krikri 0.2.1 → 0.3.1

data/lib/krikri/ldp/nr_source.rb~

@@ -0,0 +1,4 @@
+module Krikri::LDP
+  module NRSource
+  end
+end

data/lib/krikri/ldp/rdf_source.rb~

@@ -0,0 +1,24 @@
+module Krikri::LDP
+  module RdfSource
+    extend ActiveSupport::Concern
+    include Krikri::LDP::Resource
+
+    def save(headers = {})
+      result = super(dump(:ttl))
+      get
+      result
+    end
+
+    def get(*args)
+      result = super
+      reload
+      result
+    end
+
+    def reload
+      return super unless !node? && exists?
+      clear
+      self << RDF::Reader.for(:ttl).new(@http_cache.body).statements
+    end
+  end
+end

data/lib/krikri/ldp/resource.rb~

@@ -0,0 +1,49 @@
+require 'faraday'
+require 'faraday_middleware'
+
+module Krikri::LDP
+  module Resource
+    extend ActiveSupport::Concern
+
+    attr_accessor :rdf_subject, :connection_uri
+
+    def ldp_connection
+      @ldp_conn ||= Faraday.new(connection_uri) do |conn|
+        conn.use Faraday::Response::RaiseError
+        conn.use FaradayMiddleware::FollowRedirects, limit: 3
+        conn.adapter Faraday.default_adapter
+      end
+    end
+
+    def etag
+      http_head.env['response_headers']['etag']
+    end
+
+    def http_head
+      raise 'No subject set for LDP resource.' if rdf_subject.nil?
+      @http_cache ||= ldp_connection.head { |request| request.url rdf_subject }
+    end
+
+    def exists?
+      return true if rdf_subject && http_head
+      false
+    rescue Faraday::ResourceNotFound
+      false
+    end
+
+    def save(uri, body = nil, opts = {})
+      require 'pry'
+      binding.pry
+      create(uri, body, opts)
+    end
+
+    def create(uri, body = nil, opts = {})
+      response = ldp_connection.post do |request|
+        request.url uri
+        request.headers = opts if opts
+        request.body = body
+      end
+      response
+    end
+  end
+end

data/lib/krikri/mapping_dsl.rb~

@@ -0,0 +1,4 @@
+module Krikri
+  module MappingDSL
+  end
+end

data/lib/krikri/mapping_dsl/parser_methods.rb

@@ -48,11 +48,10 @@ module Krikri::MappingDSL
     #   OriginalRecord associated with the parsed record passed as its argument.
     # @todo consider a more generalized approach
     def header
-      lambda do |parsed|
-        raise "#{parsed} does not have a `header`" unless
-          parsed.respond_to? :header
-        parsed.header
-      end
+      RecordProxy.new(
+        [{ name: :header }],
+        Krikri::Parser::ValueArray,
+        true)
     end
 
     ##
@@ -65,7 +64,7 @@ module Krikri::MappingDSL
     #
     # @see Krikri::Parser::ValueArray
     class RecordProxy
-      attr_reader :value_class, :call_chain
+      attr_reader :value_class, :call_chain, :non_root
 
       ##
       # Create a new RecordProxy object.
@@ -76,15 +75,20 @@ module Krikri::MappingDSL
       # @param klass [Class] a Class that acts as the target for delayed method
       #   calls. Must respond to #build(record) and #values. Defaults to
       #   Krikri::Parser::ValueArray
+      # @param non_root [Boolean] a flag indicating whether to build the root
+      #   node of the record before beginning the call chain. The default case
+      #   call `value_class.build`; `false` allows you to skip this step and
+      #   begin the call chain directly on object passed to `#call`.
       #
       # @return [RecordProxy]
-      def initialize(call_chain = [], klass = Krikri::Parser::ValueArray)
+      def initialize(call_chain = [], klass = Krikri::Parser::ValueArray, non_root = false)
         @call_chain = call_chain
         @value_class = klass
+        @non_root = non_root
       end
 
       def dup
-        RecordProxy.new(call_chain.dup, value_class)
+        RecordProxy.new(call_chain.dup, value_class, non_root)
       end
 
       ##
@@ -96,7 +100,7 @@ module Krikri::MappingDSL
       #   value_class#build.
       # @return the values resulting from the full run of the call chain
       def call(record)
-        result = value_class.build(record)
+        result = non_root ? record : value_class.build(record)
         call_chain.each do |message|
           result = result.send(message[:name], *message[:args], &message[:block])
         end

data/lib/krikri/mapping_job.rb~

@@ -0,0 +1,13 @@
+module Krikri
+  class MappingJob < Krikri::Job
+    @queue = :mapping
+
+    private
+
+    def self.run(mapper)
+      require 'pry'
+      binding.pry
+      raise NotImplementedError
+    end
+  end
+end

data/lib/krikri/md5_minter.rb~

@@ -0,0 +1,8 @@
+module Krikri
+  class Md5Minter
+    class << self
+      def mint
+      end
+    end
+  end
+end

