excel2csv 0.0.2 → 0.0.3

data/lib/excel2csv/version.rb

@@ -1,3 +1,3 @@
 module Excel2CSV
-  VERSION = "0.0.2"
+  VERSION = "0.0.3"
 end

data/lib/excel2csv.rb

@@ -7,6 +7,7 @@ module Excel2CSV
 
   class Info
     attr_accessor :sheets
+    attr_accessor :previews
     attr_accessor :tmp_dir
     attr_accessor :working_dir
 
@@ -17,12 +18,21 @@ module Excel2CSV
     end
 
     def read
-      @sheets = Dir["#{@working_dir}/*.csv"].map do |file|
-        {path: file}
+      Dir["#{@working_dir}/*.csv"].map do |file|
+        name = File.basename(file)
+        m = /(?<sheet>\d+)-(?<rows>\d+)(-of-(?<total_rows>\d+))?/.match(name)
+        next if !m
+        total_rows   = (m[:total_rows] || m[:rows]).to_i
+        preview_rows = m[:rows].to_i
+        if name =~ /preview/
+          @previews << {path: file, total_rows:total_rows, rows:preview_rows}
+        else
+          @sheets << {path: file, total_rows:total_rows, rows:total_rows}
+        end
       end
     end 
 
-    def close
+    def clean
       FileUtils.remove_entry_secure(@tmp_dir, true) if @tmp_dir
     end
 
@@ -32,13 +42,14 @@ module Excel2CSV
       @working_dir = working_dir
       @tmp_dir = tmp_dir
       @sheets = []
+      @previews = []
     end
 
   end
   
   def foreach(path, options = {}, &block) 
     convert(path, options) do |info|
-      CSV.foreach(info.sheets.first[:path], options, &block)
+      CSV.foreach(path_to_sheet(info, options), clean_options(options), &block)
     end 
   end
 
@@ -46,30 +57,83 @@ module Excel2CSV
 
   def read(path, options = {})
     convert(path, options) do |info|
-      CSV.read(info.sheets.first[:path], options)
+      CSV.read(path_to_sheet(info, options), clean_options(options))
     end
   end
 
   module_function :read
 
   def convert(path, options = {})
+    info = options[:info]
+    if info && Dir.exists?(info.working_dir)
+      return block_given? ? yield(info) : info
+    end
     begin
-      tmp_dir = Dir.mktmpdir
-      dest_folder = options[:dest_folder] || tmp_dir
-      java_options = options[:java_options] || "-Dfile.encoding=utf8 -Xms512m -Xmx512m -XX:MaxPermSize=256m"
-      jar_path = File.join(File.dirname(__FILE__), "excel2csv.jar")
-      `java #{java_options} -jar #{jar_path} #{path} #{dest_folder}`
-      info = Info.read dest_folder, tmp_dir
+      info = create_cvs_files(path, options)
       if block_given?
         yield info
       else
         info
       end
     ensure
-      info.close if block_given?
+      info.clean if block_given? && info
     end
   end
 
   module_function :convert
 
+  def path_to_sheet(info, options = {})
+    if options[:preview]
+      options.delete(:encoding) # all previews are in utf-8
+      collection = info.previews
+    else
+      collection = info.sheets
+    end
+    index = (idx = options[:index]) ? idx : 0
+    collection[index][:path]
+  end
+
+  module_function :path_to_sheet
+
+  def create_cvs_files(path, options)
+    tmp_dir = Dir.mktmpdir
+    dest_folder = options[:dest_folder] || tmp_dir
+    limit = options[:rows_limit]
+    
+    if path =~ /\.csv/
+      total_rows = 0
+      preview_rows = []
+      opts = clean_options(options)
+      CSV.foreach(path, opts) do |row| 
+        if limit && total_rows <= limit
+          preview_rows << row
+        end
+        total_rows += 1
+      end
+      FileUtils.cp path, "#{dest_folder}/1-#{total_rows}.csv"
+      if limit
+        CSV.open("#{dest_folder}/1-#{limit}-of-#{total_rows}-preview.csv", "wb") do |csv|
+          preview_rows.each {|row| csv << row}
+        end
+      end
+    else
+      java_options = options[:java_options] || "-Dfile.encoding=utf8 -Xms512m -Xmx512m -XX:MaxPermSize=256m"
+      rows_limit = limit ? "-r #{limit}" : ""
+      jar_path = File.join(File.dirname(__FILE__), "excel2csv.jar")
+      `java #{java_options} -jar #{jar_path} #{rows_limit} #{path} #{dest_folder}`
+    end
+    
+    Info.read(dest_folder, tmp_dir)
+  end
+
+  module_function :create_cvs_files
+
+  def clean_options options
+    options.dup.delete_if do |key, value|
+      [:dest_folder, :java_options, :preview, :index, :path, :rows_limit, :info].include?(key)
+    end
+  end
+
+  module_function :clean_options
+
 end

data/spec/excel2csv_spec.rb

@@ -1,27 +1,31 @@
 #encoding: utf-8
+
 require 'excel2csv'
 
 describe Excel2CSV do
 
   let(:excel) {Excel2CSV}
+  let(:csv_basic_types)  {"spec/fixtures/basic_types.csv"}
+  let(:xls_basic_types)  {"spec/fixtures/basic_types.xls"}
+  let(:xlsx_basic_types) {"spec/fixtures/basic_types.xlsx"}
 
