geen xquery meer gebruikt, te traag

cmd/items2count/items2count.go

@@ -5,20 +5,13 @@ import (
 	// "github.com/kr/pretty"
 
 	"bufio"
-	"encoding/xml"
 	"fmt"
 	"os"
+	"regexp"
 	"sort"
 	"strings"
 )
 
-type Item struct {
-	XMLName xml.Name `xml:"i"`
-	Msg     string   `xml:"m"`
-	Tags    []string `xml:"t"`
-	Word    string   `xml:"w"`
-}
-
 type Word struct {
 	word    string
 	sortkey string
@@ -33,8 +26,10 @@ type Tag struct {
 }
 
 var (
-	x     = e.ExitErr
+	x         = e.ExitErr
-	words = make(map[string]*Word)
+	words     = make(map[string]*Word)
+	reTag     = regexp.MustCompile(`tag: "((?:\\.|[^\\"])*)"`)
+	reUnquote = regexp.MustCompile(`\\.`)
 
 	ignore = map[string]bool{
 		"Algemeen":  true,
@@ -48,26 +43,28 @@ func main() {
 
 	scanner := bufio.NewScanner(os.Stdin)
 	for scanner.Scan() {
-		var item Item
 		line := scanner.Text()
-		x(xml.Unmarshal([]byte(line), &item))
+		aa := strings.Split(line, "\t")
-		w, ok := words[item.Word]
+		word := aa[0]
+		tags := aa[1]
+		lbl := aa[2]
+		w, ok := words[word]
 		if !ok {
 			w = &Word{
-				word:    item.Word,
+				word:    word,
-				sortkey: strings.ToLower(item.Word),
+				sortkey: strings.ToLower(word),
 				tags:    make(map[string]map[string]int),
 			}
-			words[item.Word] = w
+			words[word] = w
 		}
 		w.count++
-		lbl := item.Msg[:strings.Index(item.Msg, ".")]
+		lbl = lbl[:strings.Index(lbl, ".")]
-		for _, tag := range item.Tags {
+		for _, tag := range parseTags(tags) {
 			if !ignore[tag] {
 				if _, ok := w.tags[lbl]; !ok {
 					w.tags[lbl] = make(map[string]int)
 				}
-				if tag != item.Word {
+				if tag != word {
 					w.tags[lbl][tag] = w.tags[lbl][tag] + 1
 				}
 			}
@@ -95,6 +92,15 @@ func main() {
 
 }
 
+func parseTags(s string) []string {
+	tags := make([]string, 0)
+	aa := reTag.FindAllStringSubmatch(s, -1)
+	for _, a := range aa {
+		tags = append(tags, unquote(a[1]))
+	}
+	return tags
+}
+
 func getTag(tags map[string]map[string]int) string {
 
 	all := make([]Tag, 0)
@@ -155,3 +161,9 @@ func getTag(tags map[string]map[string]int) string {
 
 	return strings.Join(aa, ", ")
 }
+
+func unquote(text string) string {
+	return reUnquote.ReplaceAllStringFunc(text, func(s string) string {
+		return s[1:]
+	})
+}

collect.sh

@@ -48,19 +48,43 @@ do
 
         # tellingen met tags
 
-        alto tq:../xquery/nieuwe_namen.xq   $files | sort | uniq | items2count > $part-nieuwe-namen-$ds-$i
+        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\t%d\t%I' $files \
+            | sed -e 's/\.[0-9][0-9]*$//' | sort | uniq \
+            | items2count > $part-nieuwe-namen-$ds-$i
         top20 $part-nieuwe-namen-$ds-$i
 
-        alto tq:../xquery/nieuwe_woorden.xq $files | sort | uniq | items2count > $part-nieuwe-woorden-$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%d\t%I' $files \
+            | sed -e 's/\.[0-9][0-9]*$//' | sort | uniq \
+            | items2count > $part-nieuwe-woorden-$ds-$i
         top20 $part-nieuwe-woorden-$ds-$i
 
-        alto tq:../xquery/locaties.xq       $files | sort | uniq | items2count > $part-locaties-$ds-$i
+        alto \
+            'fp://node[(@neclass="LOC" and not(@rel="mwp")) or (@cat="mwu" and node[@pt="spec" and @neclass="LOC"])]' \
+            'tt:%l\t%d\t%I' $files \
+            | sed -e 's/\.[0-9][0-9]*$//' | sort | uniq \
+            | items2count > $part-locaties-$ds-$i
 
