// choose http GET/method to download
func connectByHttp(p *page.Page, req *request.Request) (*http.Response, error) {
client := &http.Client{
CheckRedirect: req.GetRedirectFunc(),
}
httpreq, err := http.NewRequest(req.GetMethod(), req.GetUrl(), strings.NewReader(req.GetPostdata()))
if header := req.GetHeader(); header != nil {
httpreq.Header = req.GetHeader()
}
if cookies := req.GetCookies(); cookies != nil {
for i := range cookies {
httpreq.AddCookie(cookies[i])
}
}
var resp *http.Response
if resp, err = client.Do(httpreq); err != nil {
if e, ok := err.(*url.Error); ok && e.Err != nil && e.Err.Error() == "normal" {
// normal
} else {
mlog.LogInst().LogError(err.Error())
p.SetStatus(true, err.Error())
//fmt.Printf("client do error %v \r\n", err)
return nil, err
}
}
return resp, nil
}
func (this *HttpDownloader) downloadText(p *page.Page, req *request.Request) *page.Page {
p, destbody := this.downloadFile(p, req)
if !p.IsSucc() {
return p
}
p.SetBodyStr(destbody).SetStatus(false, "")
return p
}
func (this *HttpDownloader) downloadHtml(p *page.Page, req *request.Request) *page.Page {
var err error
p, destbody := this.downloadFile(p, req)
//fmt.Printf("Destbody %v \r\n", destbody)
if !p.IsSucc() {
//fmt.Print("Page error \r\n")
return p
}
bodyReader := bytes.NewReader([]byte(destbody))
var doc *goquery.Document
if doc, err = goquery.NewDocumentFromReader(bodyReader); err != nil {
mlog.LogInst().LogError(err.Error())
p.SetStatus(true, err.Error())
return p
}
var body string
if body, err = doc.Html(); err != nil {
mlog.LogInst().LogError(err.Error())
p.SetStatus(true, err.Error())
return p
}
p.SetBodyStr(body).SetHtmlParser(doc).SetStatus(false, "")
return p
}
// Parse html dom here and record the parse result that we want to crawl.
// Package goquery (http://godoc.org/github.com/PuerkitoBio/goquery) is used to parse html.
func (this *MyPageProcesser) Process(p *page.Page) {
if !p.IsSucc() {
println(p.Errormsg())
return
}
query := p.GetHtmlParser()
name := query.Find(".lemmaTitleH1").Text()
name = strings.Trim(name, " \t\n")
summary := query.Find(".card-summary-content .para").Text()
summary = strings.Trim(summary, " \t\n")
// the entity we want to save by Pipeline
p.AddField("name", name)
p.AddField("summary", summary)
}
func TestDownloadHtml(t *testing.T) {
//return
//request := request.NewRequest("http://live.sina.com.cn/zt/api/l/get/finance/globalnews1/index.htm?format=json&callback=t13975294&id=23521&pagesize=45&dire=f&dpc=1")
var req *request.Request
req = request.NewRequest("http://live.sina.com.cn/zt/l/v/finance/globalnews1/", "html", "", "GET", "", nil, nil, nil, nil)
var dl downloader.Downloader
dl = downloader.NewHttpDownloader()
var p *page.Page
p = dl.Download(req)
var doc *goquery.Document
doc = p.GetHtmlParser()
//fmt.Println(doc)
//body := p.GetBodyStr()
//fmt.Println(body)
var s *goquery.Selection
s = doc.Find("body")
if s.Length() < 1 {
t.Error("html parse failed!")
}
/*
doc, err := goquery.NewDocument("http://live.sina.com.cn/zt/l/v/finance/globalnews1/")
if err != nil {
fmt.Printf("%v",err)
}
s := doc.Find("meta");
fmt.Println(s.Length())
resp, err := http.Get("http://live.sina.com.cn/zt/l/v/finance/globalnews1/")
if err != nil {
fmt.Printf("%v",err)
}
defer resp.Body.Close()
doc, err = goquery.NewDocumentFromReader(resp.Body)
s = doc.Find("meta");
fmt.Println(s.Length())
*/
}
func TestDownloadJson(t *testing.T) {
//return
var req *request.Request
req = request.NewRequest("http://live.sina.com.cn/zt/api/l/get/finance/globalnews1/index.htm?format=json&id=23521&pagesize=4&dire=f&dpc=1", "json", "", "GET", "", nil, nil, nil, nil)
var dl downloader.Downloader
dl = downloader.NewHttpDownloader()
var p *page.Page
p = dl.Download(req)
var jsonMap interface{}
jsonMap = p.GetJson()
fmt.Printf("%v", jsonMap)
//fmt.Println(doc)
//body := p.GetBodyStr()
//fmt.Println(body)
}
func TestCharSetChange(t *testing.T) {
var req *request.Request
//req = request.NewRequest("http://stock.finance.sina.com.cn/usstock/api/jsonp.php/t/US_CategoryService.getList?page=1&num=60", "jsonp")
req = request.NewRequest("http://soft.chinabyte.com/416/13164916.shtml", "html", "", "GET", "", nil, nil, nil, nil)
var dl downloader.Downloader
dl = downloader.NewHttpDownloader()
var p *page.Page
p = dl.Download(req)
//hp := p.GetHtmlParser()
//fmt.Printf("%v", jsonMap)
//fmt.Println(doc)
p.GetBodyStr()
body := p.GetBodyStr()
fmt.Println(body)
}
func (this *HttpDownloader) downloadJson(p *page.Page, req *request.Request) *page.Page {
var err error
p, destbody := this.downloadFile(p, req)
if !p.IsSucc() {
return p
}
var body []byte
body = []byte(destbody)
mtype := req.GetResponceType()
if mtype == "jsonp" {
tmpstr := util.JsonpToJson(destbody)
body = []byte(tmpstr)
}
var r *simplejson.Json
if r, err = simplejson.NewJson(body); err != nil {
mlog.LogInst().LogError(string(body) + "\t" + err.Error())
p.SetStatus(true, err.Error())
return p
}
// json result
p.SetBodyStr(string(body)).SetJson(r).SetStatus(false, "")
return p
}
// Parse html dom here and record the parse result that we want to crawl.
// Package goquery (http://godoc.org/github.com/PuerkitoBio/goquery) is used to parse html.
func (this *MyPageProcesser) Process(p *page.Page) {
if !p.IsSucc() {
println(p.Errormsg())
return
}
html := p.GetBodyStr()
newUrls := urlutil.GetAllUrlIn(p.GetRequest().GetUrl(), html)
for _, newUrl := range newUrls {
newUrl = strings.Replace(newUrl, "//weibo.com/", "//tw.weibo.com/", -1)
p.AddTargetRequest(newUrl, "html")
}
mailAddrList := mailaddrutil.GetAllMailAddrIn(html)
for _, mailAddr := range mailAddrList {
if _, ok := this.mailAddrMap[mailAddr]; !ok {
this.mailAddrMap[mailAddr] = true
this.mailLogger.WriteString(mailAddr + "\n")
this.MailHandle.Push(mailAddr)
}
}
}
func (this MyPageProcesser) Process(p *page.Page) {
query := p.GetHtmlParser()
if p.GetUrlTag() == "index" {
query.Find(`div[class="main area"] div[class="lc"] ul li a`).Each(func(i int, s *goquery.Selection) {
url, isExsit := s.Attr("href")
if isExsit {
reg := regexp.MustCompile(`^do not know what is this`)
var fmtStr string
if rxYule.MatchString(url) {
reg = rxYule
fmtStr = wkSohuYule
}
if rxPic.MatchString(url) {
reg = rxPic
fmtStr = wkSohuPic
}
regxpArrag := reg.FindStringSubmatch(url)
if len(regxpArrag) == 2 {
addRequest(p, "changyan", fmt.Sprintf(fmtStr, regxpArrag[1]), "", s.Text())
}
}
})
}
if p.GetUrlTag() == "changyan" {
jsonMap := ChangyanJson{}
err := json.NewDecoder(strings.NewReader(p.GetBodyStr())).Decode(&jsonMap)
if err == nil {
content, ok := p.GetRequest().GetMeta().(string)
if ok {
fmt.Println("Title:", content, " CommentCount:", jsonMap.ListData.OuterCmtSum, " ParticipationCount:", jsonMap.ListData.ParticipationSum)
}
}
}
}
// Download file and change the charset of page charset.