   it "reads xls files" do
-    data = excel.read "spec/fixtures/basic_types.xls"
+    data = excel.read xls_basic_types
     data[0].should == ["1.00", "2011-12-23 21:00:00 UTC(+0000)", "Hello"]
     data[1].should == ["2.00", "2011-12-24 21:00:00 UTC(+0000)", "Привет"]
     data[2].should == ["3.00", "2011-12-25 21:00:00 UTC(+0000)", 'Привет, "я excel!"']
   end
 
   it "reads xlsx files" do
-    data = excel.read "spec/fixtures/basic_types.xlsx"
+    data = excel.read xlsx_basic_types
     data[0].should == ["1.00", "2011-12-23 21:00:00 UTC(+0000)", "Hello"]
     data[1].should == ["2.00", "2011-12-24 21:00:00 UTC(+0000)", "Привет"]
     data[2].should == ["3.00", "2011-12-25 21:00:00 UTC(+0000)", 'Привет, "я excel!"']
   end
 
-  it "iterates rows" do
+  it "iterates rows like CSV lib" do
     count = 0
-    excel.foreach "spec/fixtures/basic_types.xls" do |row|
+    excel.foreach xls_basic_types do |row|
       row.length.should == 3
       count += 1
     end
@@ -30,7 +34,7 @@ describe Excel2CSV do
 
   it "removes tmp dir after work" do
     tmp_dir = nil
-    excel.convert "spec/fixtures/basic_types.xlsx" do |info|
+    excel.convert xlsx_basic_types do |info|
       # puts IO.read(info.sheets.first[:path])
       tmp_dir = info.tmp_dir
     end
@@ -38,4 +42,56 @@ describe Excel2CSV do
     Dir.exists?(tmp_dir).should == false
   end
 
+  it "converts once if info is passed" do
+    info = excel.convert xlsx_basic_types
+    info.sheets.length.should == 1
+    info.previews.length.should == 0
+    info.should == excel.convert("spec/fixtures/basic_types.xlsx", info:info)
+  end
+
+  it "regenerate csv files if working_dir is removed" do
+    info = excel.convert xlsx_basic_types
+    info.clean
+    info.should_not == excel.convert(xlsx_basic_types, info:info)
+  end
+
+  it "generates preview csv files with rows limit" do
+    info = excel.convert xls_basic_types, rows_limit:1
+    info.sheets.length.should == 1
+    info.previews.length.should == 1
+
+    info.sheets.first[:total_rows].should == 3
+    info.previews.first[:total_rows].should == 3
+
+    info.sheets.first[:rows].should == 3
+    info.previews.first[:rows].should == 1
+  end
+
+  it "reads previews" do
+    data = excel.read(xls_basic_types, rows_limit:1, preview:true, index:0)
+    data.length.should == 1
+    data[0].should == ["1.00", "2011-12-23 21:00:00 UTC(+0000)", "Hello"]
+  end
+
+  it "reads csv files" do
+    data = excel.read(csv_basic_types, encoding:'windows-1251:utf-8')
+    data[0].should == ["1.00","12/24/11 12:00 AM","Hello"]
+    data[1].should == ["2.00","12/25/11 12:00 AM","Привет"]
+    data[2].should == ["3.00","12/26/11 12:00 AM",'Привет, "я excel!"']
+  end
+
+  it "reads csv files with preview" do
+    data = excel.read(csv_basic_types,
+      encoding:   'windows-1251:utf-8',
+      rows_limit: 2,
+      preview:    true
+    )
+    data[0].should == ["1.00","12/24/11 12:00 AM","Hello"]
+    data[1].should == ["2.00","12/25/11 12:00 AM","Привет"]
+  end
+
+  # Date, Boolean, String, [Phone, Percent, Email, Gender, Url]
+  # Несколько телефонов, 
+
+
 end

data/spec/fixtures/basic_types.csv

@@ -0,0 +1,3 @@
+1.00,12/24/11 12:00 AM,Hello
+2.00,12/25/11 12:00 AM,������
+3.00,12/26/11 12:00 AM,"������, ""� excel!"""

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: excel2csv
 version: !ruby/object:Gem::Version
-  version: 0.0.2
+  version: 0.0.3
   prerelease: 
 platform: ruby
 authors:
@@ -9,11 +9,11 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2011-09-15 00:00:00.000000000 Z
+date: 2011-09-20 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rspec
-  requirement: &70366534136320 !ruby/object:Gem::Requirement
+  requirement: &70270996778580 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - <=
@@ -21,7 +21,7 @@ dependencies:
         version: '2.6'
   type: :development
   prerelease: false
-  version_requirements: *70366534136320
+  version_requirements: *70270996778580
 description: gem for converting Excel files to csv
 email:
 - yury.korolev@gmail.com
@@ -38,6 +38,7 @@ files:
 - lib/excel2csv.rb
 - lib/excel2csv/version.rb
 - spec/excel2csv_spec.rb
+- spec/fixtures/basic_types.csv
 - spec/fixtures/basic_types.xls
 - spec/fixtures/basic_types.xlsx
 - spec/spec_helper.rb
@@ -67,6 +68,7 @@ specification_version: 3
 summary: extract excel worksheets to csv files
 test_files:
 - spec/excel2csv_spec.rb
+- spec/fixtures/basic_types.csv
 - spec/fixtures/basic_types.xls
 - spec/fixtures/basic_types.xlsx
 - spec/spec_helper.rb