func SaveRule(form url.Values, opUser string) (errMsg string, err error) {
rule := model.NewCrawlRule()
err = util.ConvertAssign(rule, form)
if err != nil {
logger.Errorln("rule ConvertAssign error", err)
errMsg = err.Error()
return
}
rule.OpUser = opUser
if rule.Id != 0 {
err = rule.Persist(rule)
} else {
_, err = rule.Insert()
}
if err != nil {
errMsg = "内部服务器错误"
logger.Errorln("rule save:", errMsg, ":", err)
return
}
return
}
// 获取抓取规则列表（分页）
func FindRuleByPage(conds map[string]string, curPage, limit int) ([]*model.CrawlRule, int) {
conditions := make([]string, 0, len(conds))
for k, v := range conds {
conditions = append(conditions, k+"="+v)
}
rule := model.NewCrawlRule()
limitStr := strconv.Itoa((curPage-1)*limit) + "," + strconv.Itoa(limit)
ruleList, err := rule.Where(strings.Join(conditions, " AND ")).Order("id DESC").Limit(limitStr).
FindAll()
if err != nil {
logger.Errorln("rule service FindArticleByPage Error:", err)
return nil, 0
}
total, err := rule.Count()
if err != nil {
logger.Errorln("rule service FindArticleByPage COUNT Error:", err)
return nil, 0
}
return ruleList, total
}
// 获取url对应的文章并根据规则进行解析
func ParseArticle(articleUrl string, auto bool) (*model.Article, error) {
articleUrl = strings.TrimSpace(articleUrl)
if !strings.HasPrefix(articleUrl, "http") {
articleUrl = "http://" + articleUrl
}
tmpArticle := model.NewArticle()
err := tmpArticle.Where("url=" + articleUrl).Find("id")
if err != nil || tmpArticle.Id != 0 {
logger.Errorln(articleUrl, "has exists:", err)
return nil, errors.New("has exists!")
}
urlPaths := strings.SplitN(articleUrl, "/", 5)
domain := urlPaths[2]
for k, v := range domainPatch {
if strings.Contains(domain, k) && !strings.Contains(domain, "www."+k) {
domain = v
break
}
}
rule := model.NewCrawlRule()
err = rule.Where("domain=" + domain).Find()
if err != nil {
logger.Errorln("find rule by domain error:", err)
return nil, err
}
if rule.Id == 0 {
logger.Errorln("domain:", domain, "not exists!")
return nil, errors.New("domain not exists")
}
var doc *goquery.Document
if doc, err = goquery.NewDocument(articleUrl); err != nil {
logger.Errorln("goquery newdocument error:", err)
return nil, err
}
author, authorTxt := "", ""
if rule.InUrl {
index, err := strconv.Atoi(rule.Author)
if err != nil {
logger.Errorln("author rule is illegal:", rule.Author, "error:", err)
return nil, err
}
author = urlPaths[index]
authorTxt = author
} else {
if strings.HasPrefix(rule.Author, ".") || strings.HasPrefix(rule.Author, "#") {
authorSelection := doc.Find(rule.Author)
author, err = authorSelection.Html()
if err != nil {
logger.Errorln("goquery parse author error:", err)
return nil, err
}
author = strings.TrimSpace(author)
authorTxt = strings.TrimSpace(authorSelection.Text())
} else {
// 某些个人博客，页面中没有作者的信息，因此，规则中 author 即为 作者
author = rule.Author
authorTxt = rule.Author
}
}
title := ""
doc.Find(rule.Title).Each(func(i int, selection *goquery.Selection) {
if title != "" {
return
}
tmpTitle := strings.TrimSpace(strings.TrimPrefix(selection.Text(), "原"))
tmpTitle = strings.TrimSpace(strings.TrimPrefix(tmpTitle, "荐"))
tmpTitle = strings.TrimSpace(strings.TrimPrefix(tmpTitle, "转"))
tmpTitle = strings.TrimSpace(strings.TrimPrefix(tmpTitle, "顶"))
if tmpTitle != "" {
title = tmpTitle
}
})
if title == "" {
logger.Errorln("url:", articleUrl, "parse title error:", err)
return nil, err
}
replacer := strings.NewReplacer("[置顶]", "", "[原]", "", "[转]", "")
title = strings.TrimSpace(replacer.Replace(title))
contentSelection := doc.Find(rule.Content)
// relative url -> abs url
contentSelection.Find("img").Each(func(i int, s *goquery.Selection) {
if v, ok := s.Attr("src"); ok {
if !strings.HasPrefix(v, "http") {
s.SetAttr("src", domain+v)
}
}
})
content, err := contentSelection.Html()
if err != nil {
logger.Errorln("goquery parse content error:", err)
return nil, err
}
content = strings.TrimSpace(content)
txt := strings.TrimSpace(contentSelection.Text())
txt = articleRe.ReplaceAllLiteralString(txt, " ")
txt = articleSpaceRe.ReplaceAllLiteralString(txt, " ")
// 自动抓取，内容长度不能少于 300 字
if auto && len(txt) < 300 {
logger.Infoln(articleUrl, "content is short")
return nil, errors.New("content is short")
}
pubDate := util.TimeNow()
if rule.PubDate != "" {
pubDate = strings.TrimSpace(doc.Find(rule.PubDate).First().Text())
// sochina patch
re := regexp.MustCompile("[0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}")
submatches := re.FindStringSubmatch(pubDate)
if len(submatches) > 0 {
pubDate = submatches[0]
}
}
if pubDate == "" {
pubDate = util.TimeNow()
}
article := model.NewArticle()
article.Domain = domain
article.Name = rule.Name
article.Author = author
article.AuthorTxt = authorTxt
article.Title = title
article.Content = content
article.Txt = txt
article.PubDate = pubDate
article.Url = articleUrl
article.Lang = rule.Lang
article.Ctime = util.TimeNow()
_, err = article.Insert()
if err != nil {
logger.Errorln("insert article error:", err)
return nil, err
}
return article, nil
}
