func FetchGo(runtimeConfig RuntimeConfig, taskC *chan []byte, quitC *chan bool, offsets *RoutingOffset) {
shard := offsets.Shard
log.Info("fetch task started.shard:", shard)
go func() {
for {
url := <-*taskC
log.Debug("shard:", shard, ",url received:", string(url))
if !runtimeConfig.Storage.CheckFetchedUrl(url) {
timeout := 10 * time.Second
// if(fetchFilter.Lookup(url)){
// log.Debug("hit fetch filter ,ignore,",string(url))
// continue
// }
// fetchFilter.Add(url)
log.Debug("shard:", shard, ",url cool,start fetching:", string(url))
fetchUrl(url, timeout, runtimeConfig, offsets)
if runtimeConfig.TaskConfig.FetchDelayThreshold > 0 {
log.Debug("sleep ", runtimeConfig.TaskConfig.FetchDelayThreshold, "ms to control crawling speed")
time.Sleep(time.Duration(runtimeConfig.TaskConfig.FetchDelayThreshold) * time.Millisecond)
log.Debug("wake up now,continue crawing")
}
} else {
log.Debug("hit fetch-bloomfilter,ignore,", string(url))
}
}
}()
<-*quitC
log.Info("fetch task exit.shard:", shard)
}
func ParsedSavedFileLog2(runtimeConfig RuntimeConfig, pendingFetchUrls chan []byte, url string) {
if url != "" {
log.Trace("start parse filelog:", url)
if runtimeConfig.Storage.CheckFetchedUrl([]byte(url)) {
log.Debug("hit fetch filter ignore,", url)
return
}
log.Debug("new task extracted from saved page:", url)
pendingFetchUrls <- []byte(url)
}
}
func FetchFileWithOffset(runtimeConfig RuntimeConfig, path string, skipOffset int64) {
var offset int64 = 0
time1, _ := util.FileMTime(path)
log.Debug("start touch time:", time1)
f, err := os.Open(path)
if err != nil {
log.Debug("error opening file,", path, " ", err)
return
}
r := bufio.NewReader(f)
s, e := util.Readln(r)
offset = 0
log.Trace("new offset:", offset)
for e == nil {
offset = offset + 1
//TODO use byte offset instead of lines
if offset > skipOffset {
ParsedSavedFileLog(runtimeConfig, s)
}
runtimeConfig.Storage.PersistOffset(runtimeConfig.PathConfig.SavedFileLog+".offset", offset)
s, e = util.Readln(r)
//todo store offset
}
log.Trace("end offset:", offset, "vs ", skipOffset)
waitUpdate:
time2, _ := util.FileMTime(path)
log.Trace("2nd touch time:", time2)
if time2 > time1 {
log.Trace("file has been changed,restart parse")
FetchFileWithOffset(runtimeConfig, path, offset)
} else {
log.Trace("waiting file update", path)
time.Sleep(10 * time.Millisecond)
goto waitUpdate
}
}
//format url,prepare for bloom filter
func formatUrlForFilter(url []byte) []byte {
src := string(url)
log.Debug("start to normalize url:", src)
if strings.HasSuffix(src, "/") {
src = strings.TrimRight(src, "/")
}
src = strings.TrimSpace(src)
src = strings.ToLower(src)
return []byte(src)
}
func (this *FsStore) LoadOffset(fileName string) int64 {
log.Debug("start init offsets,", fileName)
if util.CheckFileExists(fileName) {
log.Debug("found offset file,start loading,", fileName)
n, err := ioutil.ReadFile(fileName)
if err != nil {
log.Error("offset", fileName, ",", err)
return 0
}
ret, err := strconv.ParseInt(string(n), 10, 64)
if err != nil {
log.Error("offset", fileName, ",", err)
return 0
}
log.Info("init offsets successfully,", fileName, ":", ret)
return int64(ret)
}
return 0
}
//parse config setting
func parseConfig() *TaskConfig {
log.Debug("start parsing taskConfig")
taskConfig := new(TaskConfig)
taskConfig.LinkUrlExtractRegex = regexp.MustCompile(
config.GetStringConfig("CrawlerRule", "LinkUrlExtractRegex", "(\\s+(src2|src|href|HREF|SRC))\\s*=\\s*[\"']?(.*?)[\"']"))
taskConfig.SplitByUrlParameter = config.GetStringConfig("CrawlerRule", "SplitByUrlParameter", "p")