func (this *HttpDownloader) downloadFile(p *page.Page, req *request.Request) (*page.Page, string) {
var err error
var urlstr string
if urlstr = req.GetUrl(); len(urlstr) == 0 {
mlog.LogInst().LogError("url is empty")
p.SetStatus(true, "url is empty")
return p, ""
}
var resp *http.Response
if proxystr := req.GetProxyHost(); len(proxystr) != 0 {
//using http proxy
//fmt.Print("HttpProxy Enter ",proxystr,"\n")
resp, err = connectByHttpProxy(p, req)
} else {
//normal http download
//fmt.Print("Http Normal Enter \n",proxystr,"\n")
resp, err = connectByHttp(p, req)
}
if err != nil {
return p, ""
}
//b, _ := ioutil.ReadAll(resp.Body)
//fmt.Printf("Resp body %v \r\n", string(b))
p.SetHeader(resp.Header)
p.SetCookies(resp.Cookies())
// get converter to utf-8
var bodyStr string
if resp.Header.Get("Content-Encoding") == "gzip" {
bodyStr = this.changeCharsetEncodingAutoGzipSupport(resp.Header.Get("Content-Type"), resp.Body)
} else {
bodyStr = this.changeCharsetEncodingAuto(resp.Header.Get("Content-Type"), resp.Body)
}
//fmt.Printf("utf-8 body %v \r\n", bodyStr)
defer resp.Body.Close()
return p, bodyStr
}
func addRequest(p *page.Page, tag, url, cookie, content string) {
req := request.NewRequest(url, "json", tag, "GET", "", nil, nil, nil, content)
p.AddTargetRequestWithParams(req)
}
// Parse html dom here and record the parse result that we want to Page.
// Package goquery (http://godoc.org/github.com/PuerkitoBio/goquery) is used to parse html.
func (this *MyPageProcesser) Process(p *page.Page) {
if !p.IsSucc() {
println(p.Errormsg())
return
}
query := p.GetHtmlParser()
query.Find(`div[class="wx-rb bg-blue wx-rb_v1 _item"]`).Each(func(i int, s *goquery.Selection) {
name := s.Find("div.txt-box > h3").Text()
href, _ := s.Attr("href")
fmt.Printf("WeName:%v link:http://http://weixin.sogou.com%v \r\n", name, href)
// the entity we want to save by Pipeline
p.AddField("name", name)
p.AddField("href", href)
})
next_page_href, _ := query.Find("#sogou_next").Attr("href")
if next_page_href == "" {
p.SetSkip(true)
} else {
p.AddTargetRequestWithHeaderFile("http://weixin.sogou.com/weixin"+next_page_href, "html", "weixin.sogou.com.json")
}
}
// Parse html dom here and record the parse result that we want to Page.
// Package goquery (http://godoc.org/github.com/PuerkitoBio/goquery) is used to parse html.
func (this *MyPageProcesser) Process(p *page.Page) {
if p.GetUrlTag() == "site_login" {
//fmt.Printf("%v\n", p.GetCookies())
this.cookies = p.GetCookies()
// AddTargetRequestWithParams Params:
// 1. Url.
// 2. Responce type is "html" or "json" or "jsonp" or "text".
// 3. The urltag is name for marking url and distinguish different urls in PageProcesser and Pipeline.
// 4. The method is POST or GET.
// 5. The postdata is body string sent to sever.
// 6. The header is header for http request.
// 7. Cookies
// 8. Http redirect function
if len(this.cookies) != 0 {
p.AddField("info", "get cookies success")
req := request.NewRequest("http://backadmin.hucong.net/site/index", "html", "site_index", "GET", "", nil, this.cookies, nil, nil)
p.AddTargetRequestWithParams(req)
} else {
p.AddField("info", "get cookies failed")
}
} else {
//fmt.Printf("%v\n", p.GetBodyStr())
query := p.GetHtmlParser()
pageTitle := query.Find(".page-content .page-title").Text()
if len(pageTitle) != 0 {
p.AddField("page_title", pageTitle)
p.AddField("info", "login success")
} else {
p.AddField("info", "login failed")
}
}
return
if !p.IsSucc() {
println(p.Errormsg())
return
}
query := p.GetHtmlParser()
var urls []string
query.Find("h3[class='repo-list-name'] a").Each(func(i int, s *goquery.Selection) {
href, _ := s.Attr("href")
urls = append(urls, "http://github.com/"+href)
})
// these urls will be saved and crawed by other coroutines.
p.AddTargetRequests(urls, "html")
name := query.Find(".entry-title .author").Text()
name = strings.Trim(name, " \t\n")
repository := query.Find(".entry-title .js-current-repository").Text()
repository = strings.Trim(repository, " \t\n")
//readme, _ := query.Find("#readme").Html()
if name == "" {
p.SetSkip(true)
}
// the entity we want to save by Pipeline
p.AddField("author", name)
p.AddField("project", repository)
//p.AddField("readme", readme)
}
func (this *MyProcessor) Process(p *page.Page) {
if !p.IsSucc() {
mlog.LogInst().LogError(p.Errormsg())
return
}
u, err := url.Parse(p.GetRequest().GetUrl())
if err != nil {
mlog.LogInst().LogError(err.Error())
return
}
if !strings.HasSuffix(u.Host, "jiexieyin.org") {
return
}
var urls []string
query := p.GetHtmlParser()
query.Find("a").Each(func(i int, s *goquery.Selection) {
href, _ := s.Attr("href")
reJavascript := regexp.MustCompile("^javascript\\:")
reLocal := regexp.MustCompile("^\\#")
reMailto := regexp.MustCompile("^mailto\\:")
if reJavascript.MatchString(href) || reLocal.MatchString(href) || reMailto.MatchString(href) {
return
}
//处理相对路径
var absHref string
urlHref, err := url.Parse(href)
if err != nil {
mlog.LogInst().LogError(err.Error())
return
}
if !urlHref.IsAbs() {
urlPrefix := p.GetRequest().GetUrl()
absHref = urlPrefix + href
urls = append(urls, absHref)
} else {
urls = append(urls, href)
}
})
p.AddTargetRequests(urls, "html")
}
// Parse html dom here and record the parse result that we want to Page.
// Package goquery (http://godoc.org/github.com/PuerkitoBio/goquery) is used to parse html.
func (this *MyPageProcesser) Process(p *page.Page) {
if !p.IsSucc() {
println(p.Errormsg())
return
}
query := p.GetHtmlParser()
var urls []string
query.Find("h3[class='repo-list-name'] a").Each(func(i int, s *goquery.Selection) {
href, _ := s.Attr("href")
urls = append(urls, "http://github.com/"+href)
})
// these urls will be saved and crawed by other coroutines.
p.AddTargetRequests(urls, "html")
name := query.Find(".entry-title .author").Text()
name = strings.Trim(name, " \t\n")
repository := query.Find(".entry-title .js-current-repository").Text()
repository = strings.Trim(repository, " \t\n")
//readme, _ := query.Find("#readme").Html()
if name == "" {
p.SetSkip(true)
}
// the entity we want to save by Pipeline
p.AddField("author", name)
p.AddField("project", repository)
//p.AddField("readme", readme)
}
// Parse html dom here and record the parse result that we want to crawl.
// Package simplejson (https://github.com/bitly/go-simplejson) is used to parse data of json.
func (this *MyPageProcesser) Process(p *page.Page) {
if !p.IsSucc() {
println(p.Errormsg())
return
}
query := p.GetJson()
status, err := query.GetPath("result", "status", "code").Int()
if status != 0 || err != nil {
log.Panicf("page is crawled error : errorinfo=%s : status=%d : startNewsId=%d", err.Error(), status, this.startNewsId)
}
num, err := query.GetPath("result", "pageStr", "pageSize").Int()
if num == 0 || err != nil {
// Add url of next crawl
startIdstr := strconv.Itoa(this.startNewsId)
p.AddTargetRequest("http://live.sina.com.cn/zt/api/l/get/finance/globalnews1/index.htm?format=json&id="+startIdstr+"&pagesize=10&dire=f", "json")
return
}
var idint, nextid int
var nextidstr string
query = query.Get("result").Get("data")
for i := 0; i < num; i++ {
id, err := query.GetIndex(i).Get("id").String()
if id == "" || err != nil {
continue
}
idint, err = strconv.Atoi(id)
if err != nil {
continue
}
if idint <= this.startNewsId {
break
}
if i == 0 {
nextid = idint
nextidstr = id
}
content, err := query.GetIndex(i).Get("content").String()
if content == "" || err != nil {
continue
}
time, err := query.GetIndex(i).Get("created_at").String()
if err != nil {
continue
}
p.AddField(id+"_id", id)
p.AddField(id+"_content", content)
p.AddField(id+"_time", time)
}
// Add url of next crawl
this.startNewsId = nextid
p.AddTargetRequest("http://live.sina.com.cn/zt/api/l/get/finance/globalnews1/index.htm?format=json&id="+nextidstr+"&pagesize=10&dire=f", "json")
//println(p.GetTargetRequests())
}