first commit

This commit is contained in:
Peter Kleiweg
2026-03-02 15:34:37 +01:00
commit 36f051a8a9
50 changed files with 4241 additions and 0 deletions

13
Tzum/Makefile Normal file
View File

@@ -0,0 +1,13 @@
all: \
xml2txt \
metadata \
tzum
xml2txt: cmd/xml2txt/*.go
go build -o $@ $^
metadata: cmd/metadata/*.go
go build -o $@ $^
tzum: cmd/tzum/*.go
go build -o $@ $^

View File

@@ -0,0 +1,136 @@
package main
import (
"github.com/pebbe/util"
"bufio"
"encoding/xml"
"fmt"
"html"
"os"
"strings"
"time"
)
type Item struct {
XMLName xml.Name `xml:"item"`
UnixTime int64 `xml:"unixTime"`
}
var (
x = util.CheckErr
escape = html.EscapeString
data = make(map[string][]string)
location *time.Location
)
func main() {
var err error
location, err = time.LoadLocation("Europe/Amsterdam")
x(err)
files, err := os.ReadDir(".")
x(err)
for _, file := range files {
filename := file.Name()
if strings.HasSuffix(filename, ".txt") {
doText("", filename)
} else if strings.HasSuffix(filename, ".xml") {
doXml("", filename)
}
}
files, err = os.ReadDir("..")
x(err)
for _, file := range files {
filename := file.Name()
if strings.HasSuffix(filename, ".txt") {
doText("../", filename)
} else if strings.HasSuffix(filename, ".xml") {
doXml("../", filename)
}
}
files, err = os.ReadDir("xml")
x(err)
for _, file := range files {
filename := file.Name()
if !strings.HasSuffix(filename, ".xml") {
continue
}
aa := strings.Split(filename, ".")
base := strings.Join(aa[1:len(aa)-2], ".")
b, err := os.ReadFile("xml/" + filename)
x(err)
s := string(b)
i := strings.Index(s, "<alpino") + 1
i += strings.Index(s[i:], "<")
fp, err := os.Create("xml/" + filename + ".tmp")
x(err)
_, err = fp.WriteString(s[:i])
x(err)
_, err = fp.WriteString("<metadata>\n <meta type=\"text\" name=\"source\" value=\"Tzum\"/>\n")
x(err)
for _, m := range data[base] {
_, err = fp.WriteString(" " + m + "\n")
x(err)
}
_, err = fp.WriteString(" </metadata>\n ")
x(err)
_, err = fp.WriteString(stripMeta(s[i:]))
x(err)
x(fp.Close())
x(os.Rename("xml/"+filename+".tmp", "xml/"+filename))
}
}
func doText(dirname, filename string) {
base := filename[:len(filename)-4]
if _, ok := data[base]; !ok {
data[base] = make([]string, 0)
}
fp, err := os.Open(dirname + filename)
x(err)
defer func() { x(fp.Close()) }()
scanner := bufio.NewScanner(fp)
for scanner.Scan() {
line := scanner.Text()
if !strings.HasPrefix(line, "##META") {
continue
}
aa := strings.Fields(line)
if len(aa) > 4 {
data[base] = append(data[base],
fmt.Sprintf(`<meta type="%s" name="%s" value="%s"/>`,
aa[1],
escape(aa[2]),
escape(strings.Join(aa[4:], " "))))
}
}
x(scanner.Err())
}
func doXml(dirname, filename string) {
base := filename[:len(filename)-4]
if _, ok := data[base]; !ok {
data[base] = make([]string, 0)
}
b, err := os.ReadFile(dirname + filename)
x(err)
var item Item
x(xml.Unmarshal(b, &item))
t := time.Unix(item.UnixTime, 0).In(location)
data[base] = append(data[base],
fmt.Sprintf(`<meta type="date" name="pubdate" value="%d-%02d-%02d"/>`,
t.Year(),
int(t.Month()),
t.Day()))
}
func stripMeta(s string) string {
i1 := strings.Index(s, "<metadata>")
if i1 < 0 {
return s
}
i2 := i1 + strings.Index(s[i1:], "</metadata>") + 11
return s[:i1] + strings.TrimLeft(s[i2:], " \t\r\n")
}

81
Tzum/cmd/tzum/tzum.go Normal file
View File

@@ -0,0 +1,81 @@
package main
import (
"github.com/pebbe/util"
"encoding/xml"
"fmt"
"io"
"net/http"
"net/url"
"os"
"strings"
"time"
)
type Rss struct {
XMLName xml.Name `xml:"rss"`
Items []ItemT `xml:"channel>item"`
}
type ItemT struct {
PubDate string `xml:"pubDate"`
UnixTime int64 `xml:"unixTime"`
Guid string `xml:"guid"`
Data []byte `xml:",innerxml"`
}
var (
x = util.CheckErr
agent = "AhrefsBot/7.0"
)
func main() {
req, err := http.NewRequest("GET", "https://www.tzum.info/feed/", nil)
x(err)
req.Header.Set("User-Agent", agent)
client := &http.Client{}
resp, err := client.Do(req)
x(err)
body, err := io.ReadAll(resp.Body)
x(err)
x(resp.Body.Close())
var rss Rss
x(xml.Unmarshal(body, &rss))
if len(rss.Items) == 0 {
x(fmt.Errorf("len(rss.Items) == 0"))
}
for _, item := range rss.Items {
t, err := time.Parse(time.RFC1123Z, item.PubDate)
if err != nil {
t, err = time.Parse(time.RFC1123, item.PubDate)
}
x(err)
year, week := t.ISOWeek()
dirname := fmt.Sprintf("/net/corpora/nlnieuws/Tzum/%d/%02d", year, week)
basename := strings.TrimPrefix(item.Guid, "https://www.tzum.info/?p=")
if i := strings.LastIndex(basename, "/"); i > 0 {
basename = basename[:i]
}
filename := dirname + "/" + url.PathEscape(basename)
x(os.MkdirAll(dirname, 0777))
fp, err := os.Create(filename + ".xml")
x(err)
_, err = fp.WriteString("<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n<item>\n")
x(err)
_, err = fmt.Fprintf(fp, "<unixTime>%d</unixTime>", t.Unix())
x(err)
_, err = fp.Write(item.Data)
x(err)
_, err = fp.WriteString("</item>\n")
x(err)
x(fp.Close())
x(os.Chtimes(filename+".xml", t, t))
}
}