taskConfig.LinkUrlExtractRegexGroupIndex = config.GetIntConfig("CrawlerRule", "LinkUrlExtractRegexGroupIndex", 3)
taskConfig.Name = config.GetStringConfig("CrawlerRule", "Name", "GopaTask")
taskConfig.FollowSameDomain = config.GetBoolConfig("CrawlerRule", "FollowSameDomain", true)
taskConfig.FollowSubDomain = config.GetBoolConfig("CrawlerRule", "FollowSubDomain", true)
taskConfig.LinkUrlMustContain = config.GetStringConfig("CrawlerRule", "LinkUrlMustContain", "")
taskConfig.LinkUrlMustNotContain = config.GetStringConfig("CrawlerRule", "LinkUrlMustNotContain", "")
taskConfig.SkipPageParsePattern = regexp.MustCompile(config.GetStringConfig("CrawlerRule", "SkipPageParsePattern", ".*?\\.((js)|(css)|(rar)|(gz)|(zip)|(exe)|(bmp)|(jpeg)|(gif)|(png)|(jpg)|(apk))\\b")) //end with js,css,apk,zip,ignore
taskConfig.FetchUrlPattern = regexp.MustCompile(config.GetStringConfig("CrawlerRule", "FetchUrlPattern", ".*"))
taskConfig.FetchUrlMustContain = config.GetStringConfig("CrawlerRule", "FetchUrlMustContain", "")
taskConfig.FetchUrlMustNotContain = config.GetStringConfig("CrawlerRule", "FetchUrlMustNotContain", "")
taskConfig.SavingUrlPattern = regexp.MustCompile(config.GetStringConfig("CrawlerRule", "SavingUrlPattern", ".*"))
taskConfig.SavingUrlMustContain = config.GetStringConfig("CrawlerRule", "SavingUrlMustContain", "")
taskConfig.SavingUrlMustNotContain = config.GetStringConfig("CrawlerRule", "SavingUrlMustNotContain", "")
taskConfig.Cookie = config.GetStringConfig("CrawlerRule", "Cookie", "")
taskConfig.FetchDelayThreshold = config.GetIntConfig("CrawlerRule", "FetchDelayThreshold", 0)
taskConfig.TaskDataPath = config.GetStringConfig("CrawlerRule", "TaskData", runtimeConfig.PathConfig.TaskData+"/"+taskConfig.Name+"/")
defaultWebDataPath := runtimeConfig.PathConfig.WebData + "/" + taskConfig.Name + "/"
if runtimeConfig.StoreWebPageTogether {
defaultWebDataPath = runtimeConfig.PathConfig.WebData
}
taskConfig.WebDataPath = config.GetStringConfig("CrawlerRule", "WebData", defaultWebDataPath)
log.Debug("finished parsing taskConfig")
return taskConfig
}
func initBloomFilter(bloomFilterPersistFileName string) *Filter {
var bloomFilter = new(Filter)
//loading or initializing bloom filter
if util.CheckFileExists(bloomFilterPersistFileName) {
log.Debug("found bloomFilter,start reload,", bloomFilterPersistFileName)
n, err := ioutil.ReadFile(bloomFilterPersistFileName)
if err != nil {
log.Error("bloomFilter:", bloomFilterPersistFileName, err)
}
if err := bloomFilter.GobDecode(n); err != nil {
log.Error("bloomFilter:", bloomFilterPersistFileName, err)
}
log.Info("bloomFilter successfully reloaded:", bloomFilterPersistFileName)
} else {
probItems := config.GetIntConfig("BloomFilter", "ItemSize", 100000)
log.Debug("initializing bloom-filter", bloomFilterPersistFileName, ",virual size is,", probItems)
bloomFilter = NewFilter(fnv.New64(), probItems)
log.Info("bloomFilter successfully initialized:", bloomFilterPersistFileName)
}
return bloomFilter
}
func ParsedSavedFileLog(runtimeConfig RuntimeConfig, fileLog string) {
if fileLog != "" {
log.Debug("start parse filelog:", fileLog)
//load file's content,and extract links
stringArray := strings.Split(fileLog, "|||")
fileUrl := stringArray[0]
fileName := []byte(stringArray[1])
if runtimeConfig.Storage.CheckParsedFile(fileName) {
log.Debug("hit parse filter ignore,", string(fileName))
return
}
fileContent := loadFileContent(string(fileName))
runtimeConfig.Storage.AddParsedFile(fileName)
if fileContent != nil {
// log.Debug("partition:", partition, ",parse fileName:", string(fileName))
//extract urls to fetch queue.
extractLinks(runtimeConfig, fileUrl, fileName, fileContent)
// offsetV := msg.Offset()
// offset.Offset = offsetV
//
// path := taskConfig.BaseStoragePath+ "task/parse_offset_" + strconv.FormatInt(int64(partition), 10) + ".tmp"
// path_new := taskConfig.BaseStoragePath+ "task/parse_offset_" + strconv.FormatInt(int64(partition), 10)
// fout, error := os.Create(path)
// if error != nil {
// log.Error(path, error)
// return
// }
//
// defer fout.Close()
// log.Debug("partition:", partition, ",saved offset:", offsetV)
// fout.Write([]byte(strconv.FormatUint(msg.Offset(), 10)))
// utils.CopyFile(path, path_new)
}
}
}
func (this *FsStore) PersistOffset(fileName string, offset int64) {
//persist worker's offset
path := fileName + ".tmp"
fout, error := os.Create(path)
if error != nil {
log.Error(path, error)
return
}
defer fout.Close()
log.Debug("saved offset:", fileName, ":", offset)
fout.Write([]byte(strconv.FormatInt(offset, 10)))
util.CopyFile(path, fileName)
}
func shutdown(offsets []*RoutingOffset, quitChannels []*chan bool, offsets2 []*RoutingOffset, quitChannels2 []*chan bool, quit chan bool) {
log.Debug("start shutting down")
for i := range quitChannels {
log.Debug("send exit signal to channel,", i)
*quitChannels[i] <- true
}
for i := range quitChannels2 {
log.Debug("send exit signal to channel,", i)
*quitChannels2[i] <- true
}
log.Info("sent quit signal to go routings done")
// for i:=range offsets{
// //TODO
// log.Info("persist offset,",i,":",offsets[i].Offset,",",offsets[i].shard)
// }
// log.Info("persist kafka offsets done")
quit <- true
log.Debug("finished shutting down")
}
func post(url string, cookie string, postStr string) []byte {
log.Debug("let's post :" + url)
client := &http.Client{
CheckRedirect: nil,
}
postBytesReader := bytes.NewReader([]byte(postStr))
reqest, _ := http.NewRequest("POST", url, postBytesReader)
reqest.Header.Set("User-Agent", " Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/31.0.1650.63 Safari/537.36")
reqest.Header.Set("Accept", "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8")
reqest.Header.Set("Accept-Charset", "GBK,utf-8;q=0.7,*;q=0.3")
reqest.Header.Set("Accept-Encoding", "gzip,deflate,sdch")
// reqest.Header.Add("Content-Type", "application/x-www-form-urlencoded")
reqest.Header.Add("Content-Type", "application/x-www-form-urlencoded; charset=UTF-8")
reqest.Header.Set("Accept-Language", "zh-CN,zh;q=0.8")
reqest.Header.Set("Cache-Control", "max-age=0")
reqest.Header.Set("Connection", "keep-alive")
reqest.Header.Set("Referer", url)
if len(cookie) > 0 {
log.Debug("dealing with cookie:" + cookie)
array := strings.Split(cookie, ";")
for item := range array {
array2 := strings.Split(array[item], "=")
if len(array2) == 2 {
cookieObj := http.Cookie{}
cookieObj.Name = array2[0]
cookieObj.Value = array2[1]
reqest.AddCookie(&cookieObj)
} else {
log.Info("error,index out of range:" + array[item])
}
}
}
resp, err := client.Do(reqest)
if err != nil {
log.Error(url, err)
return nil
}
defer resp.Body.Close()
var reader io.ReadCloser
switch resp.Header.Get("Content-Encoding") {
case "gzip":
reader, err = gzip.NewReader(resp.Body)
if err != nil {
log.Error(url, err)
return nil
}
defer reader.Close()
default:
reader = resp.Body
}
if reader != nil {
body, err := ioutil.ReadAll(reader)
if err != nil {
log.Error(url, err)
return nil
}
return body
}
return nil
}
func main() {
flag.StringVar(&seedUrl, "seed", "http://example.com", "the seed url,where everything starts")
flag.StringVar(&logLevel, "log", "info", "setting log level,options:trace,debug,info,warn,error")
flag.Parse()
defer log.Flush()
runtime.GOMAXPROCS(runtime.NumCPU())
log.SetInitLogging(logLevel)
runtimeConfig.PathConfig = new(PathConfig)
runtimeConfig.ClusterConfig = new(ClusterConfig)
runtimeConfig.ClusterConfig.Name = config.GetStringConfig("cluster", "name", "gopa")
// per cluster:data/gopa/
runtimeConfig.PathConfig.Home = config.GetStringConfig("path", "home", "cluster/"+runtimeConfig.ClusterConfig.Name+"/")
runtimeConfig.PathConfig.Data = config.GetStringConfig("path", "data", "")
if runtimeConfig.PathConfig.Data == "" {
runtimeConfig.PathConfig.Data = runtimeConfig.PathConfig.Home + "/" + "data/"
}
runtimeConfig.PathConfig.Log = config.GetStringConfig("path", "log", "")
if runtimeConfig.PathConfig.Log == "" {
runtimeConfig.PathConfig.Log = runtimeConfig.PathConfig.Home + "/" + "log/"
}
runtimeConfig.PathConfig.WebData = config.GetStringConfig("path", "webdata", "")
if runtimeConfig.PathConfig.WebData == "" {
runtimeConfig.PathConfig.WebData = runtimeConfig.PathConfig.Data + "/" + "webdata/"
}
runtimeConfig.PathConfig.TaskData = config.GetStringConfig("path", "taskdata", "")
if runtimeConfig.PathConfig.TaskData == "" {
runtimeConfig.PathConfig.TaskData = runtimeConfig.PathConfig.Data + "/" + "taskdata/"
}
runtimeConfig.StoreWebPageTogether = config.GetBoolConfig("Global", "StoreWebPageTogether", true)
runtimeConfig.TaskConfig = parseConfig()
//set default logging
logPath := runtimeConfig.PathConfig.Log + "/" + runtimeConfig.TaskConfig.Name + "/gopa.log"
log.SetLogging(logLevel, logPath)
runtimeConfig.ParseUrlsFromSavedFileLog = config.GetBoolConfig("Switch", "ParseUrlsFromSavedFileLog", true)
runtimeConfig.LoadTemplatedFetchJob = config.GetBoolConfig("Switch", "LoadTemplatedFetchJob", true)
runtimeConfig.LoadRuledFetchJob = config.GetBoolConfig("Switch", "LoadRuledFetchJob", false)
runtimeConfig.LoadPendingFetchJobs = config.GetBoolConfig("Switch", "LoadPendingFetchJobs", true)
runtimeConfig.HttpEnabled = config.GetBoolConfig("Switch", "HttpEnabled", true)
runtimeConfig.ParseUrlsFromPreviousSavedPage = config.GetBoolConfig("Switch", "ParseUrlsFromPreviousSavedPage", false)
runtimeConfig.ArrayStringSplitter = config.GetStringConfig("CrawlerRule", "ArrayStringSplitter", ",")
runtimeConfig.GoProfEnabled = config.GetBoolConfig("CrawlerRule", "GoProfEnabled", false)
runtimeConfig.WalkBloomFilterFileName = config.GetStringConfig("BloomFilter", "WalkBloomFilterFileName", runtimeConfig.TaskConfig.TaskDataPath+"/filters/walk.bloomfilter")
runtimeConfig.FetchBloomFilterFileName = config.GetStringConfig("BloomFilter", "FetchBloomFilterFileName", runtimeConfig.TaskConfig.TaskDataPath+"/filters/fetch.bloomfilter")
runtimeConfig.ParseBloomFilterFileName = config.GetStringConfig("BloomFilter", "ParseBloomFilterFileName", runtimeConfig.TaskConfig.TaskDataPath+"/filters/parse.bloomfilter")
runtimeConfig.PendingFetchBloomFilterFileName = config.GetStringConfig("BloomFilter", "PendingFetchBloomFilterFileName", runtimeConfig.TaskConfig.TaskDataPath+"/filters/pending_fetch.bloomfilter")
runtimeConfig.PathConfig.SavedFileLog = runtimeConfig.TaskConfig.TaskDataPath + "/tasks/pending_parse.files"
runtimeConfig.PathConfig.PendingFetchLog = runtimeConfig.TaskConfig.TaskDataPath + "/tasks/pending_fetch.urls"
runtimeConfig.PathConfig.FetchFailedLog = runtimeConfig.TaskConfig.TaskDataPath + "/tasks/failed_fetch.urls"
runtimeConfig.MaxGoRoutine = config.GetIntConfig("Global", "MaxGoRoutine", 2)
if runtimeConfig.MaxGoRoutine < 2 {
runtimeConfig.MaxGoRoutine = 2
}
log.Debug("maxGoRoutine:", runtimeConfig.MaxGoRoutine)
log.Debug("path.home:", runtimeConfig.PathConfig.Home)
os.MkdirAll(runtimeConfig.PathConfig.Home, 0777)
os.MkdirAll(runtimeConfig.PathConfig.Data, 0777)
os.MkdirAll(runtimeConfig.PathConfig.Log, 0777)
os.MkdirAll(runtimeConfig.PathConfig.WebData, 0777)
os.MkdirAll(runtimeConfig.PathConfig.TaskData, 0777)
os.MkdirAll(runtimeConfig.TaskConfig.TaskDataPath, 0777)
os.MkdirAll(runtimeConfig.TaskConfig.TaskDataPath+"/tasks/", 0777)
os.MkdirAll(runtimeConfig.TaskConfig.TaskDataPath+"/filters/", 0777)
os.MkdirAll(runtimeConfig.TaskConfig.TaskDataPath+"/urls/", 0777)
os.MkdirAll(runtimeConfig.TaskConfig.WebDataPath, 0777)
runtimeConfig.RuledFetchConfig = new(RuledFetchConfig)
runtimeConfig.RuledFetchConfig.UrlTemplate = config.GetStringConfig("RuledFetch", "UrlTemplate", "")
runtimeConfig.RuledFetchConfig.From = config.GetIntConfig("RuledFetch", "From", 0)
runtimeConfig.RuledFetchConfig.To = config.GetIntConfig("RuledFetch", "To", 10)
runtimeConfig.RuledFetchConfig.Step = config.GetIntConfig("RuledFetch", "Step", 1)
runtimeConfig.RuledFetchConfig.LinkExtractPattern = config.GetStringConfig("RuledFetch", "LinkExtractPattern", "")
runtimeConfig.RuledFetchConfig.LinkTemplate = config.GetStringConfig("RuledFetch", "LinkTemplate", "")
if seedUrl == "" || seedUrl == "http://example.com" {
log.Error("no seed was given. type:\"gopa -h\" for help.")
os.Exit(1)
}
log.Info("[gopa] " + runtimeConfig.Version + " is on.")
runtimeConfig.Storage = &fsstore.FsStore{}
// if(runtimeConfig.)
runtimeConfig.Storage.InitWalkBloomFilter(runtimeConfig.WalkBloomFilterFileName)
runtimeConfig.Storage.InitFetchBloomFilter(runtimeConfig.FetchBloomFilterFileName)
runtimeConfig.Storage.InitParseBloomFilter(runtimeConfig.ParseBloomFilterFileName)
runtimeConfig.Storage.InitPendingFetchBloomFilter(runtimeConfig.PendingFetchBloomFilterFileName)
// atr:="AZaz"
// btr:=[]byte(atr)
// fmt.Println(btr)
//
// id:= getSeqStr([]byte("AA"),[]byte("ZZ"),false)
// fmt.Println(id)
//pprof serves
if runtimeConfig.GoProfEnabled {
go func() {
log.Info(http.ListenAndServe("localhost:6060", nil))
log.Info("pprof server is up,http://localhost:6060/debug/pprof")
}()
}
//http serves
if runtimeConfig.HttpEnabled {
go func() {
httpServ.Start(runtimeConfig)
}()
}
//adding default http protocol
if !strings.HasPrefix(seedUrl, "http") {
seedUrl = "http://" + seedUrl
}
maxGoRoutine := runtimeConfig.MaxGoRoutine
fetchQuitChannels := make([]*chan bool, maxGoRoutine) //shutdownSignal signals for each go routing
fetchTaskChannels := make([]*chan []byte, maxGoRoutine) //fetchTask channels
fetchOffsets := make([]*RoutingOffset, maxGoRoutine) //kafka fetchOffsets
parseQuitChannels := make([]*chan bool, 2) //shutdownSignal signals for each go routing
// parseQuitChannels := make([]*chan bool, MaxGoRoutine) //shutdownSignal signals for each go routing
parseOffsets := make([]*RoutingOffset, maxGoRoutine) //kafka fetchOffsets
shutdownSignal := make(chan bool, 1)
finalQuitSignal := make(chan bool, 1)
//handle exit event
exitEventChannel := make(chan os.Signal, 1)
signal.Notify(exitEventChannel, syscall.SIGINT)
signal.Notify(exitEventChannel, os.Interrupt)
go func() {
s := <-exitEventChannel
log.Debug("got signal:", s)
if s == os.Interrupt || s.(os.Signal) == syscall.SIGINT {
log.Warn("got signal:os.Interrupt,saving data and exit")
// defer os.Exit(0)
runtimeConfig.Storage.PersistBloomFilter()
//wait workers to exit
log.Info("waiting workers exit")
go shutdown(fetchOffsets, fetchQuitChannels, parseOffsets, parseQuitChannels, shutdownSignal)
<-shutdownSignal
log.Info("workers shutdown")
finalQuitSignal <- true
}
}()
//start fetcher
for i := 0; i < maxGoRoutine; i++ {
quitC := make(chan bool, 1)
taskC := make(chan []byte)
fetchQuitChannels[i] = &quitC
fetchTaskChannels[i] = &taskC
offset := new(RoutingOffset)
// offset.Offset = initOffset(runtimeConfig, "fetch", i)
offset.Shard = i
fetchOffsets[i] = offset
go task.FetchGo(runtimeConfig, &taskC, &quitC, offset)
}
c2 := make(chan bool, 1)
parseQuitChannels[0] = &c2
offset2 := new(RoutingOffset)
// offset2.Offset = initOffset(runtimeConfig, "parse", 0)
offset2.Shard = 0
parseOffsets[0] = offset2
pendingFetchUrls := make(chan []byte)
//fetch rule:all urls -> persisted to sotre -> fetched from store -> pushed to pendingFetchUrls -> redistributed to sharded goroutines -> fetch -> save webpage to store -> done
//parse rule:url saved to store -> local path persisted to store -> fetched to pendingParseFiles -> redistributed to sharded goroutines -> parse -> clean urls -> enqueue to url store ->done
//sending feed to task queue
go func() {
//notice seed will not been persisted
log.Debug("sending feed to fetch queue,", seedUrl)
pendingFetchUrls <- []byte(seedUrl)
}()
//start local saved file parser
if runtimeConfig.ParseUrlsFromSavedFileLog {
go task.ParseGo(pendingFetchUrls, runtimeConfig, &c2, offset2)
}
//redistribute pendingFetchUrls to sharded workers
go func() {
for {
url := <-pendingFetchUrls
if !runtimeConfig.Storage.CheckWalkedUrl(url) {
if runtimeConfig.Storage.CheckFetchedUrl(url) {
log.Warn("dont hit walk bloomfilter but hit fetch bloomfilter,also ignore,", string(url))
runtimeConfig.Storage.AddWalkedUrl(url)
continue
}
randomShard := 0
if maxGoRoutine > 1 {
randomShard = rand.Intn(maxGoRoutine - 1)
}
log.Debug("publish:", string(url), ",shard:", randomShard)
runtimeConfig.Storage.AddWalkedUrl(url)
*fetchTaskChannels[randomShard] <- url
} else {
log.Trace("hit walk or fetch bloomfilter,just ignore,", string(url))
}
}
}()
//load predefined fetch jobs
if runtimeConfig.LoadTemplatedFetchJob {
go func() {
if util.CheckFileExists(runtimeConfig.TaskConfig.TaskDataPath + "/urls/template.txt") {
templates := util.ReadAllLines(runtimeConfig.TaskConfig.TaskDataPath + "/urls/template.txt")
ids := util.ReadAllLines(runtimeConfig.TaskConfig.TaskDataPath + "/urls/id.txt")
for _, id := range ids {
for _, template := range templates {
log.Trace("id:", id)
log.Trace("template:", template)
url := strings.Replace(template, "{id}", id, -1)
log.Debug("new task from template:", url)
pendingFetchUrls <- []byte(url)
}
}
log.Info("templated download is done.")
}
}()
}
//fetch urls from saved pages
if runtimeConfig.LoadPendingFetchJobs {
c3 := make(chan bool, 1)
parseQuitChannels[1] = &c3
offset3 := new(RoutingOffset)
// offset3.Offset = initOffset(runtimeConfig, "fetch_from_saved", 0)
offset3.Shard = 0
parseOffsets[1] = offset3
go task.LoadTaskFromLocalFile(pendingFetchUrls, &runtimeConfig, &c3, offset3)
}
//parse fetch failed jobs,and will ignore the walk-filter
//TODO
if runtimeConfig.LoadRuledFetchJob {
log.Debug("start ruled fetch")
go func() {
if runtimeConfig.RuledFetchConfig.UrlTemplate != "" {
for i := runtimeConfig.RuledFetchConfig.From; i <= runtimeConfig.RuledFetchConfig.To; i += runtimeConfig.RuledFetchConfig.Step {
url := strings.Replace(runtimeConfig.RuledFetchConfig.UrlTemplate, "{id}", strconv.FormatInt(int64(i), 10), -1)
log.Debug("add ruled url:", url)
pendingFetchUrls <- []byte(url)
}
} else {
log.Error("ruled template is empty,ignore")
}
}()
}
<-finalQuitSignal
log.Info("[gopa] is down")
}
func (this *FsStore) CheckSavedFile(file string) bool {
log.Debug("start check file:", file)
return util.CheckFileExists(file)
}
func extractLinks(runtimeConfig RuntimeConfig, fileUrl string, fileName []byte, body []byte) {
// siteUrlStr := string(fileName)
// siteUrlStr = strings.TrimLeft(siteUrlStr, "data/")
// siteUrlStr = "http://" + siteUrlStr
// log.Debug("fileName to Url:", string(fileName), ",", siteUrlStr)
siteUrlStr := fileUrl
siteConfig := runtimeConfig.TaskConfig
siteUrlByte := []byte(siteUrlStr)
log.Debug("enter links extract,", siteUrlStr)
if siteConfig.SkipPageParsePattern.Match(siteUrlByte) {
log.Debug("hit SkipPageParsePattern pattern,", siteUrlStr)
return
}
log.Debug("parsing external links:", siteUrlStr, ",using:", siteConfig.LinkUrlExtractRegex)
matches := siteConfig.LinkUrlExtractRegex.FindAllSubmatch(body, -1)
log.Debug("extract links with pattern,total matchs:", len(matches), " match result,", string(fileName))
xIndex := 0
for _, match := range matches {
log.Debug("dealing with match result,", xIndex)
xIndex = xIndex + 1
url := match[siteConfig.LinkUrlExtractRegexGroupIndex]
filterUrl := formatUrlForFilter(url)
log.Debug("url clean result:", string(filterUrl), ",original url:", string(url))
filteredUrl := string(filterUrl)
//filter error link
if filteredUrl == "" {
log.Debug("filteredUrl is empty,continue")
continue
}
result1 := strings.HasPrefix(filteredUrl, "#")
if result1 {
log.Debug("filteredUrl started with: # ,continue")
continue
}
result2 := strings.HasPrefix(filteredUrl, "javascript:")
if result2 {
log.Debug("filteredUrl started with: javascript: ,continue")
continue
}
hit := false
// l.Lock();
// defer l.Unlock();
if runtimeConfig.Storage.CheckWalkedUrl(filterUrl) || runtimeConfig.Storage.CheckFetchedUrl(filterUrl) || runtimeConfig.Storage.CheckPendingFetchUrl(filterUrl) {
log.Debug("hit bloomFilter,continue")
hit = true
continue
}
if !hit {
currentUrlStr := string(url)
currentUrlStr = strings.Trim(currentUrlStr, " ")
seedUrlStr := siteUrlStr
seedURI, err := ParseRequestURI(seedUrlStr)
if err != nil {
log.Error("ParseSeedURI failed!: ", seedUrlStr, " , ", err)
continue
}
currentURI1, err := ParseRequestURI(currentUrlStr)
currentURI := currentURI1
if err != nil {
if strings.Contains(err.Error(), "invalid URI for request") {
log.Debug("invalid URI for request,fix relative url,original:", currentUrlStr)
// log.Debug("old relatived url,", currentUrlStr)
//page based relative urls
currentUrlStr = "http://" + seedURI.Host + "/" + currentUrlStr
currentURI1, err = ParseRequestURI(currentUrlStr)
currentURI = currentURI1
if err != nil {
log.Error("ParseCurrentURI internal failed!: ", currentUrlStr, " , ", err)
continue
}
log.Debug("new relatived url,", currentUrlStr)
} else {
log.Error("ParseCurrentURI failed!: ", currentUrlStr, " , ", err)
continue
}
}
// relative links
if currentURI == nil || currentURI.Host == "" {
if strings.HasPrefix(currentURI.Path, "/") {
//root based relative urls
log.Debug("old relatived url,", currentUrlStr)
currentUrlStr = "http://" + seedURI.Host + currentUrlStr
log.Debug("new relatived url,", currentUrlStr)
} else {
log.Debug("old relatived url,", currentUrlStr)
//page based relative urls
urlPath := getRootUrl(currentURI)
currentUrlStr = "http://" + urlPath + currentUrlStr
log.Debug("new relatived url,", currentUrlStr)
}
} else {
log.Debug("host:", currentURI.Host, " ", currentURI.Host == "")
//resolve domain specific filter
if siteConfig.FollowSameDomain {
if siteConfig.FollowSubDomain {
//TODO handler com.cn and .com,using a TLC-domain list
} else if seedURI.Host != currentURI.Host {
log.Debug("domain mismatch,", seedURI.Host, " vs ", currentURI.Host)
//continue
}
//TODO follow all or list of domain
}
}
if len(siteConfig.LinkUrlMustContain) > 0 {
if !util.ContainStr(currentUrlStr, siteConfig.LinkUrlMustContain) {
log.Debug("link does not hit must-contain,ignore,", currentUrlStr, " , ", siteConfig.LinkUrlMustNotContain)
continue
}
}
if len(siteConfig.LinkUrlMustNotContain) > 0 {
if util.ContainStr(currentUrlStr, siteConfig.LinkUrlMustNotContain) {
log.Debug("link hit must-not-contain,ignore,", currentUrlStr, " , ", siteConfig.LinkUrlMustNotContain)
continue
}
}
//normalize url
currentUrlStr = MustNormalizeURLString(currentUrlStr, FlagLowercaseScheme|FlagLowercaseHost|FlagUppercaseEscapes|
FlagRemoveUnnecessaryHostDots|FlagRemoveDuplicateSlashes|FlagRemoveFragment)
log.Debug("normalized url:", currentUrlStr)
currentUrlByte := []byte(currentUrlStr)
if !(runtimeConfig.Storage.CheckWalkedUrl(currentUrlByte) || runtimeConfig.Storage.CheckFetchedUrl(currentUrlByte) || runtimeConfig.Storage.CheckPendingFetchUrl(currentUrlByte)) {
//bloomFilter.Lookup(currentUrlByte) {
// if(CheckIgnore(currentUrlStr)){}
// log.Info("enqueue fetch: ", currentUrlStr)
// broker.Publish(kafka.NewMessage(currentUrlByte))
//copied form fetchTask,TODO refactor
//checking fetchUrlPattern
log.Debug("started check fetchUrlPattern,", currentUrlStr)
if siteConfig.FetchUrlPattern.Match(currentUrlByte) {
log.Debug("match fetch url pattern,", currentUrlStr)
if len(siteConfig.FetchUrlMustNotContain) > 0 {
if util.ContainStr(currentUrlStr, siteConfig.FetchUrlMustNotContain) {
log.Debug("hit FetchUrlMustNotContain,ignore,", currentUrlStr)
continue
}
}
if len(siteConfig.FetchUrlMustContain) > 0 {
if !util.ContainStr(currentUrlStr, siteConfig.FetchUrlMustContain) {
log.Debug("not hit FetchUrlMustContain,ignore,", currentUrlStr)
continue
}
}
} else {
log.Debug("does not hit FetchUrlPattern ignoring,", currentUrlStr)
continue
}
if !runtimeConfig.Storage.CheckPendingFetchUrl(currentUrlByte) {
log.Debug("log new pendingFetch url", currentUrlStr)
runtimeConfig.Storage.LogPendingFetchUrl(runtimeConfig.PathConfig.PendingFetchLog, currentUrlStr)
runtimeConfig.Storage.AddPendingFetchUrl(currentUrlByte)
} else {
log.Debug("hit new pendingFetch filter,ignore:", currentUrlStr)
}
// pendingUrls <- currentUrlByte
// TODO pendingFetchFilter bloomFilter.Add(currentUrlByte)
} else {
log.Debug("hit bloom filter,ignore:", currentUrlStr)
}
// bloomFilter.Add([]byte(filterUrl))
} else {
log.Debug("hit bloom filter,ignore,", string(url))
}
log.Debug("exit links extract,", siteUrlStr)
}
//TODO 处理ruled fetch pattern
log.Info("all links within ", siteUrlStr, " is done")
}
//fetch url's content
func fetchUrl(url []byte, timeout time.Duration, runtimeConfig RuntimeConfig, offsets *RoutingOffset) {
t := time.NewTimer(timeout)
defer t.Stop()
resource := string(url)
log.Debug("enter fetchUrl method:", resource)
config := runtimeConfig.TaskConfig
if runtimeConfig.Storage.CheckFetchedUrl(url) {
return
}
path := getSavedPath(runtimeConfig, url)
if runtimeConfig.Storage.CheckSavedFile(path) {
log.Warn("file is already saved,skip fetch.", path)
runtimeConfig.Storage.AddSavedUrl(url)
//re-parse local's previous saved page
if runtimeConfig.ParseUrlsFromPreviousSavedPage {
if !runtimeConfig.Storage.CheckParsedFile([]byte(path)) {
log.Debug("previous saved page send to parse-queue:", path)
runtimeConfig.Storage.LogSavedFile(runtimeConfig.PathConfig.SavedFileLog, resource+"|||"+path)
}
}
return
}
//checking fetchUrlPattern
log.Debug("started check fetchUrlPattern,", config.FetchUrlPattern, ",", resource)
if config.FetchUrlPattern.Match(url) {
log.Debug("match fetch url pattern,", resource)
if len(config.FetchUrlMustNotContain) > 0 {
if util.ContainStr(resource, config.FetchUrlMustNotContain) {
log.Debug("hit FetchUrlMustNotContain,ignore,", resource, " , ", config.FetchUrlMustNotContain)
return
}
}
if len(config.FetchUrlMustContain) > 0 {
if !util.ContainStr(resource, config.FetchUrlMustContain) {
log.Debug("not hit FetchUrlMustContain,ignore,", resource, " , ", config.FetchUrlMustContain)
return
}
}
} else {
log.Debug("does not hit FetchUrlPattern ignoring,", resource)
return
}
log.Debug("start fetch url,", resource)
flg := make(chan bool, 1)
go func() {
body, err := HttpGetWithCookie(resource, config.Cookie)
if err == nil {
if body != nil {
//todo parse urls from this page
log.Debug("started check savingUrlPattern,", config.SavingUrlPattern, ",", string(url))
if config.SavingUrlPattern.Match(url) {
log.Debug("match saving url pattern,", resource)
if len(config.SavingUrlMustNotContain) > 0 {
if util.ContainStr(resource, config.SavingUrlMustNotContain) {
log.Debug("hit SavingUrlMustNotContain,ignore,", resource, " , ", config.SavingUrlMustNotContain)
goto exitPage
}
}
if len(config.SavingUrlMustContain) > 0 {
if !util.ContainStr(resource, config.SavingUrlMustContain) {
log.Debug("not hit SavingUrlMustContain,ignore,", resource, " , ", config.SavingUrlMustContain)
goto exitPage
}
}
_, err := Save(runtimeConfig, path, body)
if err == nil {
log.Info("saved:", path)
//todo saved per shard
runtimeConfig.Storage.LogSavedFile(runtimeConfig.PathConfig.SavedFileLog, resource+"|||"+path)
} else {
log.Info("error while saved:", path, ",", err)
goto exitPage
}
} else {
log.Debug("does not hit SavingUrlPattern ignoring,", resource)
}
}
runtimeConfig.Storage.AddFetchedUrl(url)
exitPage:
log.Debug("exit fetchUrl method:", resource)
} else {
// runtimeConfig.Storage.AddFetchFailedUrl(url)
runtimeConfig.Storage.LogFetchFailedUrl(runtimeConfig.PathConfig.FetchFailedLog, resource)
}
flg <- true
}()
//监听通道，由于设有超时，不可能泄露
select {
case <-t.C:
log.Error("fetching url time out,", resource)
case <-flg:
log.Debug("fetching url normal exit,", resource)
return
}
}