From fac00976168c6b3c94d01d76babede147e4a0710 Mon Sep 17 00:00:00 2001
From: Patrick Simianer <p@simianer.de>
Date: Wed, 20 Jul 2016 10:55:13 +0200
Subject: noloo new rules: align all

---
 phrase2_extraction/phrase2_extraction.rb | 10 ++++++++++
 server.rb                                | 12 +++++++++++-
 2 files changed, 21 insertions(+), 1 deletion(-)

diff --git a/phrase2_extraction/phrase2_extraction.rb b/phrase2_extraction/phrase2_extraction.rb
index 1f268cd..b376953 100755
--- a/phrase2_extraction/phrase2_extraction.rb
+++ b/phrase2_extraction/phrase2_extraction.rb
@@ -178,6 +178,16 @@ class Rule
     }
     astr.strip!
 
+    #a = []
+    #source_string.strip.lstrip.split.each_with_index { |s,i|
+    #  target_string.strip.lstrip.split.each_with_index { |t,j|
+    #    if !s.match /\[X,\d+\]/ and !t.match /\[X,\d+\]/
+    #      a << "#{i}-#{j}"
+    #    end
+    #  }
+    #}
+    #astr = a.join ' '
+
     return "[X] ||| #{source_string} ||| #{target_string} ||| NewRule=1 ||| #{astr}"
   end
 
diff --git a/server.rb b/server.rb
index 4a3e8f6..752d0d5 100755
--- a/server.rb
+++ b/server.rb
@@ -307,7 +307,17 @@ def process_next reply
           }
           tmp_rules_known = tmp_rules - tmp_rules_new
           tmp_rules_known.each { |i| no_loo_known_rules << "[X] ||| #{i[0]} ||| #{i[1]} ||| KnownRule=1 ||| 0-0" }
-          tmp_rules_new.each { |i| no_loo_new_rules << "[X] ||| #{i[0]} ||| #{i[1]} ||| NewRule=1 ||| 0-0" }
+          tmp_rules_new.each { |i|
+            a = []
+            i[0].strip.lstrip.split.each_with_index { |s,ii|
+              i[1].strip.lstrip.split.each_with_index { |t,j|
+                if !s.match /\[X,\d+\]/ and !t.match /\[X,\d+\]/
+                  a << "#{ii}-#{j}"
+                end
+              }
+            }
+            no_loo_new_rules << "[X] ||| #{i[0]} ||| #{i[1]} ||| NewRule=1 ||| #{a.join ' '}"
+          }
         end
         # regular
         new_rules = PhrasePhraseExtraction.extract_rules f, e, data["align"], true
-- 
cgit v1.2.3