puree 0.20.0 → 1.0.0

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: puree
 version: !ruby/object:Gem::Version
-  version: 0.20.0
+  version: 1.0.0
 platform: ruby
 authors:
 - Adrian Albin-Clark
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2017-02-03 00:00:00.000000000 Z
+date: 2017-03-15 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: http
@@ -38,10 +38,24 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '1.6'
-description: Consumes the Pure Research Information System API and puts the metadata
-  into simple data structures.
-email:
-- a.albin-clark@lancaster.ac.uk
+- !ruby/object:Gem::Dependency
+  name: rspec
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+description: |-
+  Fetches metadata from the Pure Research Information System and
+                            extracts it into Ruby data models.
+email: a.albin-clark@lancaster.ac.uk
 executables: []
 extensions: []
 extra_rdoc_files: []
@@ -54,36 +68,85 @@ files:
 - README.md
 - Rakefile
 - lib/puree.rb
-- lib/puree/collection.rb
-- lib/puree/configuration.rb
-- lib/puree/dataset.rb
-- lib/puree/date.rb
-- lib/puree/download.rb
-- lib/puree/event.rb
-- lib/puree/journal.rb
-- lib/puree/map.rb
-- lib/puree/organisation.rb
-- lib/puree/person.rb
-- lib/puree/project.rb
-- lib/puree/publication.rb
-- lib/puree/publisher.rb
-- lib/puree/resource.rb
-- lib/puree/server.rb
+- lib/puree/api/api.rb
+- lib/puree/api/authentication.rb
+- lib/puree/api/configuration.rb
+- lib/puree/api/map.rb
+- lib/puree/api/request.rb
+- lib/puree/extractor/collection.rb
+- lib/puree/extractor/dataset.rb
+- lib/puree/extractor/download.rb
+- lib/puree/extractor/event.rb
+- lib/puree/extractor/extractor.rb
+- lib/puree/extractor/journal.rb
+- lib/puree/extractor/organisation.rb
+- lib/puree/extractor/person.rb
+- lib/puree/extractor/project.rb
+- lib/puree/extractor/publication.rb
+- lib/puree/extractor/publisher.rb
+- lib/puree/extractor/resource.rb
+- lib/puree/extractor/server.rb
+- lib/puree/model/address.rb
+- lib/puree/model/copyright_license.rb
+- lib/puree/model/dataset.rb
+- lib/puree/model/download_header.rb
+- lib/puree/model/endeavour_person.rb
+- lib/puree/model/event.rb
+- lib/puree/model/event_header.rb
+- lib/puree/model/file.rb
+- lib/puree/model/helper/validation.rb
+- lib/puree/model/journal.rb
+- lib/puree/model/legal_condition.rb
+- lib/puree/model/link.rb
+- lib/puree/model/model.rb
+- lib/puree/model/organisation.rb
+- lib/puree/model/organisation_header.rb
+- lib/puree/model/person.rb
+- lib/puree/model/person_name.rb
+- lib/puree/model/project.rb
+- lib/puree/model/publication.rb
+- lib/puree/model/publication_status.rb
+- lib/puree/model/publisher.rb
+- lib/puree/model/related_content_header.rb
+- lib/puree/model/resource.rb
+- lib/puree/model/server.rb
+- lib/puree/model/spatial_point.rb
+- lib/puree/model/structure.rb
+- lib/puree/model/temporal_range.rb
+- lib/puree/util/date.rb
+- lib/puree/util/util.rb
 - lib/puree/version.rb
+- lib/puree/xml_extractor/base.rb
+- lib/puree/xml_extractor/collection.rb
+- lib/puree/xml_extractor/dataset.rb
+- lib/puree/xml_extractor/download.rb
+- lib/puree/xml_extractor/event.rb
+- lib/puree/xml_extractor/journal.rb
+- lib/puree/xml_extractor/organisation.rb
+- lib/puree/xml_extractor/person.rb
+- lib/puree/xml_extractor/project.rb
+- lib/puree/xml_extractor/publication.rb
+- lib/puree/xml_extractor/publisher.rb
+- lib/puree/xml_extractor/resource.rb
+- lib/puree/xml_extractor/server.rb
+- lib/puree/xml_extractor/shared.rb
+- lib/puree/xml_extractor/xml_extractor.rb
 - puree.gemspec
-- spec/collection_spec.rb
-- spec/dataset_spec.rb
-- spec/download_spec.rb
-- spec/event_spec.rb
-- spec/journal_spec.rb
-- spec/organisation_spec.rb
-- spec/person_spec.rb
-- spec/project_spec.rb
-- spec/publication_spec.rb
-- spec/publisher_spec.rb
-- spec/server_spec.rb
+- spec/download_http_spec.rb
+- spec/open_api_dataset_http_spec.rb
+- spec/resource/collection_all_http_spec.rb
+- spec/resource/collection_http_spec.rb
+- spec/resource/dataset_http_spec.rb
+- spec/resource/event_http_spec.rb
+- spec/resource/journal_http_spec.rb
+- spec/resource/organisation_http_spec.rb
+- spec/resource/person_http_spec.rb
+- spec/resource/project_http_spec.rb
+- spec/resource/publication_http_spec.rb
+- spec/resource/publisher_http_spec.rb
+- spec/server_http_spec.rb
 - spec/spec_helper.rb
-homepage: https://aalbinclark.gitbooks.io/puree
+homepage: https://github.com/lulibrary/puree
 licenses:
 - MIT
 metadata: {}
@@ -106,18 +169,19 @@ rubyforge_project:
 rubygems_version: 2.2.2
 signing_key: 
 specification_version: 4
-summary: A client for the Pure Research Information System API.
