304 lines
6.5 KiB
Go
304 lines
6.5 KiB
Go
package main
|
|
|
|
import (
|
|
e "codeberg.org/pebbe/errors"
|
|
"github.com/jbowtie/gokogiri"
|
|
"golang.org/x/net/publicsuffix"
|
|
|
|
u "git.web.rug.nl/p209327/nlnieuws/internal/util"
|
|
|
|
"bytes"
|
|
"encoding/json"
|
|
"encoding/xml"
|
|
"errors"
|
|
"fmt"
|
|
"io"
|
|
"net/http"
|
|
"net/http/cookiejar"
|
|
"net/url"
|
|
"os"
|
|
"strings"
|
|
"time"
|
|
)
|
|
|
|
type Rss struct {
|
|
XMLName xml.Name `xml:"rss"`
|
|
Items []ItemT `xml:"channel>item"`
|
|
}
|
|
|
|
type ItemT struct {
|
|
Title string `xml:"title"`
|
|
PubDate string `xml:"pubDate"`
|
|
UnixTime int64 `xml:"unixTime"`
|
|
Guid string `xml:"guid"`
|
|
Link string `xml:"https://www.rssboard.org/rss-specification link"` // i.p.v. atom:link
|
|
Premium bool `xml:"premium"`
|
|
Data []byte `xml:",innerxml"`
|
|
}
|
|
|
|
type LdJson struct {
|
|
Type string `json:"@type"`
|
|
IsFree string `json:"isAccessibleForFree"`
|
|
H1 string `json:"headline"`
|
|
H2 string `json:"description"`
|
|
}
|
|
|
|
var (
|
|
p = e.PanicErr
|
|
w = e.WarnErr
|
|
agent = "AhrefsBot/7.0"
|
|
// agent = "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/145.0.0.0 Safari/537.36"
|
|
jar *cookiejar.Jar
|
|
)
|
|
|
|
func exists(filename string) bool {
|
|
_, err := os.Stat(filename)
|
|
return err == nil
|
|
}
|
|
|
|
func fileDate(filename string) string {
|
|
b, err := os.ReadFile(filename)
|
|
if err != nil {
|
|
return ""
|
|
}
|
|
s := string(b)
|
|
i1 := strings.Index(s, "<unixTime>") + 10
|
|
i2 := strings.Index(s, "</unixTime>")
|
|
if i2 < i1 {
|
|
return ""
|
|
}
|
|
return s[i1:i2]
|
|
}
|
|
|
|
func main() {
|
|
defer func() {
|
|
if e.Panicked {
|
|
_ = recover()
|
|
os.Exit(1)
|
|
}
|
|
}()
|
|
|
|
var err error
|
|
jar, err = cookiejar.New(&cookiejar.Options{PublicSuffixList: publicsuffix.List})
|
|
p(err)
|
|
|
|
myLock := "/net/corpora/nlnieuws/Sikkom/lock"
|
|
u.MkLock(myLock)
|
|
defer func() {
|
|
_ = os.Remove(myLock)
|
|
}()
|
|
|
|
resp, err := http.Get("https://www.sikkom.nl/rss")
|
|
p(err)
|
|
body, err := io.ReadAll(resp.Body)
|
|
p(err)
|
|
p(resp.Body.Close())
|
|
|
|
// zet de default namespace om onderscheid te kunnen maken met overige namespaces
|
|
b2 := bytes.Replace(body, []byte("<rss"), []byte(`<rss xmlns="https://www.rssboard.org/rss-specification"`), 1)
|
|
|
|
var rss Rss
|
|
p(xml.Unmarshal(b2, &rss))
|
|
|
|
if len(rss.Items) == 0 {
|
|
p(fmt.Errorf("len(rss.Items) == 0"))
|
|
}
|
|
|
|
for _, item := range rss.Items {
|
|
if item.Premium {
|
|
continue
|
|
}
|
|
t, err := time.Parse(time.RFC1123Z, item.PubDate)
|
|
if err != nil {
|
|
t, err = time.Parse(time.RFC1123, item.PubDate)
|
|
}
|
|
p(err)
|
|
year, week := t.ISOWeek()
|
|
dirname := fmt.Sprintf("/net/corpora/nlnieuws/Sikkom/%d/w%02d", year, week)
|
|
if exists(dirname + "/lock") {
|
|
continue
|
|
}
|
|
filename := dirname + "/" + url.PathEscape(item.Guid)
|
|
|
|
ts := fmt.Sprintf("%d", t.Unix())
|
|
needUpdate := fileDate(filename+".xml") != ts
|
|
|
|
p(os.MkdirAll(dirname, 0777))
|
|
func() {
|
|
var ok bool
|
|
defer func() {
|
|
if e.Panicked {
|
|
fmt.Fprintln(os.Stderr, "----", filename)
|
|
fmt.Fprintln(os.Stderr, "----", item.Link)
|
|
}
|
|
if !ok {
|
|
_ = os.Remove(filename + ".xml")
|
|
}
|
|
}()
|
|
fp, err := os.Create(filename + ".xml")
|
|
p(err)
|
|
p(fp.WriteString("<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n<item>\n"))
|
|
p(fmt.Fprintf(fp, "<unixTime>%d</unixTime>", t.Unix()))
|
|
p(fp.Write(item.Data))
|
|
p(fp.WriteString("</item>\n"))
|
|
p(fp.Close())
|
|
p(os.Chtimes(filename+".xml", t, t))
|
|
ok = doArticle(filename, item.Link, item.Title, t, needUpdate)
|
|
}()
|
|
}
|
|
}
|
|
|
|
