added a new regression test for --continue option of mert-moses-new.pl

git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@2569 1f5c12ca-751b-0410-a591-d2e778427230

scripts/regression-testing/run-test-suite.pl

@@ -9,6 +9,7 @@ my @tests = qw (
   mert-moses-new
   mert-moses-new-nocase
   mert-moses-new-aggregate
+  mert-moses-new-continue
 );
 
 ###########################################################

scripts/regression-testing/tests/mert-moses-new-continue/command

@@ -0,0 +1,62 @@
+#! /bin/sh
+
+moses_scripts_dir=$1; shift
+mert_dir=$1; shift
+decoder=$1; shift
+data_dir=$1; shift
+testdir=$1; shift
+
+cwd=$moses_scripts_dir/regression-testing/
+
+pushd $testdir 2> /dev/null > /dev/null
+
+rm actual_index 2> /dev/null > /dev/null
+
+local_moses_ini=`perl $cwd/create_localized_moses_ini.pl $data_dir ./moses.ini 2> /dev/null`
+
+nbest=200
+#if you want distinct nbest change this:
+nbestdistinct=""
+#nbestdistinct="--nbest-distinct"
+
+devset="input"
+reference="reference"
+reference="reference."
+config=$local_moses_ini
+
+#1-LM, 1-TM, 1-lex-RM
+lambdas="d:0.4,0-2;0.15,0-2;0.15,0-2;0.15,0-2;0.15,0-2;0.15,0-2;0.15,0-2 lm:0.5,0-2 tm:0.2,0-2;0.2,0-2;0.2,0-2;0.2,0-2;0.2,-1-1 w:-1,-1-1"
+
+rm -r working_dir 2> /dev/null > /dev/null
+mkdir working_dir 2> /dev/null
+
+cp $0 working_dir/ 2> /dev/null
+cp ./data/archive.list working_dir/ 2> /dev/null
+cp ./data/run* working_dir/ 2> /dev/null
+cp ./data/finished_step.txt working_dir/ 2> /dev/null
+cp ./data/finished_step.txt working_dir/actual.index 2> /dev/null
+gzip -d working_dir/run*.NBEST.out.gz 2> /dev/null
+
+parameters="--nonorm --closest --no-filter-phrase-table --efficient_scorenbest_flag  --predictable-seeds --continue"
+
+/Users/nicolabertoldi/workspace/mosesdecoder/trunk/scripts/regression-testing/tests/mert-moses-new-continue/mert-moses-new.pl $devset $reference $decoder $config --mertdir $mert_dir --mertargs "--binary" --working-dir working_dir --nbest $nbest $nbestdistinct --lambdas "$lambdas" $parameters > working_dir/MERT.out 2> working_dir/MERT.err
+#$moses_scripts_dir/training/mert-moses-new.pl $devset $reference $decoder $config --mertdir $mert_dir --mertargs "--binary" --working-dir working_dir --nbest $nbest $nbestdistinct --lambdas "$lambdas" $parameters > working_dir/MERT.out 2> working_dir/MERT.err
+
+# checking for weights.opt files
+cat working_dir/run[0-9].weights.txt | awk '{a++; print "weights_"a": "$0}' > tmpout$$ 
+
+# checking for bleu scores
+cat working_dir/run[0-9].mert.log | grep -i ^Best | cut -d'>' -f2 | awk '{a++; print "best_"a": "$0}' >> tmpout$$
+
+# checking for translations
+cat working_dir/run[0-9].out | awk '{a++; print "translation_"a": "$0}' >> tmpout$$ 
+
+cat tmpout$$ > /dev/stdout
+
+rm $config 2> /dev/null > /dev/null
+rm tmpout$$ tmperr$$ 2> /dev/null > /dev/null
+#rm -r working_dir 2> /dev/null > /dev/null
+
+popd 2> /dev/null > /dev/null
+
+exit