+summary: Metadata extraction from the Pure Research Information System.
 test_files:
-- spec/collection_spec.rb
-- spec/dataset_spec.rb
-- spec/download_spec.rb
-- spec/event_spec.rb
-- spec/journal_spec.rb
-- spec/organisation_spec.rb
-- spec/person_spec.rb
-- spec/project_spec.rb
-- spec/publication_spec.rb
-- spec/publisher_spec.rb
-- spec/server_spec.rb
+- spec/download_http_spec.rb
+- spec/open_api_dataset_http_spec.rb
+- spec/resource/collection_all_http_spec.rb
+- spec/resource/collection_http_spec.rb
+- spec/resource/dataset_http_spec.rb
+- spec/resource/event_http_spec.rb
+- spec/resource/journal_http_spec.rb
+- spec/resource/organisation_http_spec.rb
+- spec/resource/person_http_spec.rb
+- spec/resource/project_http_spec.rb
+- spec/resource/publication_http_spec.rb
+- spec/resource/publisher_http_spec.rb
+- spec/server_http_spec.rb
 - spec/spec_helper.rb
-has_rdoc: 

data/lib/puree/collection.rb

@@ -1,285 +0,0 @@
-module Puree
-
-  # Collection of resources
-  #
-  class Collection
-
-    # @param resource [Symbol]
-    # @param base_url [String]
-    # @param username [String]
-    # @param password [String]
-    # @param basic_auth [Boolean]
-    def initialize(resource: nil,
-                   base_url: nil,
-                   username: nil,
-                   password: nil,
-                   basic_auth: nil)
-      @resource_type = resource
-      @api_map = Puree::Map.new.get
-      @base_url = base_url.nil? ? Puree.base_url : base_url
-      @basic_auth = basic_auth.nil? ? Puree.basic_auth : basic_auth
-      if @basic_auth === true
-        @username = username.nil? ? Puree.username : username
-        @password = password.nil? ? Puree.password : password
-      end
-      @uuids = []
-    end
-
-    # Gets an array of objects of resource type specified in constructor
-    #
-    # @param limit [Integer]
-    # @param offset [Integer]
-    # @param created_start [String]
-    # @param created_end [String]
-    # @param modified_start [String]
-    # @param modified_end [String]
-    # @param full [Boolean]
-    # @param instance [Boolean]
-    # @return [Array<Object>]
-    # @return [Array<Resource subclass>]
-    def get(
-            limit:            20,
-            offset:           0,
-            created_start:    nil,
-            created_end:      nil,
-            modified_start:   nil,
-            modified_end:     nil,
-            full:             true,
-            instance:         false,
-            rendering:        :xml_long
-    )
-
-      @options = {
-          basic_auth:       @basic_auth,
-          latest_api:       true,
-          resource_type:    @resource_type.to_sym,
-          rendering:        :system,
-          limit:            limit,
-          offset:           offset,
-          created_start:    created_start,
-          created_end:      created_end,
-          modified_start:   modified_start,
-          modified_end:     modified_end,
-          full:             full,
-          instance:         instance,
-          record_rendering: rendering
-      }
-
-      reset
-
-      missing = missing_credentials
-      if !missing.empty?
