colander 0.0.3 → 0.1.0

data/.rvmrc

@@ -1 +1 @@
-rvm --create ruby-1.8.7-p302@email_extractor
+rvm --create 1.9.3-p125@colander

data/colander.gemspec

@@ -19,7 +19,7 @@ Gem::Specification.new do |s|
   s.executables   = `git ls-files -- bin/*`.split("\n").map{ |f| File.basename(f) }
   s.require_paths = ["lib"]
 
-  s.add_dependency "roo", "~> 1.10.1"
   s.add_dependency "zip", "~> 2.0.2"
   s.add_development_dependency "rspec"
+  s.add_development_dependency "pry"
 end

data/lib/colander/parser/base.rb

@@ -8,8 +8,31 @@ module Colander
       end
 
       def parse
+        @emails = collect_emails
+      rescue Exception => e
+        raise InvalidFile.new e
+      end
+
+      def payload
         raise "plz implement me in"
       end
+
+      protected
+
+      def collect_emails
+        parse_file.scan(/\b[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,4}\b/).flatten.uniq
+      end
+
+      def parse_file
+        ic = Iconv.new("UTF-8//IGNORE", "UTF-8")
+        std_out, std_err, exit_status = Open3.capture3("strings", :stdin_data => payload)
+        if exit_status == 0
+          ic.iconv(std_out)
+        else
+          raise RuntimeError.new(std_err)
+        end
+      end
+
     end
   end
-end
+end

data/lib/colander/parser/xls.rb

@@ -1,31 +1,15 @@
 require 'colander/invalid_file'
 require 'colander/parser/base'
-require 'roo'
-require 'iconv'
-
+require 'open3'
 module Colander
   module Parser
     class Xls < Base
-      def parse
-        spreadsheet = parse_file
-        @emails = collect_emails spreadsheet
-      rescue Exception => e
-        raise InvalidFile.new e
-      end
 
       protected
 
-      def parse_file
-        Excel.new(@file_path,nil,:ignore)
+      def payload
+        File.read(@file_path)
       end
-
-      def collect_emails(spreadsheet)
-        spreadsheet.sheets.map do |sheet|
-          spreadsheet.default_sheet = sheet
-          spreadsheet.to_yaml.scan(/\b[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,4}\b/)
-        end.flatten
-      end
-
     end
   end
-end
+end

data/lib/colander/parser/xlsx.rb

@@ -1,13 +1,21 @@
 require 'colander/parser/base'
-require 'roo'
-
+require 'zip'
+require 'iconv'
 module Colander
   module Parser
     class Xlsx < Xls
+
       protected
-      def parse_file
-        Excelx.new(@file_path,nil,:ignore)
+
+      def payload
+        ''.tap do |string|
+          Zip::ZipInputStream::open(@file_path) do |io|
+            while (entry = io.get_next_entry)
+              string << io.read
+            end
+          end
+        end
       end
     end
   end
-end
+end

data/lib/colander/version.rb

@@ -1,3 +1,3 @@
 module Colander
-  VERSION = "0.0.3"
+  VERSION = "0.1.0"
 end

data/spec/spec_helper.rb

@@ -1,3 +1,4 @@
 $:.push File.expand_path("../lib", __FILE__)
 
-require 'colander'
+require 'colander'
+require 'pry'

data/spec/xls_spec.rb

@@ -3,9 +3,8 @@ require 'spec_helper'
 describe Colander::Parser::Xls do
   describe "#parse" do
     it "stores found emails" do
-      Excel.stub(:new)
       parser = Colander::Parser::Xls.new("file/path")
-      parser.should_receive(:collect_emails).and_return(["bruce@wayne.com"])
+      parser.should_receive(:parse_file).and_return("bruce@wayne.com")
       parser.parse
       parser.emails.sort.should eql(["bruce@wayne.com"])
     end
@@ -26,21 +25,19 @@ describe Colander::Parser::Xls do
     end
 