data/lib/krikri/parsers/json_parser.rb~

@@ -0,0 +1,17 @@
+module Krikri
+  ##
+  # A parser for JSON records
+  # @see Krikri::Parser
+  class JsonParser < Krikri::Parser
+    def initialize(record)
+      @root = Value.new(JSON.parse(record.content))
+      super
+    end
+
+    class Value < Krikri::Parser::Value
+      def initialize(json)
+        @json = json
+      end
+    end
+  end
+end

data/lib/krikri/parsers/mods_parser.rb

@@ -4,6 +4,8 @@ module Krikri
   # metadata path.
   # @see Krikri::XmlParser
   class ModsParser < XmlParser
+    include Krikri::OaiParserHeaders
+
     def initialize(record, root_path = '//mods:mods', ns = {})
       ns = { mods: 'http://www.loc.gov/mods/v3' }.merge(ns)
       super(record, root_path, ns)

data/lib/krikri/parsers/oai_headers.rb~

@@ -0,0 +1,4 @@
+module Krikri
+  module OaiHeaders
+  end
+end

data/lib/krikri/parsers/oai_parser.rb~

@@ -0,0 +1,2 @@
+module Krikri
+  class XmlParser < Krikri::Parser

data/lib/krikri/parsers/oai_parser_headers.rb~

@@ -0,0 +1,4 @@
+module Krikri
+  module OaiParserHeaders
+  end
+end

data/lib/krikri/parsers/qdc_parser.rb

@@ -7,6 +7,8 @@ module Krikri
   #
   # @see Krikri::XmlParser
   class QdcParser < XmlParser
+    include Krikri::OaiParserHeaders
+
     def initialize(record, root_path = '//oai_qdc:qualifieddc', ns = {})
       ns = {
         qdc: 'http://epubs.cclrc.ac.uk/xmlns/qdc/',

data/lib/krikri/parsers/xml_parser.rb~

@@ -0,0 +1,61 @@
+module Krikri
+  ##
+  # An XmlParser
+  # @see Krikri::Parser
+  class XmlParser < Krikri::Parser
+    def initialize(record, root_path = '/')
+      xml = Nokogiri::XML(record.to_s)
+      ns = namespaces_from_xml(xml)
+      @root = Value.new(xml.at_xpath(root_path, ns), ns)
+      super(record)
+    end
+
+    private
+
+    def namespaces_from_xml(xml)
+      namespaces = xml.collect_namespaces.map do |k, v|
+        [k.gsub('xmlns:', ''), v]
+      end
+      namespaces.to_h
+    end
+
+    ##
+    # An XML Parser Value node class
+    # @see Krikri::Parser::Value
+    class Value < Krikri::Parser::Value
+      def initialize(node, namespaces)
+        node = node.root if node.xml?
+        @node = node
+        @namespaces = namespaces
+      end
+
+      delegate :attributes, :xpath, :at_xpath, :css, :at_css, :to => :@node
+
+      def children
+        @node.element_children.map do |child|
+          return child.name unless child.namespace
+          "#{@namespaces.key(child.namespace.href)}:#{child.name}"
+        end
+      end
+
+      def values
+        select_values.map(&:text)
+      end
+
+      def values?
+        !select_values.empty?
+      end
+
+      private
+
+      def get_child_nodes(name)
+        @node.xpath("#{@node.path}/#{name}", @namespaces)
+          .map { |node| self.class.new(node, @namespaces) }
+      end
+
+      def select_values
+        @node.children.select(&:text?)
+      end
+    end
+  end
+end

data/lib/krikri/provenance_query.rb~

@@ -0,0 +1,4 @@
+module Krikri
+  class ProvenanceQuery
+  end
+end

data/lib/krikri/provenance_query_client.rb~

@@ -0,0 +1,11 @@
+module Krikri
+  class ProvenanceQueryClient < SPARQL::Client
+    def initialize(url = Krikri::Repository, options = {}, &block)
+      super
+    end
+
+    def find_by_activity(activity_uri)
+      self.select(:s).where([:s, RDF::PROV.wasGeneratedBy, activity_uri])
+    end
+  end
+end

data/lib/krikri/util/edtf_converter.rb~

@@ -0,0 +1,26 @@
+module Krikri::Util
+  module ExtendedDateParser
+
+    ##
+    # e.g. 1990s or 199- or 199x
+    def decade(str)
+      decade
+    end
+
+    ##
+    # e.g. 1990s
+    def decade_s(str)
+      /^(\d{3})0s$/.match(str) do |m|
+        Date.edtf("#{m[1]}x")
+      end
+    end
+
+    ##
+    # e.g. 199-
+    def decade_hyphen(str)
+      /^(\d{3})-$/.match(str) do |m|
+        Date.edtf("#{m[1]}x")
+      end
+    end
+  end
+end