scripts/regression-testing/tests/mert-moses-new-continue/data/actual.index

@@ -0,0 +1 @@
+5

scripts/regression-testing/tests/mert-moses-new-continue/data/archive.list

@@ -0,0 +1,8 @@
+run1.TRANS.out run1.NBEST.out
+run2.TRANS.out run2.NBEST.out
+run3.TRANS.out run3.NBEST.out
+run4.TRANS.out run4.NBEST.out
+run5.TRANS.out run5.NBEST.out
+run6.TRANS.out run6.NBEST.out
+run7.TRANS.out run7.NBEST.out
+run8.TRANS.out run8.NBEST.out

scripts/regression-testing/tests/mert-moses-new-continue/data/finished_step.txt

@@ -0,0 +1 @@
+5

scripts/regression-testing/tests/mert-moses-new-continue/data/run5.mert.log

@@ -0,0 +1,50 @@
+Seeding random numbers with 5000
+Scorer config string: reflen:closest,case:true
+name: reflen value: closest
+name: case value: true
+Using scorer regularisation strategy: none
+Using scorer regularisation window: 0
+Using case preservation: 1
+Using reference length strategy: closest
+Data::score_type BLEU
+Data::Scorer type from Scorer: BLEU
+BleuScorer: 9
+ScoreData: number_of_scores: 9
+Loading Data from: run1.scores.dat and run1.features.dat
+loading feature data from run1.features.dat
+loading score data from run1.scores.dat
+Loading Data from: run2.scores.dat and run2.features.dat
+loading feature data from run2.features.dat
+loading score data from run2.scores.dat
+Loading Data from: run3.scores.dat and run3.features.dat
+loading feature data from run3.features.dat
+loading score data from run3.scores.dat
+Loading Data from: run4.scores.dat and run4.features.dat
+loading feature data from run4.features.dat
+loading score data from run4.scores.dat
+Loading Data from: run5.scores.dat and run5.features.dat
+loading feature data from run5.features.dat
+loading score data from run5.scores.dat
+Data loaded : [0] seconds
+Try number 1 : [0] seconds
+Try number 2 : [0] seconds
+Try number 3 : [0] seconds
+Try number 4 : [1] seconds
+Try number 5 : [1] seconds
+Try number 6 : [1] seconds
+Try number 7 : [1] seconds
+Try number 8 : [1] seconds
+Try number 9 : [2] seconds
+Try number 10 : [2] seconds
+Try number 11 : [2] seconds
+Try number 12 : [2] seconds
+Try number 13 : [2] seconds
+Try number 14 : [3] seconds
+Try number 15 : [3] seconds
+Try number 16 : [3] seconds
+Try number 17 : [3] seconds
+Try number 18 : [3] seconds
+Try number 19 : [4] seconds
+Try number 20 : [4] seconds
+Best point: 0.0836727 0.135627 0.0602366 0.0263658 0.043897 0.0146552 0.0770568 0.112996 0.0437857 0.0206718 0.133082 0.0821608 -0.101029 -0.0647629  => 0.408975
+Stopping... : [4] seconds

scripts/regression-testing/tests/mert-moses-new-continue/data/run5.weights.txt

@@ -0,0 +1 @@
+0.0836727 0.135627 0.0602366 0.0263658 0.043897 0.0146552 0.0770568 0.112996 0.0437857 0.0206718 0.133082 0.0821608 -0.101029 -0.0647629 

scripts/regression-testing/tests/mert-moses-new-continue/data/run6.TRANS.out

@@ -0,0 +1,5 @@
+till Tuesday Almeno stable and mostly sunny weather with rising temperatures , especially in maxima values . 
+A possible worsening is expected in the second half of the week . 
+A moist and temporary flow will affect the Alps in the next days determining mostly changeable cloudiness , and scattered southern sectors , with light precipitation towards the half of next week . 
+Con today evening , Thursday , a disturbance is expected , bringing widespread rains into clearing on Friday and snowfalls to 600 of m 2000 m . 
+For the weekend clear or clear sky cloudy with temperature lowering minimum temperatures . 