-        missing.each do |m|
-          puts "#{self.class.name}" + '#' + "#{__method__} missing #{m}"
-        end
-        exit
-      end
-
-      # strip any trailing slash
-      @base_url = @base_url.sub(/(\/)+$/, '')
-
-      headers = {}
-      headers['Accept'] = 'application/xml'
-
-      if @options[:basic_auth] === true
-        @auth = Base64::strict_encode64(@username + ':' + @password)
-        headers['Authorization'] = 'Basic ' + @auth
-      end
-
-      query = {}
-
-      query['rendering'] = @options[:rendering]
-
-      if @options[:limit] >= 0
-        query['window.size'] = @options[:limit]
-      end
-
-      if @options[:offset]
-        query['window.offset'] = @options[:offset]
-      end
-
-      if @options[:created_start]
-        query['createdDate.fromDate'] = @options[:created_start]
-      end
-
-      if @options[:created_end]
-        query['createdDate.toDate'] = @options[:created_end]
-      end
-
-      if @options[:modified_start]
-        query['modifiedDate.fromDate'] = @options[:modified_start]
-      end
-
-      if @options[:modified_end]
-        query['modifiedDate.toDate'] = @options[:modified_end]
-      end
-
-      if @options['rendering']
-        query['rendering'] = @options['rendering']
-      end
-
-      begin
-        url = build_url
-        req = HTTP.headers accept: 'application/xml'
-        if @options[:basic_auth]
-          req = req.auth headers['Authorization']
-        end
-        @response = req.get(url, params: query)
-        @doc = Nokogiri::XML @response.body
-        @doc.remove_namespaces!
-
-        @count = extract_count
-
-      rescue HTTP::Error => e
-        puts 'HTTP::Error '+ e.message
-      end
-
-      if @options[:full]
-        collect_resource
-      else
-        data = []
-        uuid.each do |u|
-          o = {}
-          o['uuid'] = u
-          data << o
-        end
-        data
-      end
-
-    end
-
-
-    # Count of records available for a resource type
-    #
-    # @return [Integer]
-    def count
-      @count ||= get_count
-    end
-
-    private
-
-    def get_count
-      find limit: 0
-      extract_count
-    end
-
-    def extract_count
-      path = '//count'
-      xpath_query_for_single_value(path).to_i
-    end
-
-
-    # Array of UUIDs
-    #
-    # @return [Array<String>]
-    def uuid
-      collect_uuid
-      @uuids
-    end
-
-
-    def collect_resource
-      data = []
-      resource_class = 'Puree::' + @resource_type.to_s.capitalize
-
-      # if @options[:basic_auth] === true
-      #   r = Object.const_get(resource_class).new base_url:   @base_url,
-      #                                            username:   @username,
-      #                                            password:   @password,
-      #                                            basic_auth: true
-      # else
-      #   r = Object.const_get(resource_class).new base_url:   @base_url
-      # end
-      # whitelist symbol
-      if @api_map[:resource_type].has_key?(@resource_type)
-        uuid.each do |u|
-          if @options[:basic_auth] === true
-            r = Object.const_get(resource_class).new base_url:   @base_url,
-                                                     username:   @username,
-                                                     password:   @password,
-                                                     basic_auth: true
-          else
-            r = Object.const_get(resource_class).new base_url:   @base_url
-          end
-          record = r.find uuid: u,
-                          rendering:  @options[:record_rendering]
-          # puts JSON.pretty_generate( record, :indent => '  ')
-          # p u
-          if @options[:instance]
-            data << r
-          else
-            # just the data
-            data << record
-          end
-        end
-        data
-      else
-        puts 'Invalid resource class'
-        exit
-      end
-    end
-
-    def collect_uuid
-      @uuids = []
-      path = '//renderedItem/@renderedContentUUID'
-      xpath_result = xpath_query path
-      xpath_result.each { |i| @uuids << i.text.strip }
-    end
-
-    def service_name
-      resource_type = @options[:resource_type]
-      @api_map[:resource_type][resource_type][:service]
-    end
-
-    def service_response_name
-      resource_type = @options[:resource_type]
-      @api_map[:resource_type][resource_type][:response]
-    end
-
-    def build_url
-      service = service_name
-      if @options[:latest_api] === false
-        service_api_mode = service
-      else
-        service_api_mode = service + '.current'
-      end
-      @base_url + '/' + service_api_mode
-    end
-
-    def xpath_query(path)
-      @doc.xpath path
-    end
-
-    def xpath_query_for_single_value(path)
-      xpath_result = xpath_query path
-      xpath_result ? xpath_result.text.strip : ''
-    end
-
-
-    def missing_credentials
-      missing = []
-      if @base_url.nil?
-        missing << 'base_url'
-      end
-
-      if @options[:basic_auth] === true
-        if @username.nil?
-          missing << 'username'
-        end
-        if @password.nil?
-          missing << 'password'
-        end
-      end
-
-      missing
-    end
-
-    def reset
-      @response = nil
-      @count = nil
-    end
-
-    alias :find :get
-
-  end
-end