-        alto tq:../xquery/personen.xq       $files | sort | uniq | items2count > $part-personen-$ds-$i
+        alto \
+            'fp://node[(@neclass="PER" and not(@rel="mwp")) or (@cat="mwu" and node[@pt="spec" and @neclass="PER"])]' \
+            'tt:%l\t%d\t%I' $files \
+            | sed -e 's/\.[0-9][0-9]*$//' | sort | uniq \
+            | items2count > $part-personen-$ds-$i
 
-        alto tq:../xquery/organisaties.xq   $files | sort | uniq | items2count > $part-organisaties-$ds-$i
+        alto \
+            'fp://node[(@neclass="ORG" and not(@rel="mwp")) or (@cat="mwu" and node[@pt="spec" and @neclass="ORG"])]' \
+            'tt:%l\t%d\t%I' $files \
+            | sed -e 's/\.[0-9][0-9]*$//' | sort | uniq \
+            | items2count > $part-organisaties-$ds-$i
 
-        alto tq:../xquery/overige_namen.xq  $files | sort | uniq | items2count > $part-overige-namen-$ds-$i
+        alto \
+            'fp://node[(@neclass="MISC" and not(@rel="mwp")) or (@cat="mwu" and node[@pt="spec" and @neclass="MISC"])]' \
+            'tt:%l\t%d\t%I' $files \
+            | sed -e 's/\.[0-9][0-9]*$//' | sort | uniq \
+            | items2count > $part-overige-namen-$ds-$i
 
         # tellingen met postags
 

xquery/README

@@ -0,0 +1,2 @@
+dit wordt niet meer gebruikt
+xquery is veel te traag

xquery/collect.sh.oud

@@ -0,0 +1,86 @@
+#!/bin/bash
+
+set -e
+
+unset CDPATH
+PATH=/net/corpora/nlnieuws/bin:/net/aps/bin:$PATH
+export TZ=Europe/Amsterdam
+
+if [ "$1" = "" ]
+then
+    ds=`ISOWeek -7`
+else
+    case "$1" in
+        2[0-9][0-9][0-9]-[0-5][0-9])
+            ds=$1
+            ;;
+        *)
+            echo INVALID
+            exit 1
+            ;;
+    esac
+fi
+
+cd /net/corpora/nlnieuws/data
+
+declare -A parts
+parts[alles]='.'
+parts[algemeen]='NOS|NU|NieuwsNL|RO|Sargasso'
+parts[groningen]='GG|Sikkom'
+parts[AT5]='AT5'
+parts[GG]='GG'
+parts[NOS]='NOS'
+parts[NU]='NU'
+parts[NieuwsNL]='NieuwsNL'
+parts[RO]='RO'
+parts[Sargasso]='Sargasso'
+parts[Sikkom]='Sikkom'
+parts[Tzum]='Tzum'
+parts[VRT]='VRT'
+
+for part in ${!parts[@]}
+do
+    regex=${parts[$part]}
+
+    for i in 1 4
+    do
+        files=$(find .. $(week2files $ds $i) | grep -E "$regex")
+
+        # tellingen met tags
+
+        alto tq:../xquery/nieuwe_namen.xq   $files | sort | uniq | items2count > $part-nieuwe-namen-$ds-$i
+        top20 $part-nieuwe-namen-$ds-$i
+
+        alto tq:../xquery/nieuwe_woorden.xq $files | sort | uniq | items2count > $part-nieuwe-woorden-$ds-$i
+        top20 $part-nieuwe-woorden-$ds-$i
+
+        alto tq:../xquery/locaties.xq       $files | sort | uniq | items2count > $part-locaties-$ds-$i
+
+        alto tq:../xquery/personen.xq       $files | sort | uniq | items2count > $part-personen-$ds-$i
+
+        alto tq:../xquery/organisaties.xq   $files | sort | uniq | items2count > $part-organisaties-$ds-$i
+
+        alto tq:../xquery/overige_namen.xq  $files | sort | uniq | items2count > $part-overige-namen-$ds-$i
+
+        # tellingen met postags
+
+        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\t%I' $files \
+            | sed -e 's/\.[0-9][0-9]*$//' | sort | uniq \
+            | sed -e 's/\(.*\)\t.*/\1/' | uniq -c \
+            | grep -v '^ *1 ' \
+            | sed -e 's/\([0-9]\)  */\1\t/' | sort -f -k 2 | sort -n -r -k 1,1 -s \
+            > $part-nieuwe-woorden-extra-$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\t%I' $files \
+            | sed -e 's/\.[0-9][0-9]*$//' | sort | uniq \
+            | sed -e 's/\(.*\)\t.*/\1/' | uniq -c \
+            | grep -v '^ *1 ' \
+            | sed -e 's/\([0-9]\)  */\1\t/' | sort -f -k 2 | sort -n -r -k 1,1 -s \
+            > $part-nieuwe-adjww-extra-$ds-$i
+    done
+done