     it "retreives emails from an 95-excel spreadsheet" do
-      pending "handle encoding error"
       parser = Colander::Parser::Xls.new("spec/fixtures/excel95.xls")
       parser.parse
-      parser.emails.should eql(["markus.nordin@mynewsdesk.com", "markus@hej.se", "sven@bertil.se", "Adam.A@hotmail.com", "apa@elabs.se", "liam@neeson.net", "david@mynewsdesk.com"])
+      parser.emails.sort.should eql(["markus.nordin@mynewsdesk.com", "markus@hej.se", "sven@bertil.se", "Adam.A@hotmail.com", "apa@elabs.se", "liam@neeson.net", "david@mynewsdesk.com"].sort)
     end
     it "retreives emails from an xls spreadsheet" do
       parser = Colander::Parser::Xls.new("spec/fixtures/old-format.xls")
       parser.parse
-      parser.emails.should eql(["markus.nordin@mynewsdesk.com", "markus@hej.se", "sven@bertil.se", "Adam.A@hotmail.com", "apa@elabs.se", "liam@neeson.net", "david@mynewsdesk.com"])
+      parser.emails.sort.should eql(["markus.nordin@mynewsdesk.com", "markus@hej.se", "sven@bertil.se", "Adam.A@hotmail.com", "apa@elabs.se", "liam@neeson.net", "david@mynewsdesk.com"].sort)
     end
     it "retreives emails from an 95-excel spreadsheet without file suffix" do
-      pending "handle encoding error"
       parser = Colander::Parser::Xls.new("spec/fixtures/excel95-without-file-suffix")
       parser.parse
-      parser.emails.should eql(["markus.nordin@mynewsdesk.com", "markus@hej.se", "sven@bertil.se", "Adam.A@hotmail.com", "apa@elabs.se", "liam@neeson.net", "david@mynewsdesk.com"])
+      parser.emails.sort.should eql(["markus.nordin@mynewsdesk.com", "markus@hej.se", "sven@bertil.se", "Adam.A@hotmail.com", "apa@elabs.se", "liam@neeson.net", "david@mynewsdesk.com"].sort)
     end
   end
 end

data/spec/xlsx_spec.rb

@@ -8,7 +8,7 @@ describe Colander::Parser::Xlsx do
     it "retreives emails from an xlsx spreadsheet" do
       parser = Colander::Parser::Xlsx.new("spec/fixtures/new-format.xlsx")
       parser.parse
-      parser.emails.should eql(["markus.nordin@mynewsdesk.com", "markus@hej.se", "sven@bertil.se", "Adam.A@hotmail.com", "apa@elabs.se", "liam@neeson.net", "david@mynewsdesk.com"])
+      parser.emails.sort.should eql(["markus.nordin@mynewsdesk.com", "markus@hej.se", "sven@bertil.se", "Adam.A@hotmail.com", "apa@elabs.se", "liam@neeson.net", "david@mynewsdesk.com"].sort)
     end
   end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: colander
 version: !ruby/object:Gem::Version
-  version: 0.0.3
+  version: 0.1.0
   prerelease: 
 platform: ruby
 authors:
@@ -10,33 +10,33 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2012-03-01 00:00:00.000000000 Z
+date: 2012-03-21 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
-  name: roo
-  requirement: &70276138479340 !ruby/object:Gem::Requirement
+  name: zip
+  requirement: &70203123886720 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
       - !ruby/object:Gem::Version
-        version: 1.10.1
+        version: 2.0.2
   type: :runtime
   prerelease: false
-  version_requirements: *70276138479340
+  version_requirements: *70203123886720
 - !ruby/object:Gem::Dependency
-  name: zip
-  requirement: &70276138467840 !ruby/object:Gem::Requirement
+  name: rspec
+  requirement: &70203123886060 !ruby/object:Gem::Requirement
     none: false
     requirements:
-    - - ~>
+    - - ! '>='
       - !ruby/object:Gem::Version
-        version: 2.0.2
-  type: :runtime
+        version: '0'
+  type: :development
   prerelease: false
-  version_requirements: *70276138467840
+  version_requirements: *70203123886060
 - !ruby/object:Gem::Dependency
-  name: rspec
-  requirement: &70276138466580 !ruby/object:Gem::Requirement
+  name: pry
+  requirement: &70203123885380 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -44,7 +44,7 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *70276138466580
+  version_requirements: *70203123885380
 description: Exctract an array of emails from various file formats (xls, xlsx)
 email:
 - dev@mynewsdesk.com