fixing some issues

Author: jtrmal

egs/chime5/s5/conf/chime5.cfg renamed to egs/chime5/s5/conf/beamformit.cfg

@@ -17,12 +17,19 @@
 
 [ -f ./path.sh ] && . ./path.sh
 
-uconv=`command -v uconv 2>/dev/null` \
+command -v uconv &>/dev/null \
  || { echo >&2 "uconv not found on PATH. You will have to install ICU4C"; exit 1; }
 
-srilm=`command -v ngram 2>/dev/null` \
+command -v ngram &>/dev/null \
  || { echo >&2 "srilm not found on PATH. Please use the script $KALDI_ROOT/tools/extras/install_srilm.sh to install it"; exit 1; }
 
+if [ -z ${LIBLBFGS} ]; then
+ echo >&2 "SRILM is not compiled with the support of MaxEnt models."
+ echo >&2 "You should use the script in \$KALDI_ROOT/tools/install_srilm.sh"
+ echo >&2 "which will take care of compiling the SRILM with MaxEnt support"
+ exit 1;
+fi
+
 sox=`command -v sox 2>/dev/null` \
  || { echo >&2 "sox not found on PATH. Please install it manually (you will need version 14.4.0 and higher)."; exit 1; }
 
@@ -35,10 +42,10 @@ if [ ! -z "$sox" ]; then
  fi
 fi
 
-phalign=`command -v phonetisaurus-align 2>/dev/null` \
+command -v phonetisaurus-align &>/dev/null \
  || { echo >&2 "Phonetisaurus not found on PATH. Please use the script $KALDI_ROOT/tools/extras/install_phonetisaurus.sh to install it"; exit 1; }
 
-beamformit=`command -v BeamformIt 2>/dev/null` \
+command -v BeamformIt &>/dev/null \
  || { echo >&2 "BeamformIt not found on PATH. Please use the script $KALDI_ROOT/tools/extras/install_beamformit.sh to install it"; exit 1; }
 
 exit 0

egs/chime5/s5/conf/decode.config

@@ -20,10 +20,26 @@ if [ $# -ne 3 ] ; then
  exit 1
 fi
 
+set -e -o pipefail
+
 adir=$1
 jdir=$2
 dir=$3
 
+json_count=$(find $jdir -name "*.json" | wc -l)
+wav_count=$(find $adir -name "*.wav" | wc -l)
+
+if [ "$json_count" -eq 0 ]; then
+ echo >&2 "We expect that the directory $jdir will contain json files."
+ echo >&2 "That implies you have supplied a wrong path to the data."
+ exit 1
+fi
+if [ "$wav_count" -eq 0 ]; then
+ echo >&2 "We expect that the directory $adir will contain wav files."
+ echo >&2 "That implies you have supplied a wrong path to the data."
+ exit 1
+fi
+
 echo "$0: Converting transcription to text"
 
 mkdir -p $dir

egs/chime5/s5/local/check_tools.sh

@@ -87,7 +87,6 @@ awk '{print $1}' $dir/iv_lexicon.txt | \
  if (!defined $seen{$w}) { print; }
  } ' $dir/word_counts > $dir/oov_counts.txt
 
-set -x
 echo "*Highest-count OOVs (including fragments) are:"
 head -n 10 $dir/oov_counts.txt
 echo "*Highest-count OOVs (excluding fragments) are:"

egs/chime5/s5/local/prepare_data.sh

@@ -25,11 +25,9 @@ odir=$2
 array=$3
 expdir=exp/enhan/`echo $odir | awk -F '/' '{print $NF}'`_`echo $bmf | tr ' ' '_'`
 
-if [ -z $BEAMFORMIT ] ; then
- export BEAMFORMIT=$KALDI_ROOT/tools/BeamformIt
+if ! command -v BeamformIt &>/dev/null ; then
+ echo "Missing BeamformIt, run 'cd $KALDI_ROOT/tools/; ./extras/install_beamformit.sh; cd -;'" && exit 1
 fi
-export PATH=${PATH}:$BEAMFORMIT
-! hash BeamformIt && echo "Missing BeamformIt, run 'cd $KALDI_ROOT/tools/; ./extras/install_beamformit.sh; cd -;'" && exit 1
 
 # Set bash to 'debug' mode, it will exit on :
 # -e 'error', -u 'undefined variable', -o ... 'error in pipeline', -x 'print commands',
@@ -75,7 +73,7 @@ for n in `seq $nj`; do
 cat << EOF > $expdir/log/beamform.$n.sh
 while read line; do
  $BEAMFORMIT/BeamformIt -s \$line -c $input_arrays \
- --config_file `pwd`/conf/chime5.cfg \
+ --config_file `pwd`/conf/beamformit.cfg \
  --source_dir $sdir \
  --result_dir $odir
 done < $output_wavfiles.$n

egs/chime5/s5/local/prepare_dict.sh

@@ -40,20 +40,9 @@ tgtdir=$2
 ##End of configuration
 loc=`which ngram-count`;
 if [ -z $loc ]; then
- if uname -a | grep 64 >/dev/null; then # some kind of 64 bit...
- sdir=`pwd`/../../../tools/srilm/bin/i686-m64
- else
- sdir=`pwd`/../../../tools/srilm/bin/i686
- fi
- if [ -f $sdir/ngram-count ]; then
- echo Using SRILM tools from $sdir
- export PATH=$PATH:$sdir
- else
- echo You appear to not have SRILM tools installed, either on your path,
- echo or installed in $sdir. See tools/install_srilm.sh for installation
- echo instructions.
- exit 1
- fi
+ echo >&2 "You appear to not have SRILM tools installed, either on your path,"
+ echo >&2 "Use the script \$KALDI_ROOT/tools/install_srilm.sh to install it."
+ exit 1
 fi
 
 # Prepare the destination directory
@@ -122,7 +111,6 @@ else
  echo train.txt contains `cat $tgtdir/train.txt | perl -ane 'BEGIN{$w=$s=0;}{$w+=@F; $s++;}END{print "$w words, $s sentences\n";}'`
 fi
 
-set -x
 # Kaldi transcript files contain Utterance_ID as the first word; remove it
 sed -e "s/\.CH.//" -e "s/_.\-./_/" $dev_text | sort -u | \
  perl -ane 'print join(" ", @F[1..$#F]) . "\n" if @F > 1' > $tgtdir/dev.txt
@@ -247,7 +235,10 @@ if [ ! -z ${LIBLBFGS} ]; then
  ngram -lm - -order 4 -unk -map-unk "$oov_symbol" -prune-lowprobs -write-lm - |\
  sed 's/<unk>/'${oov_symbol}'/g' | gzip -c > $tgtdir/4gram.me.gz || exit 1
 else
- echo "Skipping MaxEnt models"
+ echo >&2 "SRILM is not compiled with the support of MaxEnt models."
+ echo >&2 "You should use the script in \$KALDI_ROOT/tools/install_srilm.sh"
+ echo >&2 "which will take care of compiling the SRILM with MaxEnt support"
+ exit 1;
 fi