scripts/regression-testing/tests/mert-moses-new-continue/data/run7.TRANS.out

@@ -0,0 +1,5 @@
+till Tuesday Almeno stable and mostly sunny weather with rising temperatures , especially in maxima values . 
+A possible worsening is expected in the second half of the week . 
+A moist and temporary flow will affect the Alpine area for the next days will determine changeable cloudiness , will probably southern sectors , and with light precipitation is expected for the half of the week . 
+Con evening today , Thursday , a disturbance is expected , bringing widespread rains into clearing on Friday and snowfalls to 600 of m 2000 m . 
+For the weekend clear or clear sky cloudy with dropping of minimum temperatures . 

scripts/regression-testing/tests/mert-moses-new-continue/filter-stderr

@@ -0,0 +1,9 @@
+#!/usr/bin/perl
+
+$x=0;
+while (<>) {
+  chomp;
+
+  $x++;
+  print "STDERR_$x=$_\n";
+}

scripts/regression-testing/tests/mert-moses-new-continue/filter-stdout

@@ -0,0 +1,8 @@
+#!/usr/bin/perl
+
+$x=0;
+while (<>) {
+  chomp;
+  $x++;
+  print "STDOUT_$x=$_\n";
+}

scripts/regression-testing/tests/mert-moses-new-continue/input

@@ -0,0 +1,5 @@
+Almeno fino a martedì tempo stabile e in prevalenza soleggiato con temperature in ripresa , specie nei valori massimi .
+Un possibile peggioramento è atteso nella seconda parte della settimana .
+Aria umida e a tratti fredda interesserà l' arco alpino nei prossimi giorni determinando condizioni di nuvolosità variabile , più fitta sui settori meridionali , e con possibili deboli precipitazioni verso metà settimana .
+Con la serata di oggi , giovedì , è atteso il transito di una perturbazione , che porterà piogge diffuse in esaurimento nella serata di venerdì e neve a quote di circa 2000 m .
+Per il fine settimana cielo sereno o poco nuvoloso con brusco calo delle temperature minime .

scripts/regression-testing/tests/mert-moses-new-continue/moses.ini

@@ -0,0 +1,60 @@
+# moses.ini for regression test
+
+[ttable-file]
+0 0 5 ${TM_PATH}/phrase-table.0-0.gz
+
+# language model
+[lmodel-file]
+1 0 3 ${LM_PATH}/train.en.w5.msb.nops.blm
+
+# limit on how many phrase translations e for each phrase f are loaded
+[ttable-limit]
+#ttable element load limit 0 = all elements loaded
+20
+0
+
+# distortion (reordering) files
+[distortion-file]
+0-0 msd-bidirectional-fe 6 ${RM_PATH}/reordering-table.msd-bidirectional-fe.0.5.0-0.gz
+
+# distortion (reordering) weight
+[weight-d]
+0.4
+0.15
+0.15
+0.15
+0.15
+0.15
+0.15
+
+# language model weight
+[weight-l]
+0.5
+
+# translation model weight (phrase translation, lexical weighting)
+[weight-t]
+0.2
+0.2
+0.2
+0.2
+0.2
+
+# word penalty
+[weight-w]
+-1.0
+
+[distortion-limit]
+4
+
+[beam-threshold]
+0.03
+
+[input-factors]
+0
+
+[mapping]
+T 0
+
+[verbose]
+0
+

scripts/regression-testing/tests/mert-moses-new-continue/reference.0

@@ -0,0 +1,5 @@
+Stable weather and mostly sunny with increasing temperatures , especially maximum temperatures , at least till Tuesday .
+A possible worsening is expected in the second part of the week .
+Humid air , cold at times , will affect the Alpine area for the next days fostering changeable cloudiness , thicker over southern areas , with possible precipitation towards the middle of the week .
+Today , Thursday , in the evening the passage of a disturbance is expected which will bring widespread rainfall dying out on Friday evening and snowfall around 2000 m .
+Over the weekend clear or lightly cloudy with a sudden fall of minimum temperatures .