100
Tzum/cmd/xml2txt/xml2txt.go Normal file
View File

@@ -0,0 +1,100 @@
package main
import (
"github.com/jbowtie/gokogiri"
"github.com/pebbe/util"
"encoding/xml"
"fmt"
"os"
"regexp"
"strings"
"time"
)
type Item struct {
Title string `xml:"title"`
Text string `xml:"encoded"`
Cats []string `xml:"category"`
}
var (
x = util.CheckErr
reYearWeek = regexp.MustCompile(`^2[0-9][0-9][0-9]-[0-5][0-9]$`)
)
func main() {
var ds string
switch len(os.Args) {
case 1:
year, week := time.Now().AddDate(0, 0, -7).ISOWeek()
ds = fmt.Sprintf("%d-%02d", year, week)
case 2:
if !reYearWeek.MatchString(os.Args[1]) {
x(fmt.Errorf("arg must be yyyy-ww"))
}
ds = os.Args[1]
default:
x(fmt.Errorf("too many arguments"))
}
dp := ds[:4] + "/" + ds[5:]
x(os.Chdir("/net/corpora/nlnieuws/Tzum/" + dp))
x(os.MkdirAll("out", 0777))
files, err := os.ReadDir(".")
x(err)
for _, file := range files {
filename := file.Name()
if !strings.HasSuffix(filename, ".xml") {
continue
}
b, err := os.ReadFile(filename)
x(err)
fp, err := os.Create("out/" + filename[:len(filename)-4] + ".txt")
x(err)
var item Item
x(xml.Unmarshal(b, &item))
for _, cat := range item.Cats {
_, err = fmt.Fprintf(fp, "##META text cat = %s\n", cat)
x(err)
}
_, err = fp.WriteString(addEnd(item.Title))
x(err)
doc, err := gokogiri.ParseHtml([]byte(`<html><body>` + item.Text + `</body></html>`))
x(err)
root := doc.Root()
pp, err := root.Search(`//body/p`)
x(err)
for _, p := range pp {
s := p.Content()
if !strings.Contains(s, "verscheen eerst op Tzum.") {
_, err = fp.WriteString(addEnd(p.Content()))
x(err)
}
}
x(err)
x(fp.Close())
}
}
func addEnd(s string) string {
s = strings.TrimSpace(s)
n := len(s)
if n == 0 {
return ""
}
if n > 0 {
if strings.ContainsAny(s[n-1:], ".!?") {
return s + "\n"
}
}
if n > 1 {
s2 := s[n-2:]
if s2 == `."` || s2 == `!"` || s2 == `?"` || s2 == `.'` || s2 == `!'` || s2 == `?'` {
return s + "\n"
}
}
return s + ".\n"
}

65
Tzum/txt2corpus.sh Executable file
View File

@@ -0,0 +1,65 @@
#!/bin/bash
set -e
unset CDPATH
PATH=/net/corpora/nlnieuws/bin:/net/aps/bin:$PATH
export TZ=Europe/Amsterdam
. /net/aps/etc/alpino-activate.sh > /dev/null
if [ "$1" = "" ]
then
ds=`ISODate -7`
else
case "$1" in
2[0-9][0-9][0-9]-[0-5][0-9])
ds=$1
;;
*)
echo INVALID
exit 1
;;
esac
fi
dp=${ds//-//}
corpus=/net/corpora/nlnieuws/Tzum/corpus/$ds
cd /net/corpora/nlnieuws/Tzum/$dp
ln -s lock.$$ lock
if [ "`readlink lock`" != lock.$$ ]
then
echo Getting lock failed
exit 1
fi
rm -fr out
mkdir out
../../xml2txt $ds
rm -f $corpus.lines
for i in out/*.txt
do
b=`basename $i .txt`
perl -p -e 's/^\s*//; s/^##META.*\n//' $i | tokenize.sh \
| perl -e '$n = 0; while(<>) { $n++; print("tzum.'$b'.$n|$_"); }' \
>> $corpus.lines
done
cd out
mkdir xml
Alpino -flag treebank xml debug=1 end_hook=xml user_max=900000 -parse < $corpus.lines 2> $corpus.log
../../../metadata 2> err
rm err
cd xml
alto -o $corpus.data.dz *.xml 2> /dev/null
cd ../..
rm -fr out
rm -f lock