loose_tight_dictionary 0.1.0 → 0.1.1
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
data/benchmark/memory.rb
CHANGED
|
@@ -42,17 +42,21 @@ class LooseTightDictionary
|
|
|
42
42
|
def log(str = '') #:nodoc:
|
|
43
43
|
(options[:log] || $stderr).puts str unless options[:log] == false
|
|
44
44
|
end
|
|
45
|
-
|
|
46
|
-
def
|
|
47
|
-
|
|
48
|
-
|
|
45
|
+
|
|
46
|
+
def find_all(needle, options = {})
|
|
47
|
+
options = options.symbolize_keys.merge(:find_all => true)
|
|
48
|
+
find needle, options
|
|
49
49
|
end
|
|
50
50
|
|
|
51
51
|
# todo fix record.record confusion (should be wrapper.record or smth)
|
|
52
|
-
def find(needle,
|
|
52
|
+
def find(needle, options = {})
|
|
53
53
|
raise Freed if freed?
|
|
54
54
|
free_last_result
|
|
55
55
|
|
|
56
|
+
options = options.symbolize_keys
|
|
57
|
+
gather_last_result = options.fetch(:gather_last_result, true)
|
|
58
|
+
find_all = options.fetch(:find_all, false)
|
|
59
|
+
|
|
56
60
|
if gather_last_result
|
|
57
61
|
last_result.tighteners = tighteners
|
|
58
62
|
last_result.identities = identities
|
|
@@ -65,7 +69,13 @@ class LooseTightDictionary
|
|
|
65
69
|
last_result.needle = needle
|
|
66
70
|
end
|
|
67
71
|
|
|
68
|
-
|
|
72
|
+
if strict_blocking and blockings.none? { |blocking| blocking.encompass? needle }
|
|
73
|
+
if find_all
|
|
74
|
+
return []
|
|
75
|
+
else
|
|
76
|
+
return nil
|
|
77
|
+
end
|
|
78
|
+
end
|
|
69
79
|
|
|
70
80
|
encompassed, unencompassed = if strict_blocking and blockings.any?
|
|
71
81
|
haystack.partition do |record|
|
|
@@ -98,6 +108,10 @@ class LooseTightDictionary
|
|
|
98
108
|
last_result.certainly_different = certainly_different
|
|
99
109
|
end
|
|
100
110
|
|
|
111
|
+
if find_all
|
|
112
|
+
return possibly_identical.map { |record| record.record }
|
|
113
|
+
end
|
|
114
|
+
|
|
101
115
|
similarities = possibly_identical.map do |record|
|
|
102
116
|
needle.similarity record
|
|
103
117
|
end.sort
|
|
@@ -17,7 +17,8 @@ class TestLooseTightDictionary < Test::Unit::TestCase
|
|
|
17
17
|
|
|
18
18
|
def test_002_find_with_score
|
|
19
19
|
d = LooseTightDictionary.new %w{ NISSAN HONDA }
|
|
20
|
-
assert_equal
|
|
20
|
+
assert_equal 'NISSAN', d.find('MISSAM')
|
|
21
|
+
assert_equal 0.6, d.last_result.score
|
|
21
22
|
end
|
|
22
23
|
|
|
23
24
|
def test_003_last_result
|
|
@@ -75,4 +76,10 @@ class TestLooseTightDictionary < Test::Unit::TestCase
|
|
|
75
76
|
d.find('foobar')
|
|
76
77
|
end
|
|
77
78
|
end
|
|
79
|
+
|
|
80
|
+
def test_012_find_all
|
|
81
|
+
d = LooseTightDictionary.new [ 'X', 'X22', 'Y', 'Y4' ], :blockings => [ /X/, /Y/ ], :strict_blocking => true
|
|
82
|
+
assert_equal ['X', 'X22' ], d.find_all('X')
|
|
83
|
+
assert_equal [], d.find_all('A')
|
|
84
|
+
end
|
|
78
85
|
end
|