From 6368b781d5aad9a70a929b324e48f6673e87a93c Mon Sep 17 00:00:00 2001
From: redpony <redpony@ec762483-ff6d-05da-a07a-a48fb63a330f>
Date: Thu, 28 Oct 2010 18:02:35 +0000
Subject: fix, rename

git-svn-id: https://ws10smt.googlecode.com/svn/trunk@696 ec762483-ff6d-05da-a07a-a48fb63a330f
---
 .../support/generate-per-sentence-grammars.pl      | 48 ----------------
 .../support/generate_per_sentence_grammars.pl      | 64 ++++++++++++++++++++++
 2 files changed, 64 insertions(+), 48 deletions(-)
 delete mode 100644 word-aligner/support/generate-per-sentence-grammars.pl
 create mode 100755 word-aligner/support/generate_per_sentence_grammars.pl

(limited to 'word-aligner')

diff --git a/word-aligner/support/generate-per-sentence-grammars.pl b/word-aligner/support/generate-per-sentence-grammars.pl
deleted file mode 100644
index d621213e..00000000
--- a/word-aligner/support/generate-per-sentence-grammars.pl
+++ /dev/null
@@ -1,48 +0,0 @@
-#!/usr/bin/perl -w
-use strict;
-use utf8;
-
-die "Usage: $0 f.voc corpus.f-e grammar.f-e.gz\n" unless scalar @ARGV == 3;
-
-my $MAX_INMEM = 3000;
-
-open FV,"<$ARGV[0]" or die "Can't read $ARGV[0]: $!";
-open C,"<$ARGV[1]" or die "Can't read $ARGV[1]: $!";
-open G,"gunzip -c $ARGV[2]|" or die "Can't read $ARGV[2]: $!";
-
-binmode FV, ":utf8";
-binmode C, ":utf8";
-binmode G, ":utf8";
-
-my $vc = 0;
-my %most_freq;
-$most_freq{"<eps>"} = 1;
-while(my $f = <FV>) {
-  chomp $f;
-  $most_freq{$f}=1;
-  $vc++;
-  last if $vc == $MAX_INMEM;
-}
-close FV;
-
-print STDERR "Loaded $vc vocabulary items for permanent translation cache\n";
-
-my %grammar;
-my $memrc = 0;
-my $loadrc = 0;
-while(<G>) {
-  chomp;
-  my ($f, $e, $feats) = split / \|\|\| /;
-  if ($most_freq{$f}) {
-    #print "$_\n";
-    $memrc++;
-  } else {
-    $loadrc++;
-    $grammar{$f}="$e ||| $feats";
-  }
-}
-
-print STDERR "  mem rc: $memrc\n";
-print STDERR " load rc: $loadrc\n";
-
-
diff --git a/word-aligner/support/generate_per_sentence_grammars.pl b/word-aligner/support/generate_per_sentence_grammars.pl
new file mode 100755
index 00000000..c644ec6d
--- /dev/null
+++ b/word-aligner/support/generate_per_sentence_grammars.pl
@@ -0,0 +1,64 @@
+#!/usr/bin/perl -w
+use strict;
+use utf8;
+
+die "Usage: $0 f.voc corpus.f-e grammar.f-e.gz\n" unless scalar @ARGV == 3;
+
+my $MAX_INMEM = 3000;
+
+open FV,"<$ARGV[0]" or die "Can't read $ARGV[0]: $!";
+open C,"<$ARGV[1]" or die "Can't read $ARGV[1]: $!";
+open G,"gunzip -c $ARGV[2]|" or die "Can't read $ARGV[2]: $!";
+
+binmode FV, ":utf8";
+binmode C, ":utf8";
+binmode G, ":utf8";
+
+my $vc = 0;
+my %most_freq;
+$most_freq{"<eps>"} = 1;
+while(my $f = <FV>) {
+  chomp $f;
+  $most_freq{$f}=1;
+  $vc++;
+  last if $vc == $MAX_INMEM;
+}
+close FV;
+
+print STDERR "Loaded $vc vocabulary items for permanent translation cache\n";
+
+my %grammar;
+my $memrc = 0;
+my $loadrc = 0;
+while(<G>) {
+  chomp;
+  my ($f, $e, $feats) = split / \|\|\| /;
+  if ($most_freq{$f}) {
+    #print "$_\n";
+    $memrc++;
+  } else {
+    $loadrc++;
+    my $r = $grammar{$f};
+    if (!defined $r) {
+      $r = [];
+      $grammar{$f} = $r;
+    }
+    push @$r, "$e ||| $feats";
+  }
+}
+close G;
+print STDERR "  mem rc: $memrc\n";
+print STDERR " load rc: $loadrc\n";
+
+while(<C>) {
+  my ($f,$e) = split / \|\|\| /;
+  my @fwords = split /\s+/, $f;
+  my $tot = 0;
+  for my $f (@fwords) {
+    my $r = $grammar{$f};
+    die "No translations for: $f" unless $r;
+    my $num = scalar @$r;
+    $tot += $num;
+  }
+  print "RULES: $tot\n";
+}
-- 
cgit v1.2.3