p209327/nlnieuws
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

score; collect.sh

Browse Source
This commit is contained in:
Peter Kleiweg
2026-03-19 12:37:35 +01:00
parent 01438b69ac
commit 025c134c07
5 changed files with 174 additions and 34 deletions
Download Patch File Download Diff File Expand all files Collapse all files

77
collect.sh
View File

@@ -21,40 +21,61 @@ else
esac
fi
files=$(find /net/corpora/nlnieuws/ $(week2files $ds))
cd /net/corpora/nlnieuws/data
alto \
'fp://node[((@cat="mwu" and node[@pt="spec"]) or (@pt and @*="eigen" and not(@rel="mwp"))) and not(@his="normal") and not(@his_1="decap" or @his_1="0")]' \
tt:%w $files | sort | uniq -c | grep -v '^ *1 ' | sort -nr > nieuw-namen-$ds
for i in 1 4
do
alto \
'fp://node[@his and not(@rel="mwp" or @cat="mwu") and not(@his="normal" or @his="name" or @his="prefix_name" or @his_1="decap" or @
his_1="0" or @his="skip" or @his="robust_skip" or @his="w_dia" or @his="wo_dia" or @his="within_word_conjunct")]' \
tt:%w $files | sort | uniq -c | grep -v '^ *1 ' | sort -nr > nieuw-woorden-$ds
files=$(find .. $(week2files $ds $i))
alto \
'fp://node[@his and not(@rel="mwp" or @cat="mwu") and not(@his="normal" or @his="name" or @his="prefix_name" or @his_1="decap" or @
his_1="0" or @his="skip" or @his="robust_skip" or @his="w_dia" or @his="wo_dia" or @his="within_word_conjunct")]' \
'tt:%w\t%l\t%P' $files | sort | uniq -c | grep -v '^ *1 ' | sort -nr > nieuw-extra-$ds
alto \
'fp://node[((@cat="mwu" and node[@pt="spec"]) or (@pt and @*="eigen" and not(@rel="mwp"))) and not(@his="normal") and not(@his_1="decap" or @his_1="0")]' \
tt:%w $files | sort | uniq -c | grep -v '^ *1 ' | sed -e 's/\([0-9]\) */\1\t/' | sort -nr > nieuw-namen-$ds-$i
alto \
'fp://node[(@neclass="LOC" and @his="normal" and not(@rel="mwp")) or (@cat="mwu" and node[@pt="spec" and @neclass="LOC"] and @his="normal")]' \
tt:%l $files | sort | uniq -c | grep -v '^ *1 ' | sort -nr > locaties-$ds
score nieuw-namen-$ds-$i > nieuw-namen-$ds-$i.score
alto \
'fp://node[(@neclass="PER" and @his="normal" and not(@rel="mwp")) or (@cat="mwu" and node[@pt="spec" and @neclass="PER"] and @his="normal")]' \
tt:%l $files | sort | uniq -c | grep -v '^ *1 ' | sort -nr > personen-$ds
alto \
'fp://node[@his and not(@rel="mwp" or @cat="mwu") and not(@his="normal" or @his="name" or @his="prefix_name" or @his_1="decap" or @
his_1="0" or @his="skip" or @his="robust_skip" or @his="w_dia" or @his="wo_dia" or @his="within_word_conjunct")]' \
tt:%w $files | sort | uniq -c | grep -v '^ *1 ' | sed -e 's/\([0-9]\) */\1\t/' | sort -nr > nieuw-woorden-$ds-$i
alto \
'fp://node[(@neclass="ORG" and @his="normal" and not(@rel="mwp")) or (@cat="mwu" and node[@pt="spec" and @neclass="ORG"] and @his=" normal")]' \
tt:%l $files | sort | uniq -c | grep -v '^ *1 ' | sort -nr > organisatie-$ds
score nieuw-woorden-$ds-$i > nieuw-woorden-$ds-$i.score
alto \
'fp://node[(@neclass="MISC" and @his="normal" and not(@rel="mwp")) or (@cat="mwu" and node[@pt="spec" and @neclass="MISC"] and @his="normal")]' \
tt:%l $files | sort | uniq -c | grep -v '^ *1 ' | sort -nr > overigen-$ds
alto \
'fp://node[@his and not(@rel="mwp" or @cat="mwu") and not(@his="normal" or @his="name" or @his="prefix_name" or @his_1="decap" or @
his_1="0" or @his="skip" or @his="robust_skip" or @his="w_dia" or @his="wo_dia" or @his="within_word_conjunct")]' \
'tt:%w\t%l\t%P' $files | sort | uniq -c | grep -v '^ *1 ' | sed -e 's/\([0-9]\) */\1\t/' | sort -nr > nieuw-extra-$ds-$i
alto \
'fp://node[@pt and @his and not(../@his="normal" or @rel="mwp" or ../@his="name" or ../@his_1="decap") and not(@his="normal" or @his="name" or @his="skip" or @his="robust_skip" or @his="w_dia" or @his="wo_dia" or @his="decap" or @his="within_word_conjunct") and not(@pt="n") ]' \
'tt:%w\t%P' $files | sort | uniq -c | grep -v '^ *1 ' | sort -nr > overigen-extra-$ds
score nieuw-extra-$ds-$i > nieuw-extra-$ds-$i.score
alto \
'fp://node[(@neclass="LOC" and @his="normal" and not(@rel="mwp")) or (@cat="mwu" and node[@pt="spec" and @neclass="LOC"] and @his="normal")]' \
tt:%l $files | sort | uniq -c | grep -v '^ *1 ' | sed -e 's/\([0-9]\) */\1\t/' | sort -nr > locaties-$ds-$i
score locaties-$ds-$i > locaties-$ds-$i.score
alto \
'fp://node[(@neclass="PER" and @his="normal" and not(@rel="mwp")) or (@cat="mwu" and node[@pt="spec" and @neclass="PER"] and @his="normal")]' \
tt:%l $files | sort | uniq -c | grep -v '^ *1 ' | sed -e 's/\([0-9]\) */\1\t/' | sort -nr > personen-$ds-$i
score personen-$ds-$i > personen-$ds-$i.score
alto \
'fp://node[(@neclass="ORG" and @his="normal" and not(@rel="mwp")) or (@cat="mwu" and node[@pt="spec" and @neclass="ORG"] and @his=" normal")]' \
tt:%l $files | sort | uniq -c | grep -v '^ *1 ' | sed -e 's/\([0-9]\) */\1\t/' | sort -nr > organisaties-$ds-$i
score organisaties-$ds-$i > organisaties-$ds-$i.score
alto \
'fp://node[(@neclass="MISC" and @his="normal" and not(@rel="mwp")) or (@cat="mwu" and node[@pt="spec" and @neclass="MISC"] and @his="normal")]' \
tt:%l $files | sort | uniq -c | grep -v '^ *1 ' | sed -e 's/\([0-9]\) */\1\t/' | sort -nr > overigen-$ds-$i
score overigen-$ds-$i > overigen-$ds-$i.score
alto \
'fp://node[@pt and @his and not(../@his="normal" or @rel="mwp" or ../@his="name" or ../@his_1="decap") and not(@his="normal" or @his="name" or @his="skip" or @his="robust_skip" or @his="w_dia" or @his="wo_dia" or @his="decap" or @his="within_word_conjunct") and not(@pt="n") ]' \
'tt:%w\t%P' $files | sort | uniq -c | grep -v '^ *1 ' | sed -e 's/\([0-9]\) */\1\t/' | sort -nr > overigen-extra-$ds-$i
score overigen-extra-$ds-$i > overigen-extra-$ds-$i.score
done