Class: Linkage::ResultSet

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

Instance Method Summary collapse

Constructor Details

#initialize(config) ⇒ ResultSet

Returns a new instance of ResultSet.



3
4
5
6
7
# File 'lib/linkage/result_set.rb', line 3

def initialize(config)
  @config = config
  @next_group_id = 1
  @next_group_mutex = Mutex.new
end

Instance Method Details

#add_group(group, dataset_id = nil) ⇒ Object



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
77
# File 'lib/linkage/result_set.rb', line 48

def add_group(group, dataset_id = nil)
  if @config.decollation_needed?
    original_values = group.values
    values = group.decollated_values
    if !@groups_buffer
      groups_headers = [:id] + values.keys
      @groups_buffer = ImportBuffer.new(database[@config.groups_table_name],
        groups_headers)

      original_groups_headers = [:id] + original_values.keys
      @original_groups_buffer = ImportBuffer.new(
        database[@config.original_groups_table_name],
        original_groups_headers)
    end

    group_id = next_group_id
    @groups_buffer.add([group_id] + values.values)
    @original_groups_buffer.add([group_id] + original_values.values)
  else
    # Non-DRY for minute speed improvements
    values = group.values
    if !@groups_buffer
      groups_headers = [:id] + values.keys
      @groups_buffer = ImportBuffer.new(database[@config.groups_table_name],
        groups_headers)
    end
    group_id = next_group_id
    @groups_buffer.add([group_id] + values.values)
  end
end

#add_match(record_1_id, record_2_id, total_score) ⇒ Object



88
89
90
91
92
93
94
95
# File 'lib/linkage/result_set.rb', line 88

def add_match(record_1_id, record_2_id, total_score)
  if !@matches_buffer
    matches_headers = [:record_1_id, :record_2_id, :total_score]
    @matches_buffer = ImportBuffer.new(database[@config.matches_table_name],
      matches_headers)
  end
  @matches_buffer.add([record_1_id, record_2_id, total_score])
end

#add_score(comparator_id, record_1_id, record_2_id, score) ⇒ Object



79
80
81
82
83
84
85
86
# File 'lib/linkage/result_set.rb', line 79

def add_score(comparator_id, record_1_id, record_2_id, score)
  if !@scores_buffer
    scores_headers = [:comparator_id, :record_1_id, :record_2_id, :score]
    @scores_buffer = ImportBuffer.new(database[@config.scores_table_name],
      scores_headers)
  end
  @scores_buffer.add([comparator_id, record_1_id, record_2_id, score])
end

#create_tables!Object



21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
# File 'lib/linkage/result_set.rb', line 21

def create_tables!
  if @config.groups_table_needed?
    schema = @config.groups_table_schema
    if @config.decollation_needed?
      database.create_table(@config.original_groups_table_name) do
        schema.each { |col| column(*col) }
      end
    end

    database.create_table(@config.groups_table_name) do
      schema.each { |col| column(*col) }
    end
  end

  if @config.scores_table_needed?
    schema = @config.scores_table_schema
    database.create_table(@config.scores_table_name) do
      schema.each { |col| column(*col) }
    end
  end

  schema = @config.matches_table_schema
  database.create_table(@config.matches_table_name) do
    schema.each { |col| column(*col) }
  end
end

#databaseObject



13
14
15
16
17
18
19
# File 'lib/linkage/result_set.rb', line 13

def database
  # FIXME: If the results database is the same as one of the datasets
  # being linked, there will be two connections to said database. This
  # could result in unexpected locking for non-concurrent databases (like
  # SQLite).
  @database ||= Sequel.connect(@config.results_uri, @config.results_uri_options)
end

#flush!Object



97
98
99
100
101
102
# File 'lib/linkage/result_set.rb', line 97

def flush!
  @groups_buffer.flush if @groups_buffer
  @original_groups_buffer.flush if @original_groups_buffer
  @scores_buffer.flush if @scores_buffer
  @matches_buffer.flush if @matches_buffer
end

#get_group(index) ⇒ Object



104
105
106
107
# File 'lib/linkage/result_set.rb', line 104

def get_group(index)
  values = groups_dataset.order(:id).limit(1, index).first
  Group.from_row(values)
end

#groups_datasetObject



9
10
11
# File 'lib/linkage/result_set.rb', line 9

def groups_dataset
  @groups_dataset ||= Dataset.new(database[:groups])
end

#groups_records_datasets(group) ⇒ Object



109
110
111
112
# File 'lib/linkage/result_set.rb', line 109

def groups_records_datasets(group)
  datasets = @config.datasets_with_applied_simple_expectations
  datasets.collect! { |ds| ds.dataset_for_group(group) }
end