data/lib/krikri/util/extended_date_parser.rb~

@@ -0,0 +1,70 @@
+module Krikri::Util
+  module ExtendedDateParser
+    module_function
+
+    def parse(date_str)
+      date = Date.edtf(date_str)
+      date ||= partial_edtf(date_str)
+      date ||= decade_hyphen(date_str)
+      date ||= month_year(date_str)
+      date ||= decade_s(date_str)
+      date ||= hyphenated_partial_range(date_str)
+      date ||= begin
+                 Date.parse(date_str)
+               rescue ArgumentError
+                 nil
+               end
+    end
+
+    def range_match(str)
+      regexp = %r{
+        ([\d\-\.xu\?\~]*\d{4}[\d\-\.xu\?\~]*)
+        \s*[-\.]+\s*
+        ([\d\-\.xu\?\~]*\d{4}[\d\.\-xu\?\~]*)
+      }x
+      regexp.match(str) do |m|
+        [m[1], m[2]]
+      end
+    end
+
+    ##
+    # e.g. 01-2045
+    def month_year(str)
+      /(\d{2})-(\d{4})/.match(str) do |m|
+        Date.edtf("#{m[2]}-#{m[1]}")
+      end
+    end
+
+    ##
+    # e.g. 1990-92
+    def hyphenated_partial_range(str)
+      /^(\d{2})(\d{2})-(\d{2})$/.match(str) do |m|
+        Date.edtf("#{m[1]}#{m[2]}/#{m[1]}#{m[3]}")
+      end
+    end
+
+    ##
+    # e.g. 1970-08-01/02 or 1970-12/10
+    def partial_edtf(str)
+      /^(\d{4}(-\d{2})*)-(\d{2})\/(\d{2})$/.match(str) do |m|
+        Date.edtf("#{m[1]}-#{m[3]}/#{m[1]}-#{m[4]}")
+      end
+    end
+
+    ##
+    # e.g. 1990s
+    def decade_s(str)
+      /^(\d{3})0s$/.match(str) do |m|
+        Date.edtf("#{m[1]}x")
+      end
+    end
+
+    ##
+    # e.g. 199-
+    def decade_hyphen(str)
+      /^(\d{3})-$/.match(str) do |m|
+        Date.edtf("#{m[1]}x")
+      end
+    end
+  end
+end

data/lib/krikri/version.rb

@@ -1,3 +1,3 @@
 module Krikri
-  VERSION = "0.2.1"
+  VERSION = "0.3.1"
 end

data/lib/tasks/jetty.rake

@@ -2,7 +2,7 @@ require 'fileutils'
 require 'jettywrapper'
 
 namespace :jetty do
-  Jettywrapper.url = 'https://github.com/dpla/marmotta-jetty/archive/3.3.0-solr-4.9.0.zip'
+  Jettywrapper.url = 'https://github.com/dpla/marmotta-jetty/archive/3.3.0-solr-4.10.3.zip'
 
   DEFAULT_MARMOTTA = File.expand_path(File.join(Jettywrapper.app_root,
                                                 'jetty',

data/lib/tasks/jetty.rake~

@@ -0,0 +1,12 @@
+require 'fileutils'
+
+namespace :jetty do
+
+  desc 'Configure solr schema'
+  task :solr_config do
+    cp('solr_conf/schema.xml', 'jetty/solr/blacklight-core/conf/schema.xml')
+    cp('solr_conf/solrconfig.xml',
+       'jetty/solr/blacklight-core/conf/solcconfig.xml')
+  end
+
+end

data/lib/tasks/krikri.rake

@@ -11,6 +11,14 @@ require "#{krikri_dir}/app/models/krikri/original_record"
 require "#{krikri_dir}/spec/factories/krikri_original_record"
 
 namespace :krikri do
+
+  def index_aggregation(agg)
+    graph = agg.to_jsonld['@graph'].first
+    indexer = Krikri::IndexService.new
+    indexer.add graph.to_json
+    indexer.commit
+  end
+
   namespace :samples do
 
     # Note: The content of the sample original record does not actually
@@ -32,10 +40,7 @@ namespace :krikri do
       # Save to Marmotta
       agg.save unless agg.exists?
 
-      # Save to Solr
-      graph = agg.to_jsonld['@graph'].first
-      Krikri::IndexService.add graph.to_json
-      Krikri::IndexService.commit
+      index_aggregation(agg)
     end
 
     desc 'Save an invalid sample record to Marmotta and Solr'
@@ -54,10 +59,7 @@ namespace :krikri do
       # Save to Marmotta
       agg.save unless agg.exists?
 
-      # Save to Solr
-      graph = agg.to_jsonld['@graph'].first
-      Krikri::IndexService.add graph.to_json
-      Krikri::IndexService.commit
+      index_aggregation(agg)
     end
 
     desc 'Delete all sample records from Marmotta and Solr'
@@ -79,8 +81,9 @@ namespace :krikri do
       original_record.delete! if original_record.exists?
 
       # Delete all sample records from Solr
-      Krikri::IndexService.delete_by_query 'id:*krikri_sample*'
-      Krikri::IndexService.commit
+      indexer = Krikri::IndexService.new
+      indexer.delete_by_query 'id:*krikri_sample*'
+      indexer.commit
     end
 
     desc 'Save sample institution and harvest source'