func doArticle(filename string, url string, title string, timestamp time.Time, needUpdate bool) bool {
|
|
if exists(filename + ".skip") {
|
|
return true
|
|
}
|
|
if needUpdate {
|
|
_ = os.Remove(filename + ".err")
|
|
_ = os.Remove(filename + ".html")
|
|
_ = os.Remove(filename + ".json")
|
|
_ = os.Remove(filename + ".txt")
|
|
} else {
|
|
if exists(filename+".json") && exists(filename+".txt") {
|
|
return true
|
|
}
|
|
}
|
|
|
|
time.Sleep(2 * time.Second)
|
|
|
|
req, err := http.NewRequest("GET", url, nil)
|
|
p(err)
|
|
req.Header.Set("User-Agent", agent)
|
|
|
|
client := &http.Client{Jar: jar}
|
|
resp, err := client.Do(req)
|
|
p(err)
|
|
body, err := io.ReadAll(resp.Body)
|
|
p(err)
|
|
p(resp.Body.Close())
|
|
|
|
body = u.HtmlFix(body)
|
|
|
|
doc, err := gokogiri.ParseHtml(body)
|
|
p(err)
|
|
|
|
root := doc.Root()
|
|
|
|
ldjsons, err := root.Search(`//script[@type="application/ld+json"]`)
|
|
p(err)
|
|
found := false
|
|
var jsmain LdJson
|
|
for _, ldjson := range ldjsons {
|
|
s := ldjson.Content()
|
|
var js LdJson
|
|
p(json.Unmarshal([]byte(s), &js))
|
|
if js.Type == "NewsArticle" {
|
|
found = true
|
|
fp, err := os.Create(filename + ".json")
|
|
p(err)
|
|
p(fp.WriteString(s))
|
|
p(fp.Close())
|
|
p(os.Chtimes(filename+".json", timestamp, timestamp))
|
|
jsmain = js
|
|
break
|
|
}
|
|
}
|
|
if !found {
|
|
_ = w(fmt.Errorf("script jsonld not found: %s", url))
|
|
|
|
fp, err := os.Create(filename + ".err")
|
|
p(err)
|
|
p(fmt.Fprintf(fp, "script jsonld not found: %s\n", url))
|
|
p(fp.Close())
|
|
p(os.Chtimes(filename+".err", timestamp, timestamp))
|
|
|
|
fp, err = os.Create(filename + ".html")
|
|
p(err)
|
|
p(fp.Write(body))
|
|
p(fp.Close())
|
|
p(os.Chtimes(filename+".html", timestamp, timestamp))
|
|
return false
|
|
}
|
|
|
|
locations, err := root.Search(`//span[starts-with(@class, "location")]`)
|
|
p(err)
|
|
for _, location := range locations {
|
|
location.Remove()
|
|
}
|
|
|
|
articles, err := root.Search(`//article[starts-with(@class, "article")]`)
|
|
p(err)
|
|
if len(articles) != 1 {
|
|
msg := fmt.Sprintf("len(articles) == %d in %s", len(articles), url)
|
|
_ = w(errors.New(msg))
|
|
|
|
fp, err := os.Create(filename + ".err")
|
|
p(err)
|
|
p(fmt.Fprintln(fp, msg))
|
|
p(fp.Close())
|
|
p(os.Chtimes(filename+".err", timestamp, timestamp))
|
|
|
|
fp, err = os.Create(filename + ".html")
|
|
p(err)
|
|
p(fp.Write(body))
|
|
p(err)
|
|
p(fp.Close())
|
|
p(os.Chtimes(filename+".html", timestamp, timestamp))
|
|
|
|
return false
|
|
}
|
|
|
|
art := articles[0]
|
|
|
|
text := make([]string, 0)
|
|
if jsmain.H1 != "" {
|
|
text = append(text, jsmain.H1)
|
|
} else {
|
|
text = append(text, title)
|
|
}
|
|
if jsmain.H2 != "" {
|
|
text = append(text, jsmain.H2)
|
|
}
|
|
|
|
abb, err := art.Search(`.//div[starts-with(@class, "article-body")]`)
|
|
p(err)
|
|
|
|
found = false
|
|
for _, ab := range abb {
|
|
ell, err := ab.Search(`.//p | .//h4`)
|
|
p(err)
|
|
for _, el := range ell {
|
|
if el.Name() == "p" {
|
|
found = true
|
|
}
|
|
text = append(text, el.Content())
|
|
}
|
|
}
|
|
|
|
if !found {
|
|
_ = w(fmt.Errorf("empty body: %s", url))
|
|
|
|
fp, err := os.Create(filename + ".err")
|
|
p(err)
|
|
p(fmt.Fprintf(fp, "empty body: %s\n", url))
|
|
p(fp.Close())
|
|
p(os.Chtimes(filename+".err", timestamp, timestamp))
|
|
|
|
fp, err = os.Create(filename + ".html")
|
|
p(err)
|
|
p(fp.Write(body))
|
|
p(fp.Close())
|
|
p(os.Chtimes(filename+".html", timestamp, timestamp))
|
|
|
|
return false
|
|
|
|
}
|
|
|
|
fp, err := os.Create(filename + ".txt")
|
|
p(err)
|
|
for _, t := range text {
|
|
p(fp.WriteString(u.AddEnd(u.FixSpace(t))))
|
|
}
|
|
p(fp.Close())
|
|
return true
|
|
}
|