grote reorganisatie:

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

Sargasso/cmd/sargasso/sargasso.go

@@ -77,7 +77,7 @@ func main() {
 		}
 		p(err)
 		year, week := t.ISOWeek()
-		dirname := fmt.Sprintf("/net/corpora/nlnieuws/Sargasso/%d/%02d", year, week)
+		dirname := fmt.Sprintf("/net/corpora/nlnieuws/Sargasso/%d/w%02d", year, week)
 		if exists(dirname + "/lock") {
 			continue
 		}

Sargasso/cmd/xml2txt/xml2txt.go

@@ -23,7 +23,7 @@ type Item struct {
 var (
 	x = e.ExitErr
 
-	reYearWeek = regexp.MustCompile(`^2[0-9][0-9][0-9]-[0-5][0-9]$`)
+	reYearWeek = regexp.MustCompile(`^2[0-9][0-9][0-9]\.[0-5][0-9]$`)
 )
 
 func main() {
@@ -32,16 +32,16 @@ func main() {
 	switch len(os.Args) {
 	case 1:
 		year, week := time.Now().AddDate(0, 0, -7).ISOWeek()
-		ds = fmt.Sprintf("%d-%02d", year, week)
+		ds = fmt.Sprintf("%d.%02d", year, week)
 	case 2:
 		if !reYearWeek.MatchString(os.Args[1]) {
-			x(fmt.Errorf("arg must be yyyy-ww"))
+			x(fmt.Errorf("arg must be yyyy.ww"))
 		}
 		ds = os.Args[1]
 	default:
 		x(fmt.Errorf("too many arguments"))
 	}
-	dp := ds[:4] + "/" + ds[5:]
+	dp := ds[:4] + "/w" + ds[5:]
 
 	x(os.Chdir("/net/corpora/nlnieuws/Sargasso/" + dp))
 	x(os.MkdirAll("out", 0777))

Sargasso/txt2corpus.sh

@@ -2,17 +2,20 @@
 
 set -e
 
+BASE=/net/corpora/nlnieuws
+PART=$BASE/Sargasso
+
 unset CDPATH
-PATH=/net/corpora/nlnieuws/bin:/net/aps/bin:$PATH
+PATH=$PART:$BASE/bin:$BASE:/net/aps/bin:$PATH
 export TZ=Europe/Amsterdam
 . /net/aps/etc/alpino-activate.sh > /dev/null
 
 if [ "$1" = "" ]
 then
-    ds=`date -d -7days +%G-%V`
+    ds=`date -d -7days +%G.%V`
 else
     case "$1" in
-        2[0-9][0-9][0-9]-[0-5][0-9])
+        2[0-9][0-9][0-9].[0-5][0-9])
             ds=$1
             ;;
         *)
@@ -22,11 +25,13 @@ else
     esac
 fi
 
-dp=${ds//-//}
+year=${ds%.*}
+week=${ds#*.}
+dp=$year/w$week
+corpus=$PART/corpus/$year/$ds
+mkdir -p $PART/corpus/$year
 
-corpus=/net/corpora/nlnieuws/Sargasso/corpus/$ds
-
-cd /net/corpora/nlnieuws/Sargasso/$dp
+cd $PART/$dp
 
 ln -s lock.$$ lock
 if [ "`readlink lock`" != lock.$$ ]
@@ -38,7 +43,7 @@ fi
 rm -fr out
 mkdir out
 
-../../xml2txt $ds
+xml2txt $ds
 
 rm -f $corpus.lines
 for i in out/*.txt
@@ -53,14 +58,14 @@ 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