Class: Solrizer::Solrizer

Inherits:
Object
  • Object
show all
Defined in:
lib/solrizer.rb

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(opts = {}) ⇒ Solrizer

This method initializes the indexer If passed an argument of :index_full_text=>true, it will perform full-text indexing instead of indexing fields only.



21
22
23
24
25
26
27
28
29
 
# File 'lib/solrizer.rb', line 21

def initialize( opts={} )
  @@index_list = false unless defined?(@@index_list)
  if opts[:index_full_text] == true || opts[:index_full_text] == "true"
    @index_full_text = true 
  else
    @index_full_text = false 
  end
  @indexer = Indexer.new( :index_full_text=>@index_full_text )
end

Instance Attribute Details

#index_full_text ⇒ Object

Returns the value of attribute index_full_text.



15
16
17
 
# File 'lib/solrizer.rb', line 15

def index_full_text
  @index_full_text
end

#indexer ⇒ Object

Returns the value of attribute indexer.



15
16
17
 
# File 'lib/solrizer.rb', line 15

def indexer
  @indexer
end

Instance Method Details

#solrize(obj) ⇒ Object

This method solrizes the given Fedora object’s full-text and facets into the search index



34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
 
# File 'lib/solrizer.rb', line 34

def solrize( obj )
  # retrieve the Fedora object based on the given unique id
    
    begin
    
    start = Time.now
    print "Retrieving object #{obj} ..."

    case obj
    when ActiveFedora::Base
      # do nothing
    when Fedora::FedoraObject
      obj = Repository.get_object( obj.pid )
    when String
      obj = Repository.get_object( obj )
    else
      raise "you must pass either a ActiveFedora::Base, Fedora::RepositoryObject, or a String.  You submitted a #{obj.class}"
    end
    
    # obj = obj.kind_of?(ActiveFedora::Base) ? obj : Repository.get_object( obj )
      
        obj_done = Time.now
        obj_done_elapse = obj_done - start
        puts  " completed. Duration: #{obj_done_elapse}"
        
       print "\t Indexing object #{obj.pid} ... "
       # add the keywords and facets to the search index
       index_start = Time.now
       indexer.index( obj )
       
       index_done = Time.now
       index_elapsed = index_done - index_start
       
        puts "completed. Duration:  #{index_elapsed} ."
      
    
    rescue Exception => e
         p "unable to index #{obj}.  Failed with #{e.inspect}"
      
    
    end #begin

end

#solrize_objects ⇒ Object

This method retrieves a comprehensive list of all the unique identifiers in Fedora and solrizes each object’s full-text and facets into the search index



81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
 
# File 'lib/solrizer.rb', line 81

def solrize_objects
  # retrieve a list of all the pids in the fedora repository
  num_docs = 1000000   # modify this number to guarantee that all the objects are retrieved from the repository
  puts "WARNING: You have turned off indexing of Full Text content.  Be sure to re-run indexer with @@index_full_text set to true in main.rb" if index_full_text == false

  if @@index_list == false
    
    objects = Fedora::Repository.instance.find_objects(:limit=>num_docs)

    puts "Shelving #{objects.length} Fedora objects"
    objects.each do |object|
      solrize( object )
    end
   
  else
  
     if File.exists?(@@index_list)
        arr_of_pids = FasterCSV.read(@@index_list, :headers=>false)
        
        puts "Indexing from list at #{@@index_list}"
        puts "Shelving #{arr_of_pids.length} Fedora objects"
        
       arr_of_pids.each do |row|
          pid = row[0]
          solrize( pid )
end #FASTERCSV
      else
        puts "#{@@index_list} does not exists!"
      end #if File.exists
   
  end #if Index_LISTS
end