Get_Cosme_comments.R

library(XML)

#获取当前链接的所有评论
get_one_page_comment<-function(url){
  
  html<-htmlParse(url)
  resnodes<-getNodeSet(html,"//div[@class='review-sec']//div[@class='inner']//div[@class='body']//p[@class='read']//span[@class='read-more']//a[@class='cmn-viewmore']")
  #获取当前页面中所有comment的长文链接  
  comments_url<-NULL
  if(is.null(resnodes)){
    return(NULL) }else{ 
      for(i in 1:length(resnodes)){
        comments_url[i]<-xmlGetAttr(resnodes[[i]],name="href")
      } }
  
  #获取commets_url中的单一页面评论
  comments_node<-NULL
  comments<-NULL
  for (i in 1:length(resnodes)){
    comments_node[i]<-getNodeSet(htmlParse(comments_url[i]),"//p[@class='read']")
    comments[i]<-xmlValue(comments_node[[i]]) }
  return(comments)
}

# 获取产品页面总评论页数
get_comments_page_num<-function(url){
  html<-htmlParse(url)
  page_num_end<-getNodeSet(html,"//div[@class='cmn-paging clearfix']//ul[@class='clearfix']//li[last()-1]//a")[[1]]
  end_num<-as.numeric(xmlValue(page_num_end))
  return(end_num)
}



#  http://www.cosme.net/product/product_id/10010831/reviews
comment_all<-function(url){
  end_num<-get_comments_page_num(url)
  urlpart<-strsplit(url,"/")
  product_id<-urlpart[[1]][6]
  urlpre<-paste0("http://www.cosme.net/product/product_id/",product_id,"/reviews/p/")
  
  comments_all=NULL
  as.list(comments_all) 
  for(i in 1:end_num){
    print(paste0("正在获取第",i,"页所有评论")) 
    a<-paste0(urlpre,i) 
    b<-get_one_page_comment(a) 
    gsub(pattern = "\n", replacement = "",b) # 正则表达去除 \n
    comments_all[[i]]<-b
  }
  return(comments_all)

}



#####################################################################################################

# 注意网址结构                             /ProductID/reviews/
url<-"http://www.cosme.net/product/product_id/10010831/reviews/"
                                                                           # 表示第一页
#    http://www.cosme.net/product/product_id/10010831/reviews/p/1
# 如果网址像这样的话http://www.cosme.net/product/product_id/10010831/reviews/p/1 
            # 要改成http://www.cosme.net/product/product_id/10010831/reviews/p/ 


(comments<-comment_all(url))

# comments为表单结构
#comments[[a]][b] 表示第a页 第b条评论
#在工作区间可以找到