jobparser 0.7.2 → 0.8.0

data/lib/jobparser/parseschema.rb

@@ -1,9 +1,16 @@
 # encoding: utf-8
 module JobParser
   class ParseSchema < Parser
+
+    def initialize(html, from_url)
+      Parser::ACCEPTED_ELEMENTS.push("span")
+      super(html, from_url)
+    end
     def job
       res = super
       res[:schema] = true
+      res[:benefits] = job_benefits
+      res[:education_requirements] = job_education_requirements
       res
     end
     private
@@ -23,21 +30,38 @@ module JobParser
       Facets::Apply.new(@doc, @url, @plain_text).parse
     end
 
+    def job_benefits
+      get_content_at_prop("benefits")
+    end
+
+    def job_education_requirements
+      get_content_at_prop("educationRequirements")
+    end
+
     def job_salary_string
       get_content_at_prop("baseSalary")
     end
 
     def job_location
-      # some sites don't use the address stuff properly
-      if is_content_at_prop?("addressLocality")
-        get_content_at_prop("addressLocality")
+      if @doc.css("*").any? { |elem| elem['itemtype'] == "http://schema.org/PostalAddress" }
+        fields = %w{ streetAddress addressLocality addressRegion addressCountry postalCode }
+        address = []
+        fields.each do |field|
+          content = get_content_at_prop(field)
+          address.push(content) unless content.empty?
+        end
+        address.join(", ")
       else
-        get_content_at_prop("jobLocation")
+        # some sites don't use the address stuff properly
+        if is_content_at_prop?("addressLocality")
+          get_content_at_prop("addressLocality")
+        else
+          get_content_at_prop("jobLocation")
+        end
       end
     end
 
     def deadline
-
     end
 
     def does_use_schema?
@@ -49,15 +73,15 @@ module JobParser
     def get_content_at_prop(prop)
       elem = find_with_itemprop(prop)
       if elem
-        Cleaner.strip_all_white_space(find_with_itemprop(prop).content)
+        Cleaner.strip_all_white_space(elem.content)
       else
         ""
       end
     end
 
     def is_content_at_prop?(prop)
-      elem = find_with_itemprop("prop")
-      elem && !elem.empty?
+      elem = find_with_itemprop(prop)
+      elem && !elem.content.empty?
     end
 
     def find_with_itemprop(prop)

data/lib/jobparser/version.rb

@@ -1,3 +1,3 @@
 module JobParser
-  VERSION = "0.7.2"
+  VERSION = "0.8.0"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: jobparser
 version: !ruby/object:Gem::Version
-  version: 0.7.2
+  version: 0.8.0
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-08-06 00:00:00.000000000 Z
+date: 2013-08-07 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler