collect.sh: bugfix pubdate

xquery/new2old.go

@@ -0,0 +1,66 @@
+package main
+
+import (
+	e "codeberg.org/pebbe/errors"
+
+	"bufio"
+	"encoding/xml"
+	"fmt"
+	"os"
+	"regexp"
+	"strings"
+)
+
+type Item struct {
+	XMLName xml.Name `xml:"i"`
+	Msg     string   `xml:"m"`
+	Tags    []string `xml:"t"`
+	Word    string   `xml:"w"`
+}
+
+var (
+	x         = e.ExitErr
+	reTag     = regexp.MustCompile(`tag: "((?:\\.|[^\\"])*)"`)
+	reUnquote = regexp.MustCompile(`\\.`)
+)
+
+func main() {
+
+	scanner := bufio.NewScanner(os.Stdin)
+	for scanner.Scan() {
+		line := scanner.Text()
+		aa := strings.Split(line, "\t")
+		item := Item{
+			// Msg:  aa[2][:strings.LastIndex(aa[2], ".")],
+			Msg:  aa[2],
+			Word: aa[0],
+			Tags: make([]string, 0),
+		}
+		for _, tag := range parseTags(aa[1]) {
+			item.Tags = append(item.Tags, tag)
+		}
+		b, err := xml.Marshal(item)
+		x(err)
+		fmt.Println(
+			strings.ReplaceAll(
+				strings.ReplaceAll(string(b), "'", "'"),
+				""", `"`))
+	}
+	x(scanner.Err())
+
+}
+
+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 unquote(text string) string {
+	return reUnquote.ReplaceAllStringFunc(text, func(s string) string {
+		return s[1:]
+	})
+}