xquery/items2count.go.oud

@@ -0,0 +1,157 @@
+package main
+
+import (
+	e "codeberg.org/pebbe/errors"
+	// "github.com/kr/pretty"
+
+	"bufio"
+	"encoding/xml"
+	"fmt"
+	"os"
+	"sort"
+	"strings"
+)
+
+type Item struct {
+	XMLName xml.Name `xml:"i"`
+	Msg     string   `xml:"m"`
+	Tags    []string `xml:"t"`
+	Word    string   `xml:"w"`
+}
+
+type Word struct {
+	word    string
+	sortkey string
+	count   int
+	tags    map[string]map[string]int
+}
+
+type Tag struct {
+	tag     string
+	sortkey string
+	count   int
+}
+
+var (
+	x     = e.ExitErr
+	words = make(map[string]*Word)
+
+	ignore = map[string]bool{
+		"Algemeen":  true,
+		"Artikelen": true,
+		"Nieuws":    true,
+		"Recensies": true,
+	}
+)
+
+func main() {
+
+	scanner := bufio.NewScanner(os.Stdin)
+	for scanner.Scan() {
+		var item Item
+		line := scanner.Text()
+		x(xml.Unmarshal([]byte(line), &item))
+		w, ok := words[item.Word]
+		if !ok {
+			w = &Word{
+				word:    item.Word,
+				sortkey: strings.ToLower(item.Word),
+				tags:    make(map[string]map[string]int),
+			}
+			words[item.Word] = w
+		}
+		w.count++
+		lbl := item.Msg[:strings.Index(item.Msg, ".")]
+		for _, tag := range item.Tags {
+			if !ignore[tag] {
+				if _, ok := w.tags[lbl]; !ok {
+					w.tags[lbl] = make(map[string]int)
+				}
+				if tag != item.Word {
+					w.tags[lbl][tag] = w.tags[lbl][tag] + 1
+				}
+			}
+		}
+	}
+	x(scanner.Err())
+
+	wordlist := make([]*Word, 0, len(words))
+	for _, value := range words {
+		if value.count > 1 {
+			wordlist = append(wordlist, value)
+		}
+	}
+
+	sort.Slice(wordlist, func(a, b int) bool {
+		if wordlist[a].count != wordlist[b].count {
+			return wordlist[a].count > wordlist[b].count
+		}
+		return wordlist[a].sortkey < wordlist[b].sortkey
+	})
+
+	for _, w := range wordlist {
+		fmt.Printf("%6d\t%s\t%s\n", w.count, w.word, getTag(w.tags))
+	}
+
+}
+
+func getTag(tags map[string]map[string]int) string {
+
+	all := make([]Tag, 0)
+
+	for _, tagv := range tags {
+		n := 0
+		tt := make([]string, 0)
+		for key, value := range tagv {
+			if value > n {
+				n = value
+				tt = []string{key}
+			} else if value == n {
+				tt = append(tt, key)
+			}
+		}
+		for _, t := range tt {
+			all = append(all, Tag{tag: t, count: n, sortkey: strings.ToLower(t)})
+		}
+	}
+
+	sort.Slice(all, func(a, b int) bool {
+		if all[a].count != all[b].count {
+			return all[a].count > all[b].count
+		}
+		if all[a].sortkey != all[b].sortkey {
+			return all[a].sortkey < all[b].sortkey
+		}
+		return all[a].tag < all[b].tag
+	})
+
+	needSort := false
+	for i := 1; i < len(all); i++ {
+		if all[i-1].sortkey == all[i].sortkey {
+			all[i-1].count += all[i].count
+			all = append(all[:i], all[i+1:]...)
+			i--
+			needSort = true
+		}
+	}
+	if needSort {
+		sort.Slice(all, func(a, b int) bool {
+			if all[a].count != all[b].count {
+				return all[a].count > all[b].count
+			}
+			if all[a].sortkey != all[b].sortkey {
+				return all[a].sortkey < all[b].sortkey
+			}
+			return all[a].tag < all[b].tag
+		})
+	}
+
+	aa := make([]string, 0, len(all))
+	for _, n := range all {
+		if n.count > 1 {
+			aa = append(aa, n.tag)
+		}
+	}
+
+	return strings.Join(aa, ", ")
+}