ocr 0.1.0 → 0.2.0

data/README.md

@@ -9,6 +9,8 @@
     * identify: Username and license code as password
     * languages: :brazilian, :byelorussian, :bulgarian, :catalan, :croatian, :czech, :danish, :dutch, :english, :estonian, :finnish, :french, :german, :greek, :hungarian, :indonesian, :italian, :latin, :latvian, :lithuanian, :moldavian, :polish, :portuguese, :romanian, :russian, :serbian, :slovakian, :slovenian, :spanish, :swedish, :turkish, :ukrainian
     * output formats: :doc, :pdf, :excel, :html, :txt, :rtf
+  - [Free OCR online webservice](http://www.free-ocr.co.uk/)
+    * identify: Username
 
 #Installation
 ##From the command line
@@ -32,15 +34,42 @@ gem 'ocr'
 ```
 
 ## Using
+  - Get a OCR: ocr = OCR.use <OCR_NAME>
+  - Set the login Credentials: ocr.login <YOUR_USER>, <YOUR_PASSWORD>, [<EXTRA_LOGIN_DATA>]
+  - Set proxy configuration: ocr.proxy p_addr, p_port = nil, p_user = nil, p_pass = nil
+  - Set image to work: ocr.file= <FILE_NAME_AND_PATH>
+  - Set languaje: ocr.file= <FILE_NAME_AND_PATH>
+  - Set output format: ocr.format= <FORMAT_NAME>
+  - Set output file: ocr.outputfile= <FILE_NAME_AND_PATH>
+  - Test error: error = ocr.error if ocr.error?
+  - Results: text = ocr.text unless ocr.error?
 
 ### OCR Web Service
-  [OCR Web Service](http://www.ocrwebservice.com/)
+  More info at [OCR Web Service](http://www.ocrwebservice.com/).
+
+  Extra properties convert_to_bw=<BOOLEAN>, multi_page_doc=<BOOLEAN>.
 
 ```ruby
   ocr = OCR.use :onlineocr
 
-  ocr.login <YOUR_USER>, <LICENSE_CODE>
-  ocr.file= 'pkg/text_test2.jpg'
+  ocr.login <YOUR_USER>, <YOUR_LICENSE_CODE>
+  ocr.file= 'text_image.jpg'
+  ocr.format= :pdf
+  ocr.outputfile= 'text_doc.pdf'
+  ocr.recognize
+
+  puts "ERROR: #{ocr.error}" if ocr.error?
+  puts "RESULT: #{ocr.text}" unless ocr.error?
+```
+
+### Free OCR online webservice
+  More info at [Free OCR online webservice](http://www.free-ocr.co.uk/).
+
+```ruby
+  ocr = OCR.use :free_ocr
+
+  ocr.login <YOUR_USER_NAME>
+  ocr.file= 'text_image.jpg'
   ocr.recognize
 
   puts "ERROR: #{ocr.error}" if ocr.error?

data/lib/ocr.rb

@@ -2,7 +2,6 @@ require File.expand_path('../ocr/factory', __FILE__)
 require File.expand_path('../ocr/ocr', __FILE__)
 require File.expand_path('../ocrs/dummy', __FILE__)
 require File.expand_path('../ocrs/onlineocr', __FILE__)
-require File.expand_path('../ocrs/onlineocr2', __FILE__)
 require File.expand_path('../ocrs/free_ocr', __FILE__)
 
 

data/lib/ocr/ocr.rb

@@ -7,7 +7,7 @@ module OCR
     attr_reader :text, :error
     attr_accessor :username, :password, :extra_login_data
     attr_accessor :proxy_addr, :proxy_port, :proxy_user, :proxy_pass
-    attr_accessor :file, :language, :format
+    attr_accessor :file, :outputfile, :language, :format
     attr_accessor :debug
 
     def initialize(*args)
@@ -22,6 +22,7 @@ module OCR
       file= false
       lang= false
       format= false
+      outputfile= false
       debug= false
     end
 
