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

Tzum: <em>Titel</em> -> "Titel"

Browse Source
This commit is contained in:
Peter Kleiweg
2026-06-03 17:32:01 +02:00
parent 14590570ba
commit efa301cc4a
1 changed files with 21 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

23
Tzum/cmd/xml2txt/xml2txt.go
View File

@@ -24,6 +24,8 @@ var (
x = e.ExitErr x = e.ExitErr
reYearWeek = regexp.MustCompile(`^2[0-9][0-9][0-9]\.[0-5][0-9]$`) reYearWeek = regexp.MustCompile(`^2[0-9][0-9][0-9]\.[0-5][0-9]$`)
reEM = regexp.MustCompile(`::EM::.*?::/EM::`)
reTitle = regexp.MustCompile(`^\p{Lu}`)
) )
func main() { func main() {
@@ -66,7 +68,7 @@ func main() {
x(fmt.Fprintf(fp, "##META text tag = %s\n", t)) x(fmt.Fprintf(fp, "##META text tag = %s\n", t))
} }
x(fp.WriteString(u.AddEnd(u.FixSpace(item.Title)))) x(fp.WriteString(u.AddEnd(u.FixSpace(item.Title))))
doc, err := gokogiri.ParseHtml([]byte(`<html><body>` + u.HtmlFixString(item.Text) + `</body></html>`)) doc, err := gokogiri.ParseHtml([]byte(`<html><body>` + em1(u.HtmlFixString(item.Text)) + `</body></html>`))
x(err) x(err)
root := doc.Root() root := doc.Root()
pp, err := root.Search(`//body/p`) pp, err := root.Search(`//body/p`)
@@ -74,9 +76,26 @@ func main() {
for _, p := range pp { for _, p := range pp {
s := p.Content() s := p.Content()
if !strings.Contains(s, "verscheen eerst op Tzum.") { if !strings.Contains(s, "verscheen eerst op Tzum.") {
x(fp.WriteString(u.AddEnd(u.FixSpace(p.Content())))) x(fp.WriteString(em2(u.AddEnd(u.FixSpace(p.Content())))))
} }
} }
x(fp.Close()) x(fp.Close())
} }
} }
func em1(s string) string {
return strings.ReplaceAll(
strings.ReplaceAll(s, "<em>", " ::EM::"),
"</em>",
"::/EM:: ")
}
func em2(s string) string {
return reEM.ReplaceAllStringFunc(s, func(s1 string) string {
s1 = s1[6 : len(s1)-7]
if reTitle.MatchString(s1) {
return `"` + s1 + `"`
}
return s1
})
}