# HG changeset patch
# User Igor Chubin <igor@chub.in>
# Date 1270492495 -10800
# Node ID 7e3a52db54adfd165c330be3f3e2dfa9b611e052
# Parent  35eeaf2620ce807c255c3520c6934ba58844fb51
Среднее количество слов и неизвестных слов (*10) в предложении в статистике.

Пример использования:

LANG  KNOWN%  UNKNOWN%  KNOWN     TOTAL     WPS  UWPS*10
en    89.8    10.2      167021    185840    21   21

diff -r 35eeaf2620ce -r 7e3a52db54ad new-words.sh
--- a/new-words.sh	Sun Apr 04 20:11:22 2010 +0300
+++ b/new-words.sh	Mon Apr 05 21:34:55 2010 +0300
@@ -75,10 +75,14 @@
     total_unknown="`cat $after|awk '{s=s+$1}END{print s}'`"
     total_known="`echo $total-$total_unknown|bc`"
     percentage="`echo '100*('$total-$total_unknown')'/$total | bc -l | sed 's/\\.\(.\).*/.\1/'`"
+    #sentences="`cat $after | perl -e 'local $/; $_=<>; s@http://[a-zA-Z&_.:/0-9%?=,\#+()\[\]~-]*@@g; s@\n@@g; s@(Mr|Mrs)\.@\1POINT@g; @sentences=split /\\./;print $#sentences;'`"
+    sentences="`cat $ORIGINAL_TEXT | perl -e 'local $/; $_=<>; s/[^.]//msg; print length($_);'`"
+
+
     if [ "$STAT_ONLY" = "YES" ]
     then
-        echo "LANG  KNOWN%  UNKNOWN%  KNOWN    TOTAL"
-        echo "$LANGUAGE    $percentage    `echo \(100-$percentage\) | bc -l`      $total_known    $total"
+        echo "LANG  KNOWN%  UNKNOWN%  KNOWN     TOTAL     WPS  UWPS*10"
+        echo "$LANGUAGE    $percentage    `echo \(100-$percentage\) | bc -l`      $total_known    $total    `echo $total/$sentences|bc`   `echo 10*$total_unknown/$sentences|bc` "
         rm $after
         return 0
     else