@@ -42,6 +43,10 @@ module OCR
       @error != false
     end
 
+    def outputfile?
+      outputfile != false
+    end
+
     def recognize
       clear_error
       clear_text

data/lib/ocr/version.rb

@@ -1,3 +1,3 @@
 module OCR
-  VERSION = "0.1.0"
+  VERSION = "0.2.0"
 end

data/lib/ocrs/free_ocr.rb

@@ -1,6 +1,4 @@
-require 'soap/wsdlDriver'
-require 'base64'
-
+require 'savon'
 
 module OCR
   class Free_ocr < OCR::Ocr
@@ -14,45 +12,24 @@ module OCR
 
     def ocr_recognize
       raise Exception, 'You should set image file' unless @file
-#      request = {
-#        'image' => File.open(@file, 'rb') { |f| [f.read].pack('m*') }
-#      }
-#puts request
-#return
-      #client = Savon::Client.new('http://www.free-ocr.co.uk/ocr.asmx?WSDL')
-#puts client.wsdl.soap_actions
-#return
-      #begin
-      #  response = client.request(:analyze) do
-      #    soap.body = request
-      #  end
-      #rescue
-      #  puts "EX"
-      #end
-
-client = SOAP::WSDLDriverFactory.new( 'http://www.free-ocr.co.uk/ocr.asmx?WSDL' ).create_rpc_driver
-result = client.analyze(:image => File.open(@file, 'rb') { |f| [f.read].pack('m*') })
-
-puts response. analyzeResult
-
-puts result
-return
-
-puts
-puts "BODY: #{response.body}"
+      request = {
+        'image' => File.open(@file, 'rb') { |f| [f.read].pack('m*') }
+      }
+      request[:username] = @username if @username
+
+      client = Savon::Client.new('http://www.free-ocr.co.uk/ocr.asmx?WSDL')
+
+      response = client.request(:analyze) do
+        soap.body = request
+      end
 
       return false if have_error? response.body
-puts 'sigue'
-      set_text !response[:ocr_text].nil? ? response[:ocr_text] : ''
+      set_text response.body[:analyze_response][:analyze_result]
     end
 
     def have_error? response
-      return true && set_error("No response") unless response.has_key?(:ocr_web_service_recognize_response)
-      return true && set_error("No response") unless response[:ocr_web_service_recognize_response].has_key?(:ocrws_response)
-      return true && set_error("No response") unless response[:ocr_web_service_recognize_response].has_key?(:ocrws_response)
-      return false if response[:ocr_web_service_recognize_response][:ocrws_response][:error_message].nil?
-      set_error response[:ocr_web_service_recognize_response][:ocrws_response][:error_message]
-      true
+      return true && set_error("No response") unless response.has_key?(:analyze_response)
+      return true && set_error("No response") unless response[:analyze_response].has_key?(:analyze_result)
     end
   end
 end

data/lib/ocrs/onlineocr.rb

@@ -1,11 +1,9 @@
 require 'savon'
-require 'base64'
-
 
 module OCR
   class Onlineocr < OCR::Ocr
 
-    attr_accessor :convert_to_bw
+    attr_accessor :convert_to_bw, :multi_page_doc
 
     private
     def init
@@ -13,6 +11,7 @@ module OCR
       self.language= :english
       self.format= :txt
       self.convert_to_bw= false
+      self.multi_page_doc= false
     end
 
     def ocr_recognize
@@ -29,8 +28,8 @@ module OCR
           'outputDocumentFormat' => self.format.to_s.upcase,
           'convertToBW' => self.convert_to_bw.to_s,
           'getOCRText' => true.to_s,
-          'createOutputDocument' => false.to_s,
-          'multiPageDoc' => false.to_s,
+          'createOutputDocument' => outputfile?.to_s,
+          'multiPageDoc' => self.multi_page_doc.to_s,
           'ocrWords' => false.to_s
         },
       }
@@ -50,6 +49,12 @@ module OCR
 
       return false if have_error? response.body
 
+      if outputfile?
+        File.open(outputfile, 'w+') {|f|
+          f.puts Base64.decode64(response[:ocr_web_service_recognize_response][:ocrws_response][:file_data])
+        }
+      end
+
       set_text response[:ocr_web_service_recognize_response][:ocrws_response][:ocr_text][:array_of_string][:string]
     end
 

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: ocr
 version: !ruby/object:Gem::Version
-  version: 0.1.0
+  version: 0.2.0
   prerelease: 
 platform: ruby
 authors:
@@ -9,11 +9,11 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2012-02-27 00:00:00.000000000Z
+date: 2012-02-28 00:00:00.000000000Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: savon
-  requirement: &19566960 !ruby/object:Gem::Requirement
+  requirement: &20600860 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -21,10 +21,10 @@ dependencies:
         version: 0.9.9
   type: :runtime
   prerelease: false
-  version_requirements: *19566960
+  version_requirements: *20600860
 - !ruby/object:Gem::Dependency
   name: rspec
-  requirement: &19558440 !ruby/object:Gem::Requirement
+  requirement: &20588220 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -32,7 +32,7 @@ dependencies:
         version: 2.7.0
   type: :development
   prerelease: false
-  version_requirements: *19558440
+  version_requirements: *20588220
 description: Recognize text and characters from image files using web services.
 email:
 - mabarroso@mabarroso.com
@@ -44,7 +44,6 @@ files:
 - lib/ocrs/free_ocr.rb
 - lib/ocrs/onlineocr.rb
 - lib/ocrs/dummy.rb
-- lib/ocrs/onlineocr2.rb
 - lib/ocr/version.rb
 - lib/ocr/ocr.rb
 - lib/ocr/factory.rb

