', 'value', 'value'], ['value', 'value', 'value']]
+    def list_values_from_rectangle(cell_letters, start_index, end_index)
+      start_col = column_letter_to_number(cell_letters.first)
+      end_col = column_letter_to_number(cell_letters.last)
+      @row_cache.row_range(start_index, end_index).map do |row|
+        (start_col..end_col).map do |col_index|
+          col_letter = column_number_to_letter(col_index)
+          row["#{col_letter}#{row.id}"].value
+        end
+      end
+    end
+
+    def list_values_from_column(column_letter, start_index, end_index)
+      @row_cache.row_range(start_index, end_index).map do |row|
+        row["#{column_letter}#{row.id}"].value
+      end
+    end
+
+    def list_values_from_cell(cell_ref)
+      row_index = cell_ref.gsub(/[^\d:]/, '').split(':').map(&:to_i).first
+      row = row(row_index)
+      return if row.blank?
+      [row[cell_ref].value]
     end
 
     def in_merged_cells?(cell_id)
       column_letter, column_index = cell_id.partition(/\d+/)
-      range = merged_cells_range.find { |column_range, index_range| column_range.cover?(column_letter) && index_range.cover?(column_index) }
+      range = merged_cells.find { |column_range, index_range| column_range.cover?(column_letter) && index_range.cover?(column_index) }
       range.present?
     end
 
-    private
-    def fetch_row_by_id(row_id)
-      rows.find { |row| row.id == row_id.to_s}
+    def self.load_from_stream(xml_stream, shared_strings)
+      self.new(xml_stream, shared_strings)
     end
 
-    def merged_cells_range
+    private
+
+    def merged_cells
       @merged_cells ||= @xml.xpath('//mergeCells/mergeCell').map do |merged_cell|
         # <mergeCell ref="Q381:R381"/>
         start_reference, end_reference = merged_cell.attributes["ref"].try(:value).split(':')

data/ooxml_excel.gemspec

@@ -26,9 +26,10 @@ Gem::Specification.new do |spec|
   spec.require_paths = ["lib"]
   spec.add_dependency 'activesupport'
   spec.add_dependency 'nokogiri', '~> 1'
-  spec.add_dependency 'rubyzip', '~> 1.0', '< 2.0.0'
+  spec.add_dependency 'rubyzip', '~> 1.3.0', '< 2.0.0'
 
-  spec.add_development_dependency "bundler", "~> 1.12"
+  spec.add_development_dependency "bundler"
+  spec.add_development_dependency "pry-byebug"
   spec.add_development_dependency "rake", "~> 10.0"
   spec.add_development_dependency "rspec", "~> 3.0"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: ooxl
 version: !ruby/object:Gem::Version
-  version: 0.0.1.4.24
+  version: 0.0.1.5.3
 platform: ruby
 authors:
 - James Mones
-autorequire: 
+autorequire:
 bindir: exe
 cert_chain: []
-date: 2017-02-28 00:00:00.000000000 Z
+date: 2020-10-01 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: activesupport
@@ -44,7 +44,7 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.0'
+        version: 1.3.0
     - - "<"
       - !ruby/object:Gem::Version
         version: 2.0.0
@@ -54,7 +54,7 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.0'
+        version: 1.3.0
     - - "<"
       - !ruby/object:Gem::Version
         version: 2.0.0
@@ -62,16 +62,30 @@ dependencies:
   name: bundler
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ">="
       - !ruby/object:Gem::Version
-        version: '1.12'
+        version: '0'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ">="
       - !ruby/object:Gem::Version
-        version: '1.12'
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: pry-byebug
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
 - !ruby/object:Gem::Dependency
   name: rake
   requirement: !ruby/object:Gem::Requirement
@@ -129,6 +143,7 @@ files:
 - lib/ooxl/xl_objects/number_formatting.rb
 - lib/ooxl/xl_objects/relationships.rb
 - lib/ooxl/xl_objects/row.rb
+- lib/ooxl/xl_objects/row_cache.rb
 - lib/ooxl/xl_objects/sheet.rb
 - lib/ooxl/xl_objects/sheet/data_validation.rb
 - lib/ooxl/xl_objects/styles.rb
@@ -138,7 +153,7 @@ files:
 homepage: https://github.com/halcjames/ooxl
 licenses: []
 metadata: {}
-post_install_message: 
+post_install_message:
 rdoc_options: []
 require_paths:
 - lib
@@ -153,9 +168,9 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubyforge_project: 
-rubygems_version: 2.6.8
-signing_key: 
+rubyforge_project:
+rubygems_version: 2.6.12
+signing_key:
 specification_version: 4
 summary: OOXL Excel - Parse Excel Spreadsheets (xlsx, xlsm).
 test_files: []