better error message when no corpus defined, better integration of IRSTLM training

1 files changed, 16 insertions, 8 deletions

diff --git a/scripts/generic/trainlm-irst.perl b/scripts/generic/trainlm-irst.perl
index 15e8d7ee9..d6ac3aaf6 100755
--- a/scripts/generic/trainlm-irst.perl
+++ b/scripts/generic/trainlm-irst.perl
@@ -17,34 +17,42 @@ use Getopt::Long;
 my $order;
 my $corpusPath;
 my $lmPath;
-my $cores;
+my $cores = 2;
 my $irstPath;
+my $tempPath = "tmp";
 
 GetOptions("order=s"  => \$order,
            "text=s"   => \$corpusPath,
            "lm=s"     => \$lmPath,
            "cores=s"  => \$cores,
            "irst-dir=s"  => \$irstPath,
+           "temp-dir=s"  => \$tempPath
 	   ) or exit 1;
 
+die("ERROR: please set order") unless defined($order);
+die("ERROR: please set text") unless defined($corpusPath);
+die("ERROR: please set lm") unless defined($lmPath);
+die("ERROR: please set irst-dir") unless defined($irstPath);
+
 my $ext = ($corpusPath =~ m/([^.]+)$/)[0];
 print "extension is $ext\n";
 
-mkdir 'temp';
+$tempPath .= "/irstlm-build-tmp.$$";
+`mkdir -p $tempPath`;
 
 my $cmd;
 if ($ext eq "gz")
 {
-    $cmd = "zcat $corpusPath | $irstPath/bin/add-start-end.sh | gzip -c > temp/monolingual.setagged.gz";
+    $cmd = "zcat $corpusPath | $irstPath/add-start-end.sh | gzip -c > $tempPath/monolingual.setagged.gz";
 }
 else
 {
-    $cmd = "cat $corpusPath | $irstPath/bin/add-start-end.sh | gzip -c > temp/monolingual.setagged.gz";
+    $cmd = "cat $corpusPath | $irstPath/add-start-end.sh | gzip -c > $tempPath/monolingual.setagged.gz";
 }
 print STDERR "EXECUTING $cmd\n";
 `$cmd`;
 
-$cmd = "IRSTLM=$irstPath $irstPath/bin/build-lm.sh -t stat4 -i \"gunzip -c temp/monolingual.setagged.gz\" -n $order -p -o temp/iarpa.gz -k $cores";
+$cmd = "IRSTLM=$irstPath/.. $irstPath/build-lm.sh -t $tempPath/stat4 -i \"gunzip -c $tempPath/monolingual.setagged.gz\" -n $order -p -o $tempPath/iarpa.gz -k $cores";
 print STDERR "EXECUTING $cmd\n";
 `$cmd`;
 
@@ -53,17 +61,17 @@ print "extension is $ext\n";
 
 if ($ext eq "gz")
 {
-    $cmd = "$irstPath/bin/compile-lm temp/iarpa.gz --text yes /dev/stdout | gzip -c > $lmPath";
+    $cmd = "$irstPath/compile-lm $tempPath/iarpa.gz --text yes /dev/stdout | gzip -c > $lmPath";
 }
 else
 {
-    $cmd = "$irstPath/bin/compile-lm temp/iarpa.gz --text yes $lmPath";
+    $cmd = "$irstPath/compile-lm $tempPath/iarpa.gz --text yes $lmPath";
 }
 
 print STDERR "EXECUTING $cmd\n";
 `$cmd`;
 
-$cmd = "rm -rf temp stat4";
+$cmd = "rm -rf $tempPath";
 print STDERR "EXECUTING $cmd\n";
 `$cmd`;