data/lib/ocrs/onlineocr2.rb

@@ -1,109 +0,0 @@
-require 'net/http'
-require 'base64'
-
-
-module OCR
-  class Onlineocr2 < OCR::Ocr
-
-    attr_accessor :convert_to_bw
-
-    private
-    def init
-      super()
-      self.language= :english
-      self.format= :txt
-      self.convert_to_bw= false
-    end
-
-    def ocr_recognize
-      raise Exception, 'You should set username and license as password' unless @username && @password
-      request = {
-        'user_name' => @username,
-        'license_code' => @password,
-        'OCRWSSettings' => {
-#          'ocrLanguages' => [self.language.to_s.upcase],
-          'outputDocumentFormat' => self.format.to_s.upcase,
-          'convertToBW' => self.convert_to_bw,
-          'getOCRText' => true,
-          'createOutputDocument' => false,
-          'multiPageDoc' => false,
-          'ocrWords' => true
-        },
-        'OCRWSInputImage' => {
-          'fileName' => File.basename(@file),
-#          'fileData' => File.open(@file, 'rb') { |f| [f.read].pack('m*') }
-        },
-      }
-#puts request
-#return
-      data =<<EOT
-        <?xml version="1.0" encoding="utf-8"?>
-        <soap:Envelope xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:soap="http://schemas.xmlsoap.org/soap/envelope/">
-          <soap:Body>
-            <OCRWebServiceRecognize xmlns="http://stockservice.contoso.com/wse/samples/2005/10">
-              <user_name>#{@username}</user_name>
-              <license_code>#{@password}</license_code>
-              <OCRWSInputImage>
-                <fileName>string</fileName>
-                <fileData>base64Binary</fileData>
-              </OCRWSInputImage>
-              <OCRWSSetting>
-                <ocrLanguages>#{self.language.to_s.upcase}</ocrLanguages>
-                <outputDocumentFormat>#{self.format.to_s.upcase}</outputDocumentFormat>
-                <convertToBW>#{self.convert_to_bw}</convertToBW>
-                <getOCRText>true</getOCRText>
-                <createOutputDocument>false</createOutputDocument>
-                <multiPageDoc>false</multiPageDoc>
-                <ocrWords>true</ocrWords>
-              </OCRWSSetting>
-            </OCRWebServiceRecognize>
-          </soap:Body>
-        </soap:Envelope>
-EOT
-
-      headers = {
-         'Host' => 'www.ocrwebservice.com',
-         'Content-Type' => 'application/soap+xml; charset=utf-8',
-#         'Content-Type' => 'text/xml',
-         'Content-Length' => "#{data.length}",
-         'SOAPAction' => '"http://stockservice.contoso.com/wse/samples/2005/10/OCRWebServiceRecognize"'
-      }
-
-      host = 'www.ocrwebservice.com'
-      url = '/services/OCRWebService.asmx'
-
-#data = data.squeeze.tr "\n", ''
-puts data
-puts headers
-
-#return
-      http = Net::HTTP.new(host, 80)
-      http.use_ssl = false
-      resp = http.post(url, data, headers)
-
-puts resp
-puts 'Code = ' + resp.code
-puts 'Message = ' + resp.message
-
-
-
-return
-
-puts
-puts "BODY: #{response.body}"
-
-      return false if have_error? response.body
-puts 'sigue'
-      set_text !response[:ocr_text].nil? ? response[:ocr_text] : ''
-    end
-
-    def have_error? response
-      return true && set_error("No response") unless response.has_key?(:ocr_web_service_recognize_response)
-      return true && set_error("No response") unless response[:ocr_web_service_recognize_response].has_key?(:ocrws_response)
-      return true && set_error("No response") unless response[:ocr_web_service_recognize_response].has_key?(:ocrws_response)
-      return false if response[:ocr_web_service_recognize_response][:ocrws_response][:error_message].nil?
-      set_error response[:ocr_web_service_recognize_response][:ocrws_response][:error_message]
-      true
-    end
-  end
-end