From f91c57877a60d69c8e779101e1c245656ee6bc8f Mon Sep 17 00:00:00 2001
From: Patrick Simianer <p@simianer.de>
Date: Fri, 10 Nov 2017 22:28:48 +0100
Subject: rr: fix

---
 rr | 31 ++++++++++++++++++++++++-------
 1 file changed, 24 insertions(+), 7 deletions(-)

diff --git a/rr b/rr
index 30bf2d3..2faec7b 100755
--- a/rr
+++ b/rr
@@ -9,20 +9,37 @@ while line = STDIN.gets
   if cur_sz >= 1000
     windows << cur
     cur = []
-  else
-    cur << line.strip
-    cur_size += cur[-1].split.size
+    cur_sz = 0 
   end
+  cur << line.strip
+  cur_sz += cur.last.split.size
 end
 
+enums = [0.0]*4
+denoms = [0.0]*4
 windows.each { |w|
-  h = {}
-  h.default = 0
+  ng_by_n = [{}]*4
   w.each { |seg|
     ngrams(seg, 4) { |ng|
-      h[ng] += 1
+      if ng_by_n[ng.size-1].has_key? ng
+        ng_by_n[ng.size-1][ng] += 1
+      else
+        ng_by_n[ng.size-1][ng] = 1
+      end
     }
   }
-  puts h
+  ng_by_n.each_with_index { |ng,j|
+    puts ng.to_s
+    singletons = ng.reject { |k,v| v > 1 }.size
+    enums[j] += ng.size - singletons
+    denoms[j] += ng.size.to_f
+  }
+}
+
+rr = 1.0
+enums.each_with_index { |i,j|
+  rr *= i/denoms[j] 
 }
 
+puts ((rr**0.25)*100).round 2
+
-- 
cgit v1.2.3