Welcome to mirror list, hosted at ThFree Co, Russian Federation.

github.com/moses-smt/mosesdecoder.git - Unnamed repository; edit this file 'description' to name the repository.
summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorHieu Hoang <hieuhoang@gmail.com>2015-04-28 14:03:59 +0300
committerHieu Hoang <hieuhoang@gmail.com>2015-04-28 14:03:59 +0300
commit8f9bf7ea386feb1aef5413730bd627a1161c5928 (patch)
tree16b9162f3f3a0a819f586c4aff4ea9ed1e9a7e88 /scripts
parent7c19add821266bc11e1e82176275940c4f2cdebc (diff)
add -config
Diffstat (limited to 'scripts')
-rwxr-xr-xscripts/training/wrappers/madamira-wrapper.perl12
1 files changed, 9 insertions, 3 deletions
diff --git a/scripts/training/wrappers/madamira-wrapper.perl b/scripts/training/wrappers/madamira-wrapper.perl
index 9866e6421..5c1d0404f 100755
--- a/scripts/training/wrappers/madamira-wrapper.perl
+++ b/scripts/training/wrappers/madamira-wrapper.perl
@@ -15,6 +15,7 @@ my $TMPDIR = "tmp";
my $SCHEME = "D2";
my $KEEP_TMP = 0;
my $MADA_DIR;
+my $CONFIG;
my $FACTORS_STR;
my @FACTORS;
@@ -24,9 +25,14 @@ GetOptions(
"tmpdir=s" => \$TMPDIR,
"keep-tmp" => \$KEEP_TMP,
"mada-dir=s" => \$MADA_DIR,
- "factors=s" => \$FACTORS_STR
+ "factors=s" => \$FACTORS_STR,
+ "config=s" => \$CONFIG
) or die("ERROR: unknown options");
+if (!defined($CONFIG)) {
+ $CONFIG = "$MADA_DIR/samples/sampleConfigFile.xml";
+}
+
$TMPDIR = abs_path($TMPDIR);
print STDERR "TMPDIR=$TMPDIR \n";
@@ -65,7 +71,7 @@ else {
$cmd = "$SPLIT_EXEC -l 10000 -a 7 -d $TMPDIR/input $TMPDIR/split/x";
`$cmd`;
-$cmd = "cd $MADA_DIR && parallel --jobs 4 java -Xmx2500m -Xms2500m -XX:NewRatio=3 -jar $MADA_DIR/MADAMIRA.jar -rawinput {} -rawoutdir $TMPDIR/out -rawconfig $MADA_DIR/samples/sampleConfigFile.xml ::: $TMPDIR/split/x*";
+$cmd = "cd $MADA_DIR && parallel --jobs 4 java -Xmx2500m -Xms2500m -XX:NewRatio=3 -jar $MADA_DIR/MADAMIRA.jar -rawinput {} -rawoutdir $TMPDIR/out -rawconfig $CONFIG ::: $TMPDIR/split/x*";
print STDERR "Executing: $cmd\n";
`$cmd`;
@@ -77,7 +83,7 @@ print STDERR "Executing: $cmd\n";
open(MADA_OUT,"<$infile.mada");
#binmode(MADA_OUT, ":utf8");
while(my $line = <MADA_OUT>) {
- chop($line);
+ chomp($line);
#print STDERR "line=$line \n";
if (index($line, "SENTENCE BREAK") == 0) {