jobparser 0.13.8 → 0.13.9

data/lib/jobparser/cache/mongostore.rb

@@ -4,51 +4,46 @@ module JobParser
     class MongoStore
 
       def has_cache_for_url?(url)
-        Job.where(:url => url).count > 0
+        job_for_url(url).count > 0
       end
 
       def store(hash)
-        Job.where(:url => hash[:url]).delete
-        hash = strip_job_fields(hash)
+        job_for_url(hash[:url]).delete
+        hash = strip_fields_not_stored(hash)
         Job.create(hash)
       end
 
       def cache_expired?(url)
-        job = Job.where(:url => url).first
+        job = job_for_url(url).first
         expire_time = (job.created_at + JobParser.config[:cache_expire])
         Time.now > expire_time
       end
 
       def get(url)
-        job = Job.where(:url => url).first
-        new_obj = {}
-        job.attributes.each { |k, v|
-          unless %w{created_at _id updated_at}.include?(k)
-            new_obj[k.to_sym] = v
+        job = job_for_url(url).first
+        {}.tap do |job_obj|
+          job.attributes.each do |k, v|
+            job_obj[k.to_sym] = v unless %w{created_at _id updated_at}.include?(k)
           end
-        }
-        new_obj[:from_cache] = true
-        new_obj
+          job_obj[:from_cache] = true
+        end
       end
 
       def clear_all
-        MongoStore::Job.each do |job|
-          job.delete
-        end
+        MongoStore::Job.each(&:delete)
       end
 
       def view_cache
-        res = []
-        Job.each do |job|
-          res.push({
-            :url => job.url,
-            :created => job.created_at
-          })
+        [].tap do |res|
+          Job.each do |job|
+            res.push({
+              :url => job.url,
+              :created => job.created_at
+            })
+          end
         end
-        res
       end
 
-
       class Job
         include Mongoid::Document
         include Mongoid::Timestamps
@@ -57,6 +52,7 @@ module JobParser
           benefits education_requirements incentives industry
           occupational_category qualifications responsibilities skills special_commitments work_hours
         }
+
         EXTRA_SCHEMA_TEXT_FIELDS.each do |f|
           field f.to_sym, :type => String
         end
@@ -68,20 +64,22 @@ module JobParser
         field :salary_string, :type => String
         field :location,      :type => String
         field :deadline,      :type => String
-        field :postcode,  :type => String
+        field :postcode,      :type => String
         field :schema,        :type => Boolean, :default => false
-
       end
 
       private
 
-      def strip_job_fields(hash)
-        new_hash = {}
-        excluded_fields = [:from_cache]
-        hash.each { |k, v| new_hash[k] = v unless excluded_fields.include?(k) }
-        new_hash
+      def job_for_url(url)
+        Job.where(:url => url)
       end
 
+      def strip_fields_not_stored(hash)
+        {}.tap do |new_hash|
+          excluded_fields = [:from_cache]
+          hash.each { |k, v| new_hash[k] = v unless excluded_fields.include?(k) }
+        end
+      end
     end
   end
 end

data/lib/jobparser/cache/textfile.rb

@@ -4,20 +4,17 @@ module JobParser
   class Cache
     class TextFile
       def has_cache_for_url?(url)
-        path = path_for_url(url)
-        File.exist?(path)
+        File.exist?(path_for_url(url))
       end
 
       def store(job_hash)
-        url = job_hash[:url]
-        write_to_file(path_for_url(url), job_hash.to_json)
+        write_to_file(path_for_url(job_hash[:url]), job_hash.to_json)
       end
 
       def get(url)
         path = path_for_url(url)
         obj = JSON.parse(IO.read(path))
-        sym_obj = {}
-        obj.each { |k, v| sym_obj[k.to_sym] = v }
+        sym_obj = make_object_keys_symbols(obj)
         sym_obj[:from_cache] = true
         sym_obj
       end
@@ -33,19 +30,25 @@ module JobParser
       end
 
       def view_cache
-        res = []
-        cache_files.each do |f|
-          contents = JSON.parse(IO.read(f))
-          res.push({
-            :url => contents["url"],
-            :created => File.mtime(f)
-          })
+        [].tap do |res|
+          cache_files.each do |f|
+            contents = JSON.parse(IO.read(f))
+            res.push({
+              :url => contents["url"],
+              :created => File.mtime(f)
+            })
+          end
         end
-        res
       end
 
       private
 
+      def make_object_keys_symbols(obj)
+        {}.tap do |sym_obj|
+          obj.each { |k, v| sym_obj[k.to_sym] = v }
+        end
+      end
+
       def cache_files
         Dir[File.join(JobParser.config[:cache_location], "*.txt")]
       end

data/lib/jobparser/parsehtml.rb

@@ -10,31 +10,31 @@ module JobParser
     private
 
     def job_location
-      Facets::Location.new(@doc, @url, @plain_text).parse
+      Facets::Location.new(*facet_args).parse
     end
 
     def job_salary_string
-      Facets::SalaryString.new(@doc, @url, @plain_text).parse
+      Facets::SalaryString.new(*facet_args).parse
     end
 
     def job_salary
-      Facets::Salary.new(@doc, @url, @plain_text).parse
+      Facets::Salary.new(*facet_args).parse
     end
 
     def job_title
-      Facets::Title.new(@doc, @url, @pplain_text).parse
+      Facets::Title.new(*facet_args).parse
     end
 
     def apply_link
-      Facets::Apply.new(@doc, @url, @plain_text).parse
+      Facets::Apply.new(*facet_args).parse
     end
 
     def deadline
-      Facets::Deadline.new(@doc, @url, @plain_text).parse
+      Facets::Deadline.new(*facet_args).parse
     end
 
     def job_postcode
-      Facets::Postcode.new(@doc, @url, @plain_text).parse
+      Facets::Postcode.new(*facet_args).parse
     end
   end
 end

data/lib/jobparser/parser.rb

@@ -6,6 +6,8 @@ module JobParser
     def initialize(html, from_url)
       @url = from_url
       @html = html
+      @doc = strip_bad_elements(Nokogiri::HTML(@html))
+      @plain_text = get_plain_text
     end
 
     def job
@@ -13,10 +15,7 @@ module JobParser
         return JobParser.cache.fetch_result_for_url(@url)
       end
 
-      @doc = strip_bad_elements(Nokogiri::HTML(@html))
-      @plain_text = get_plain_text
-
-      result = { :url => @url,
+      { :url => @url,
         :salary => job_salary,
         :title => job_title,
         :apply => apply_link,
@@ -25,11 +24,14 @@ module JobParser
         :deadline => deadline,
         :postcode => job_postcode
       }
-      result
     end
 
     private
 
+    def facet_args
+      [@doc, @url, @plain_text]
+    end
+
     def cache(result)
       if JobParser.config[:cache_on]
         store_result_to_cache(result)

data/lib/jobparser/version.rb

@@ -1,3 +1,3 @@
 module JobParser
-  VERSION = "0.13.8"
+  VERSION = "0.13.9"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: jobparser
 version: !ruby/object:Gem::Version
-  version: 0.13.8
+  version: 0.13.9
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-08-22 00:00:00.000000000 Z
+date: 2013-08-23 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler