diff options
author | bojar <bojar@1f5c12ca-751b-0410-a591-d2e778427230> | 2010-02-03 17:35:21 +0300 |
---|---|---|
committer | bojar <bojar@1f5c12ca-751b-0410-a591-d2e778427230> | 2010-02-03 17:35:21 +0300 |
commit | 9b10946f108e00670e0a0e3f51242d10273d34a7 (patch) | |
tree | 6509fe4c8118de0eba50bc14bd82d65ac98f7baf /scripts/analysis | |
parent | 8891069ed80b49a33ffdb0bee65eb24767cd35b8 (diff) |
fixed regexes to read current -osg format
verbose at bad lines
git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@2850 1f5c12ca-751b-0410-a591-d2e778427230
Diffstat (limited to 'scripts/analysis')
-rwxr-xr-x | scripts/analysis/sg2dot.perl | 12 |
1 files changed, 9 insertions, 3 deletions
diff --git a/scripts/analysis/sg2dot.perl b/scripts/analysis/sg2dot.perl index 1c5cf085d..f6a5dff49 100755 --- a/scripts/analysis/sg2dot.perl +++ b/scripts/analysis/sg2dot.perl @@ -26,8 +26,10 @@ $cpt = 0; $line=<>; #skip first line (the empty hypothesis, no arc in fact) +my $nr = 0; while(($line=<>) ) { + $nr++; $from = ""; $to = ""; $label = ""; @@ -40,7 +42,7 @@ while(($line=<>) ) #0 hyp=5 stack=1 back=0 score=-0.53862 transition=-0.53862 forward=181 fscore=-205.36 covered=0-0 out=I am , pC=-0.401291, c=-0.98555 #256 hyp=6566 stack=2 back=23 score=-2.15644 transition=-0.921959 recombined=6302 forward=15519 fscore=-112.807 covered=2-2 out=countries , , pC=-0.640574, c=-1.07215 - if($line =~ /hyp=(\d+).+stack=(\d+).+back=(\d+).+transition=([^ ]*).+recombined=(\d+).+out=(.*), pC/) + if($line =~ /hyp=(\d+).+stack=(\d+).+back=(\d+).+transition=([^ ]*).+recombined=(\d+).+out=(.*)(, pC|$)/) { #print STDERR "hyp=$1, stack=$2, from=$3, transition=$4, recombined=$5, out=$6\n"; $to = $1; @@ -56,7 +58,7 @@ while(($line=<>) ) #$stack++; #$stacks{$stack}{$recombined} = $recombined; } - elsif($line =~ /hyp=(\d+).+stack=(\d+).+back=(\d+).+transition=([^ ]*).+out=(.*), pC/) + elsif($line =~ /hyp=(\d+).+stack=(\d+).+back=(\d+).+transition=([^ ]*).+out=(.*)(, pC|$)/) { #print STDERR "hyp=$1, stack=$2, from=$3, transition=$4, out=$5\n"; $to = $1; @@ -69,7 +71,9 @@ while(($line=<>) ) #$stack++; #$stacks{$stack}{$to} = $to; } - else{ print STDERR "Bad file format ..."; exit(); } + else{ + die "$nr:Bad line: $line\n"; + } $o =~ s/\"/\\"/g ; #print STDERR "out = $o after regexp\n"; @@ -94,3 +98,5 @@ if ($organize_to_stacks) { } print STDOUT "\n}\n"; + +print STDERR "Converted $nr lines.\n"; |