init

1 files changed, 19 insertions, 0 deletions

diff --git a/overlapping_rules/word_pair_keys.rb b/overlapping_rules/word_pair_keys.rb
new file mode 100755
index 0000000..8b54441
--- /dev/null
+++ b/overlapping_rules/word_pair_keys.rb
@@ -0,0 +1,19 @@
+#!/usr/bin/env ruby
+
+require_relative './util.rb'
+
+
+fn = 'newstest2008-grammar+index'
+approx_lines_per_shard = 12285856#/23
+STDOUT.sync = true
+
+i = 0
+read_rules_from_file1(STDIN, 'stdin') { |r|
+  i += 1
+  shard_percentage = ((i*100.0)/approx_lines_per_shard).round 2
+  STDERR.write "reporter:status:word pair key ##{i} #{shard_percentage}%\n"
+  r.fe_word_pairs.each { |p|
+    puts "#{p}\t#{r.id} ||| #{r.to_s}"
+  }
+}
+