first commit

NU/cmd/nu/nu.go

@@ -0,0 +1,240 @@
+package main
+
+import (
+	"github.com/pebbe/util"
+
+	"bytes"
+	"encoding/json"
+	"encoding/xml"
+	"fmt"
+	"html"
+	"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"`
+	Link     string `xml:"link"`
+	Data     []byte `xml:",innerxml"`
+}
+
+type Doc struct {
+	Graph []GItem `json:"@graph"`
+}
+
+type GItem struct {
+	ArticleBody    string   `json:"articleBody"`
+	ArticleSection []string `json:"articleSection"`
+}
+
+var (
+	x     = util.CheckErr
+	w     = util.WarnErr
+	agent = "AhrefsBot/7.0"
+)
+
+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>")
+	return s[i1:i2]
+}
+
+func main() {
+	resp, err := http.Get("https://www.nu.nl/rss")
+	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/NU/%d/%02d", year, week)
+		filename := dirname + "/" + url.PathEscape(item.Guid)
+
+		ts := fmt.Sprintf("%d", t.Unix())
+		needUpdate := fileDate(filename+".xml") != ts
+
+		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))
+		if !doArticle(filename, item.Link, t, needUpdate) {
+			x(os.Remove(filename + ".xml"))
+		}
+	}
+}
+
+func doArticle(filename string, url string, timestamp time.Time, needUpdate bool) bool {
+	if needUpdate {
+		_ = os.Remove(filename + ".err")
+		_ = os.Remove(filename + ".html")
+		_ = os.Remove(filename + ".json")
+		_ = os.Remove(filename + ".txt")
+		_ = os.Remove(filename + ".skip")
+	} else {
+		// voor sommige berichten is geen .txt, alleen .json
+		if exists(filename+".json") || exists(filename+".skip") {
+			return true
+		}
+	}
+
+	time.Sleep(2 * time.Second)
+
+	req, err := http.NewRequest("GET", url, 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())
+
+	s := string(body)
+	ok := true
+	i1 := strings.Index(s, `<script type="application/ld+json"`)
+	if i1 < 0 {
+		ok = false
+	} else {
+		i1 += strings.Index(s[i1:], `>`) + 1
+		i2 := i1 + strings.Index(s[i1:], `</script>`)
+		if i2 < i1 {
+			ok = false
+		} else {
+			s = s[i1:i2]
+		}
+	}
+	if !ok {
+		_ = w(fmt.Errorf("script jsonld not found: %s", url))
+
+		fp, err := os.Create(filename + ".err")
+		x(err)
+		_, err = fmt.Fprintf(fp, "script jsonld not found: %s\n", url)
+		x(err)
+		x(fp.Close())
+		x(os.Chtimes(filename+".err", timestamp, timestamp))
+
+		fp, err = os.Create(filename + ".html")
+		x(err)
+		_, err = fp.Write(body)
+		x(err)
+		x(fp.Close())
+		x(os.Chtimes(filename+".html", timestamp, timestamp))
+
+		return false
+	}
+	fp, err := os.Create(filename + ".json")
+	x(err)
+	_, err = fp.WriteString(s)
+	x(err)
+	x(fp.Close())
+	x(os.Chtimes(filename+".json", timestamp, timestamp))
+
+	var doc Doc
+	if err = json.Unmarshal([]byte(s), &doc); err != nil {
+		_ = w(err, url)
+
+		fp, err := os.Create(filename + ".err")
+		x(err)
+		_, err = fmt.Fprintf(fp, "%s: %v\n", url, err)
+		x(err)
+		x(fp.Close())
+		x(os.Chtimes(filename+".err", timestamp, timestamp))
+
+		return false
+	}
+
+	cats := make([]string, 0)
+	var buffer bytes.Buffer
+	for _, i := range doc.Graph {
+		_, err = buffer.WriteString(html.UnescapeString(i.ArticleBody))
+		x(err)
+		cats = append(cats, i.ArticleSection...)
+	}
+	text := buffer.String()
+
+	// sommige berichten bevatten geen tekst, maar een video bijvoorbeeld
+	// dit is geen fout
+	if len(text) > 0 {
+
+		// text bevat kopjes zonder punt aan het eind
+		lines := strings.Split(text, "\n")
+		for i, line := range lines {
+			n := len(line)
+			if n > 0 {
+				if strings.ContainsAny(line[n-1:], ".!?") {
+					continue
+				}
+			}
+			if n > 1 {
+				s := line[n-2:]
+				if s == `."` || s == `!"` || s == `?"` {
+					continue
+				}
+			}
+			lines[i] = line + "."
+		}
+		text = strings.Join(lines, "\n") + "\n"
+
+		fp, err := os.Create(filename + ".txt")
+		x(err)
+		if len(cats) == 0 {
+			_, err := fmt.Fprintln(fp, "##META text cat =")
+			x(err)
+		} else {
+			for _, cat := range cats {
+				_, err := fmt.Fprintf(fp, "##META text cat = %s\n", cat)
+				x(err)
+			}
+		}
+		_, err = fp.WriteString(text)
+		x(err)
+		x(fp.Close())
+		x(os.Chtimes(filename+".txt", timestamp, timestamp))
+	}
+
+	return true
+}