@@ -170,7 +170,7 @@ class Flay next unless node.any? { |sub| Sexp === sub } next if node.mass < self.mass_threshold - self.hashes[node.fuzzy_hash] << node + self.hashes[node.structural_hash] << node end end @@ -182,7 +182,7 @@ class Flay all_hashes = {} self.hashes.values.each do |nodes| nodes.each do |node| - node.all_subhashes.each do |h| + node.all_structural_subhashes.each do |h| all_hashes[h] = true end end @@ -292,50 +292,14 @@ class String end class Sexp - def mass - @mass ||= self.structure.flatten.size + def structural_hash + @structural_hash ||= self.structure.hash end - alias :uncached_structure :structure - def structure - @structure ||= self.uncached_structure - end - - def similarity o - l, s, r = self.compare_to o - (2.0 * s) / (2.0 * s + l + r) - end - - def compare_to they - l = s = r = 0 - - l_sexp, l_lits = self.partition { |o| Sexp === o } - r_sexp, r_lits = they.partition { |o| Sexp === o } - - l += (l_lits - r_lits).size - s += (l_lits & r_lits).size - r += (r_lits - l_lits).size - - # TODO: I think this is wrong, since it isn't positional. What to do? - l_sexp.zip(r_sexp).each do |l_sub, r_sub| - next unless l_sub && r_sub # HACK - l2, s2, r2 = l_sub.compare_to r_sub - l += l2 - s += s2 - r += r2 - end - - return l, s, r - end - - def fuzzy_hash - @fuzzy_hash ||= self.structure.hash - end - - def all_subhashes + def all_structural_subhashes hashes = [] self.deep_each do |node| - hashes << node.fuzzy_hash + hashes << node.structural_hash end hashes end @@ -355,32 +319,3 @@ class Sexp end end end - -class Array - def intersection other - intersection, start = [], 0 - other_size = other.length - self.each_with_index do |m, i| - (start...other_size).each do |j| - n = other.at j - if m == n then - intersection << m - start = j + 1 - break - end - end - end - intersection - end - - def triangle # TODO: use? - max = self.size - (0...max).each do |i| - o1 = at(i) - (i+1...max).each do |j| - o2 = at(j) - yield o1, o2 - end - end - end -end lib/flay.rb @@ -3,6 +3,8 @@ require 'test/unit' require 'flay' +$: << "../../sexp_processor/dev/lib" + require 'pp' # TODO: remove ON_1_9 = RUBY_VERSION =~ /1\.9/ @@ -28,36 +30,7 @@ class TestSexp < Test::Unit::TestCase s(:call, nil, :d, s(:arglist))) end - def test_mass - assert_equal 1, s(:a).mass - assert_equal 3, s(:a, s(:b), s(:c)).mass - assert_equal 7, @s.mass - end - - def test_compare_to - s1 = s(:a, :b, :c) - s2 = s(:d, :e, :f) - assert_equal [3, 0, 3], s1.compare_to(s2), "100% different" - - s1 = s(:a, :b, :c) - s2 = s(:a, :b, :c) - assert_equal [0, 3, 0], s1.compare_to(s2), "100% same" - - s1 = s(:a, :b, :c, :d) - s2 = s(:a, :b, :c, :e) - assert_equal [1, 3, 1], s1.compare_to(s2), "1 element different on each" - - s1 = s(:a, :d, :b, :c) - s2 = s(:a, :b, :c, :e) - assert_equal [1, 3, 1], s1.compare_to(s2), "positionally different" - - - s1 = s(:a, s(:d), :b, :c) - s2 = s(:a, :b, :c, s(:e)) - assert_equal [1, 3, 1], s1.compare_to(s2), "simple subtree difference" - end - - def test_fuzzy_hash + def test_structural_hash s = s(:iter, s(:call, nil, :a, s(:arglist, s(:lit, 1))), s(:lasgn, :c), @@ -65,22 +38,22 @@ class TestSexp < Test::Unit::TestCase hash = 955256285 - assert_equal hash, s.fuzzy_hash, "hand copy" - assert_equal hash, @s.fuzzy_hash, "ivar from setup" - assert_equal hash, @s.deep_clone.fuzzy_hash, "deep clone" - assert_equal hash, s.deep_clone.fuzzy_hash, "copy deep clone" + assert_equal hash, s.structural_hash, "hand copy" + assert_equal hash, @s.structural_hash, "ivar from setup" + assert_equal hash, @s.deep_clone.structural_hash, "deep clone" + assert_equal hash, s.deep_clone.structural_hash, "copy deep clone" end unless SKIP_1_9 - def test_all_subhashes + def test_all_structural_subhashes expected = [-704571402, -282578980, -35395725, 160138040, 815971090, 927228382] - assert_equal expected, @s.all_subhashes.sort.uniq + assert_equal expected, @s.all_structural_subhashes.sort.uniq x = [] @s.deep_each do |o| - x << o.fuzzy_hash + x << o.structural_hash end assert_equal expected, x.sort.uniq @@ -146,12 +119,3 @@ class TestSexp < Test::Unit::TestCase assert flay.hashes.empty? end end - -class ArrayIntersectionTests < Test::Unit::TestCase - def test_real_array_intersection - assert_equal [2], [2, 2, 2, 3, 7, 13, 49] & [2, 2, 2, 5, 11, 107] - assert_equal [2, 2, 2], [2, 2, 2, 3, 7, 13, 49].intersection([2, 2, 2, 5, 11, 107]) - assert_equal ['a', 'c'], ['a', 'b', 'a', 'c'] & ['a', 'c', 'a', 'd'] - assert_equal ['a', 'a'], ['a', 'b', 'a', 'c'].intersection(['a', 'c', 'a', 'd']) - end -end test/test_flay.rb d02dcf1f2cd1f1f1aef2d02942ee5b358ad39467 Ryan Davis ryand@zenspider.com http://github.com/seattlerb/flay/commit/9b5e4c6765788c2f241f1c925bac4898110ba662 9b5e4c6765788c2f241f1c925bac4898110ba662 2009-08-05T17:03:11-07:00 2009-08-05T17:03:11-07:00 + Renamed fuzzy_hash to structural_hash. + Renamed all_subhashes to all_structural_subhashes. + Pushed Sexp#mass up to sexp_processor. + Removed #similarity #compare_to, #intersection, #triangle, and other cruft. [git-p4: depot-paths = "//src/flay/dev/": change = 5361] 371219b15cd6869bb9b1c706e557cd179b6e019b Ryan Davis ryand@zenspider.com