// removeDateFromAuthor removes any dates from a string s that should only
// contain the authors of the article
func removeDateFromAuthor(s string) string {
s = regexp.MustCompile(`\d+,\s`).ReplaceAllString(s, "")
s = dateRegexp().ReplaceAllString(s, "")
//s = strings.Replace(s, ",", " ", -1)
return strtrans.Spaces(s)
}
// sanatizeDate removes any authors from a string s that should only
// contain the date of the article
func sanatizeDate(s string) string {
matches := dateRegexp().FindAllString(s, -1)
date := strings.Join(matches, " ")
if date == "" {
date = "unknown date"
}
//log.Printf("match: |%+q|\n", date)
return strtrans.Spaces(date)
}
// sanatizeTitle removes any host site affiliations after a pipe "|" from a
// string s that should only contain the title of the article
// E.g.: The civic drama of Socrates trial | Aeon Essays
// --> The civic drama of Socrates trial
func sanatizeTitle(s string) string {
s = regexp.MustCompile(`\s\|\s.+`).ReplaceAllString(s, "")
return strtrans.Spaces(s)
}