Monkeypatch of SolrEad to try to make indexing faster

arclight_monkeypatch.rb

require "URI"
require 'solr_ead'

class SolrEad::Indexer

  def additional_component_fields(node, addl_fields = Hash.new)

    p_ids        = parent_id_list(node)
    p_unittitles = parent_unittitle_list(node)

    addl_fields["id"]                                                        = [eadid(node), node.attr("id")].join
    addl_fields[Solrizer.solr_name("ead", :stored_sortable)]                 = eadid(node)
    addl_fields[Solrizer.solr_name("parent", :stored_sortable)]              = node.parent.attr("id") unless node.parent.attr("id").nil?
    addl_fields[Solrizer.solr_name("parent", :displayable)]                  = p_ids
    addl_fields[Solrizer.solr_name("parent_unittitles", :displayable)]       = p_unittitles
    addl_fields[Solrizer.solr_name("parent_unittitles", :searchable)]        = p_unittitles
    addl_fields[Solrizer.solr_name("component_level", :type => :integer)]    = p_ids.length + 1
    addl_fields[Solrizer.solr_name("component_children", :type => :boolean)] = component_children?(node)
    addl_fields[Solrizer.solr_name("collection", :facetable)]                = collection(node)
    addl_fields[Solrizer.solr_name("collection", :displayable)]              = collection(node)
    addl_fields[Solrizer.solr_name("repository", :facetable)]                = repository(node)
    addl_fields[Solrizer.solr_name("repository", :displayable)]              = repository(node)
    addl_fields
  end

  # can these be made to use absolute xpaths?
  def repository(node)
    @cached_repo ||= node.xpath("/ead/archdesc/did/repository").text.strip
  end

  def collection(node)
    @cached_collection ||= node.xpath("/ead/archdesc/did/unittitle").text
  end

  def eadid(node)
    @cached_eadid ||= node.xpath("/ead/eadheader/eadid").text
  end

  # Reset the caches when we get a new file
  alias_method :original_components, :components

  def components(*args)
    STDERR.puts "****** Using the monkeypatch on file #{args}"
    @memtitle          = nil
    @cached_eadid      = nil
    @cached_collection = nil
    @cached_repo       = nil
    original_components(*args)
  end

  def parent_unittitle_list(node, results = ::Array.new)
    while node.parent.name == "c"
      parent = node.parent
      results << get_title(parent)
      node = parent
    end
    results.reverse
  end

  def get_title(node)
    @memtitle ||= Hash.new {|h, node| h[node.object_id] = _get_title(node)}
    @memtitle[node]
  end

  def _get_title(node)
    title = node.at_xpath("./did/unittitle")
    date  = node.at_xpath("./did/unitdate")
    if !title.nil? and !title.content.empty?
      return ead_to_html(title.content)
    elsif !date.nil? and !date.content.empty?
      return ead_to_html(date.content)
    else
      return "[No title available]"
    end
  end

end


##### Testing #########
# Only run this code if *this* file is run, not if it's
# required/loaded from another file.
if __FILE__==$0
  
  class SolrEad::Indexer
    public :om_component_from_node
  end
  
  require 'benchmark'
  ENV["SOLR_URL"] = "http://not.gonna.happen/solr"
  filename        = "umich-bhl-87365-2.xml"
  indexer         = SolrEad::Indexer.new

  x = Benchmark.measure do
    # Mimic making a solr document, without sending it to solr.
    solr_doc = SolrEad::Document.from_xml(File.new(filename))
    indexer.components(filename).each_with_index do |node, counter|
      solr_doc = indexer.om_component_from_node(node).to_solr(indexer.additional_component_fields(node))
      solr_doc.merge!({Solrizer.solr_name("sort", :sortable, :type => :integer) => counter.to_s})
    end
  end

  puts x
end