scripts/regression-testing/tests/mert-moses-new-continue/reference.1

@@ -0,0 +1,5 @@
+Stable weather and mostly sunny with increasing temperatures , especially maximum temperatures , at least until Tuesday .
+A worsening is possible in the second half of the week .
+Humid air and cold at times will affect the Alpine area for the next days fostering unstable cloudiness , thicker over southern areas , with possible precipitation towards the middle of the week .
+Today Thursday in the evening the passage of a disturbance is expected which will bring widespread rainfall dying out on Friday evening and snowfall around 2000 metres .
+During the weekend clear or lightly cloudy with a sudden fall of minimum temperatures .

scripts/regression-testing/tests/mert-moses-new-continue/reference.2

@@ -0,0 +1,5 @@
+Stable and mostly sunny weather with increasing temperatures , especially maximum temperatures , at least until Tuesday .
+A possible worsening is expected in the second part of the week .
+Humid and cold air will affect the Alps in the next days fostering changeable cloudiness , thicker in the Southern areas , with possible precipitation towards the middle of the week .
+Today , Thursday , in the evening the passage of a disturbance is expected which will bring widespread rainfall dying out on Friday evening and snowfall around 2000 m .
+Over the weekend clear or lightly cloudy with a sudden fall of minimum temperatures .

scripts/regression-testing/tests/mert-moses-new-continue/truth/results.dat

@@ -0,0 +1,17 @@
+STDOUT_1=weights_1: 0.0836727 0.135627 0.0602366 0.0263658 0.043897 0.0146552 0.0770568 0.112996 0.0437857 0.0206718 0.133082 0.0821608 -0.101029 -0.0647629 
+STDOUT_2=weights_2: 0.0662178 0.138211 0.0613841 0.0268681 0.0447332 0.0149344 0.0785247 0.115149 0.0446198 0.0210656 0.135617 0.0837259 -0.102954 -0.0659966 
+STDOUT_3=weights_3: 0.0662177 0.138211 0.061384 0.0268681 0.0447331 0.0149344 0.0785246 0.115149 0.0446197 0.0210656 0.135617 0.0837258 -0.102954 -0.0659965 
+STDOUT_4=best_1:  0.408975
+STDOUT_5=best_2:  0.410349
+STDOUT_6=best_3:  0.410349
+STDOUT_7=translation_1: till Tuesday Almeno stable and mostly sunny weather with rising temperatures , especially in maxima values . 
+STDOUT_8=translation_2: A possible worsening is expected in the second half of the week . 
+STDOUT_9=translation_3: A moist and temporary flow will affect the Alps in the next days determining mostly changeable cloudiness , and scattered southern sectors , with light precipitation towards the half of next week . 
+STDOUT_10=translation_4: Con today evening , Thursday , a disturbance is expected , bringing widespread rains into clearing on Friday and snowfalls to 600 of m 2000 m . 
+STDOUT_11=translation_5: For the weekend clear or clear sky cloudy with temperature lowering minimum temperatures . 
+STDOUT_12=translation_6: till Tuesday Almeno stable and mostly sunny weather with rising temperatures , especially in maxima values . 
+STDOUT_13=translation_7: A possible worsening is expected in the second half of the week . 
+STDOUT_14=translation_8: A moist and temporary flow will affect the Alpine area for the next days will determine changeable cloudiness , will probably southern sectors , and with light precipitation is expected for the half of the week . 
+STDOUT_15=translation_9: Con evening today , Thursday , a disturbance is expected , bringing widespread rains into clearing on Friday and snowfalls to 600 of m 2000 m . 
+STDOUT_16=translation_10: For the weekend clear or clear sky cloudy with dropping of minimum temperatures . 
+TOTAL_WALLTIME ~ 10