grote reorganisatie:

- HLN, NOS, NU, VRT: per week -> per dag
- yyyy-ww -> yyyy.ww
- yyyy*  -> yyyy/yyyy*
etc

NieuwsNL/txt2corpus.sh

@@ -2,8 +2,11 @@
 
 set -e
 
+BASE=/net/corpora/nlnieuws
+PART=$BASE/NieuwsNL
+
 unset CDPATH
-PATH=/net/corpora/nlnieuws/bin:/net/aps/bin:$PATH
+PATH=$PART:$BASE/bin:$BASE:/net/aps/bin:$PATH
 export TZ=Europe/Amsterdam
 . /net/aps/etc/alpino-activate.sh > /dev/null
 
@@ -25,10 +28,11 @@ else
 fi
 
 dp=${ds//-//}
+year=${ds%%-*}
+corpus=$PART/corpus/$year/$ds
+mkdir -p $PART/corpus/$year
 
-corpus=/net/corpora/nlnieuws/NieuwsNL/corpus/$ds
-
-cd /net/corpora/nlnieuws/NieuwsNL/$dp
+cd $PART/$dp
 
 ln -s lock.$$ lock
 if [ "`readlink lock`" != lock.$$ ]
@@ -53,14 +57,14 @@ cd out
 mkdir xml
 Alpino -flag treebank xml debug=1 end_hook=xml user_max=900000 -parse < $corpus.lines 2> $corpus.log
 
-../../../../metadata
+metadata
 
 cd xml
 rm -f $corpus.data.dz $corpus.index
 alto -q -o $corpus.data.dz *.xml
 
 # telling per bericht, niet per zin
-/net/corpora/nlnieuws/namen.sh -x T -s $corpus.data.dz > $corpus.tag.txt
+query.sh -x T -s $corpus.data.dz > $corpus.tag.txt
 
 cd ../..
 rm -fr out