1 files changed, 179 insertions, 0 deletions
diff --git a/overlapping_rules/merge_rules.rb b/overlapping_rules/merge_rules.rb
new file mode 100755
index 0000000..987d417
--- /dev/null
+++ b/overlapping_rules/merge_rules.rb
@@ -0,0 +1,179 @@
+#!/usr/bin/env ruby
+
+require_relative './util.rb'
+
+# ISSUES
+# new rule already in global g?
+#  => uniq!
+# singletons? => uniq -D
+# R+S = R => X
+# variants?
+#
+
+
+def is_sub(first, second, dir='left')
+  if dir=='right'
+    first = first.reverse; second = second.reverse
+  end
+
+  second_index = 0
+  match = false
+  first_range = Range.new; second_range = Range.new
+  first.each_with_index { |i,first_index|
+    break if i.match('\[X,\d\]') #(/\[X,\d\]/)
+    break if second_index > second.size-1
+    if i == second[second_index]
+      if !match
+        first_range.from = first_index
+        first_range.to   = first_index
+        second_range.from = second_index
+        second_range.to   = second_index
+        match = true
+      else
+        first_range.to = first_index
+        second_range.to = second_index
+      end
+      second_index += 1
+    else
+      first_range.from = first_range.to = second_range.from = second_range.to = nil
+    end
+  }
+  if dir=='right' && first_range.from&&first_range.to&&second_range.from&&second_range.to
+    first_range.correct(first.size-1)
+    second_range.correct(second.size-1)
+  end
+  return first_range, second_range if (first_range.from&&first_range.to&&second_range.from&&second_range.to)&&
+                                      (first_range.from>0||first_range.to>0||second_range.from>0||second_range.to>0)
+  return false
+end
+
+def merge(r, s, r_f_range, s_f_range, r_e_range, s_e_range, dir)
+  #ret = []  ?
+  new_rule = Rule.new
+  new_rule.f = Array.new(s.f)
+  new_rule.e = Array.new(s.e)
+  new_rule.nt = r.nt
+  if dir == 'left'
+    return nil if r_f_range.from==0&&r_e_range.from==0
+    (r_f_range.from-1).downto(0) { |i| new_rule.f.unshift r.f[i] }
+    (r_e_range.from-1).downto(0) { |i| new_rule.e.unshift r.e[i] }
+  elsif dir == 'right'
+    return nil if r_f_range.from==r.f.size-1&&r_e_range.from==r.e.size-1
+    (r_f_range.to+1).upto(r.f.size-1) { |i| new_rule.f <<  r.f[i] }
+    (r_e_range.to+1).upto(r.e.size-1) { |i| new_rule.e <<  r.e[i] }
+  end
+  return new_rule
+end
+
+def test
+  a = ["der", "eurozone", "[X,1]", "die", "[X,2]"]
+  b = ["eurozone"]
+  x = ["the", "eurozone", "[X,1]", "[X,2]"]
+  y = ["eurozone", "members"]
+
+  r,s = is_sub(a,b)
+  puts "#{r} #{s}"
+
+  r,s = is_sub(b,a)
+  puts "#{r} #{s}"
+
+  r,s = is_sub(x,y)
+  puts "#{r} #{s}"
+
+  r,s = is_sub(y,x)
+  puts "#{r} #{s}"
+
+  puts
+  puts
+
+  a = ["schuld", "an", "[X,1]", "inflation"]
+  b = ["schuld"]
+  x = ["blamed", "for", "[X,1]", "inflation"]
+  y = ["for", "responsible"]
+
+  r,s = is_sub(a,b)
+  puts "#{r} #{s}"
+
+  r,s = is_sub(b,a)
+  puts "#{r} #{s}"
+
+  r,s = is_sub(x,y)
+  puts "#{r} #{s}"
+
+  r,s = is_sub(y,x)
+  puts "#{r} #{s}"
+  exit
+end
+test if ARGV[0] == 'test'
+
+# main
+def main
+  strings = {}
+  while line = STDIN.gets
+    ids, data = line.split "\t", 2
+    r_id, s_id = ids.split '+', 2
+    r_id = r_id.to_s
+    s_id = s_id.to_s
+    r_raw, s_raw = data.split ' <<<>>> ', 2
+    r = Rule.new r_raw, r_id
+    s = Rule.new s_raw, s_id
+    #STDERR.puts "reporter:status:processing rules #{r_id} and #{s_id}"
+
+    # left, R->S
+    range_f_first, range_f_second = is_sub(r.f, s.f)
+    range_e_first, range_e_second = is_sub(r.e, s.e)
+    if range_f_first&&range_f_second&&range_e_first&&range_e_second
+      new_rule = merge(r, s, range_f_first, range_f_first, range_e_first, range_e_second, 'left')
+      if new_rule && !strings.has_key?(new_rule.rule_id_string)
+        #puts "R: #{r.f} ||| #{r.e}\nS: #{s.f} ||| #{s.e}"
+        #puts "f:(#{range_f_first} #{range_f_second}) e:(#{range_e_first} #{range_e_second})\n"
+        puts new_rule.to_s
+        strings[new_rule.rule_id_string] = true
+        #puts "\n"
+      end
+    end
+
+    # left, S->R
+    range_f_first, range_f_second = is_sub(s.f, r.f)
+    range_e_first, range_e_second = is_sub(s.e, r.e)
+    if range_f_first&&range_f_second&&range_e_first&&range_e_second
+      new_rule = merge(s, r, range_f_first, range_f_first, range_e_first, range_e_second, 'left')
+      if new_rule && !strings.has_key?(new_rule.rule_id_string)
+        #puts "S: #{s.f} ||| #{s.e}\nR: #{r.f} ||| #{r.e}"
+        #puts "f:(#{range_f_first} #{range_f_second}) e:(#{range_e_first} #{range_e_second})\n"
+        puts new_rule.to_s
+        strings[new_rule.rule_id_string] = true
+        #puts "\n"
+      end
+    end
+
+    # right, R->S
+    range_f_first, range_f_second = is_sub(r.f, s.f, 'right')
+    range_e_first, range_e_second = is_sub(r.e, s.e, 'right')
+    if range_f_first&&range_f_second&&range_e_first&&range_e_second
+      new_rule = merge(r, s, range_f_first, range_f_first, range_e_first, range_e_second, 'right')
+      if new_rule && !strings.has_key?(new_rule.rule_id_string)
+        #puts "Rr: #{r.f} ||| #{r.e}\nSr: #{s.f} ||| #{s.e}"
+        #puts "f:(#{range_f_first} #{range_f_second}) e:(#{range_e_first} #{range_e_second})\n"
+        puts new_rule.to_s
+        strings[new_rule.rule_id_string] = true
+        #puts "\n"
+      end
+    end
+
+    # right, S->R
+    range_f_first, range_f_second = is_sub(s.f, r.f, 'right')
+    range_e_first, range_e_second = is_sub(s.e, r.e, 'right')
+    if range_f_first&&range_f_second&&range_e_first&&range_e_second
+      new_rule = merge(s, r, range_f_first, range_f_first, range_e_first, range_e_second, 'right')
+      if new_rule && !strings.has_key?(new_rule.rule_id_string)
+        #puts "Sr: #{s.f} ||| #{s.e}\nRr: #{r.f} ||| #{r.e}"
+        #puts "f:(#{range_f_first} #{range_f_second}) e:(#{range_e_first} #{range_e_second})\n"
+        puts new_rule.to_s
+        strings[new_rule.rule_id_string] = true if new_rule
+        #puts "\n"
+      end
+    end
+  end
+end
+