Exemple #1
0
func main() {
	log.Println("Running tocrawl tool, to generate crawling list")
	log.Println("NonCrawlHosts: ", gcse.NonCrawlHosts)
	log.Println("CrawlGithubUpdate: ", gcse.CrawlGithubUpdate)
	log.Println("CrawlByGodocApi: ", gcse.CrawlByGodocApi)
	// Load CrawlerDB
	cDB = gcse.LoadCrawlerDB()

	if gcse.CrawlGithubUpdate || gcse.CrawlByGodocApi {
		// load pkgUTs
		pkgUTs, err := loadPackageUpdateTimes(
			sophie.LocalFsPath(gcse.DocsDBPath.S()))
		if err != nil {
			log.Fatalf("loadPackageUpdateTimes failed: %v", err)
		}

		if gcse.CrawlGithubUpdate {
			touchByGithubUpdates(pkgUTs)
		}

		if gcse.CrawlByGodocApi {
			httpClient := gcse.GenHttpClient("")
			pkgs, err := gcse.FetchAllPackagesInGodoc(httpClient)
			if err != nil {
				log.Fatalf("FetchAllPackagesInGodoc failed: %v", err)
			}
			log.Printf("FetchAllPackagesInGodoc returns %d entries", len(pkgs))
			for _, pkg := range pkgs {
				cDB.AppendPackage(pkg, func(pkg string) bool {
					_, ok := pkgUTs[pkg]
					return ok
				})
			}
		}
		syncDatabases()
	}

	log.Printf("Package DB: %d entries", cDB.PackageDB.Count())
	log.Printf("Person DB: %d entries", cDB.PersonDB.Count())

	pathToCrawl := gcse.DataRoot.Join(gcse.FnToCrawl)

	kvPackage := kv.DirOutput(sophie.LocalFsPath(
		pathToCrawl.Join(gcse.FnPackage).S()))
	kvPackage.Clean()
	if err := generateCrawlEntries(cDB.PackageDB, gcse.HostOfPackage, kvPackage); err != nil {
		log.Fatalf("generateCrawlEntries %v failed: %v", kvPackage.Path, err)
	}

	kvPerson := kv.DirOutput(sophie.LocalFsPath(
		pathToCrawl.Join(gcse.FnPerson).S()))

	kvPerson.Clean()
	if err := generateCrawlEntries(cDB.PersonDB, func(id string) string {
		site, _ := gcse.ParsePersonId(id)
		return site
	}, kvPerson); err != nil {
		log.Fatalf("generateCrawlEntries %v failed: %v", kvPerson.Path, err)
	}
}
Exemple #2
0
func main() {
	defer func() {
		tmpFn := villa.Path("/tmp/gddo")
		if err := tmpFn.RemoveAll(); err != nil {
			log.Printf("Delete %v failed: %v", tmpFn, err)
		}
	}()

	singlePackge := ""
	singleETag := ""
	flag.StringVar(&singlePackge, "pkg", singlePackge, "Crawling single package")
	flag.StringVar(&singleETag, "etag", singleETag, "ETag for single package crawling")

	flag.Parse()

	httpClient := gcse.GenHttpClient("")

	if singlePackge != "" {
		log.Printf("Crawling single package %s ...", singlePackge)
		p, err := gcse.CrawlPackage(httpClient, singlePackge, singleETag)
		if err != nil {
			fmtp.Printfln("Crawling package %s failured: %v", singlePackge, err)
		} else {
			fmtp.Printfln("Package %s: %+v", singlePackge, p)
		}
		return
	}

	log.Println("crawler started...")

	// Load CrawlerDB
	cDB = gcse.LoadCrawlerDB()

	fpDataRoot := sophie.FsPath{
		Fs:   sophie.LocalFS,
		Path: gcse.DataRoot.S(),
	}

	fpDocs := fpDataRoot.Join(gcse.FnDocs)
	if err := loadAllDocsPkgs(kv.DirInput(fpDocs)); err != nil {
		log.Fatalf("loadAllDocsPkgs: %v", err)
	}
	log.Printf("%d docs loaded!", len(allDocsPkgs))

	AppStopTime = time.Now().Add(gcse.CrawlerDuePerRun)

	//pathToCrawl := gcse.DataRoot.Join(gcse.FnToCrawl)
	fpCrawler := fpDataRoot.Join(gcse.FnCrawlerDB)
	fpToCrawl := fpDataRoot.Join(gcse.FnToCrawl)

	fpNewDocs := fpCrawler.Join(gcse.FnNewDocs)
	fpNewDocs.Remove()

	pkgEnd := make(chan error, 1)
	go crawlPackages(httpClient, fpToCrawl.Join(gcse.FnPackage), fpNewDocs,
		pkgEnd)

	psnEnd := make(chan error, 1)
	go crawlPersons(httpClient, fpToCrawl.Join(gcse.FnPerson), psnEnd)

	errPkg, errPsn := <-pkgEnd, <-psnEnd
	if errPkg != nil || errPsn != nil {
		log.Fatalf("Some job may failed, package: %v, person: %v",
			errPkg, errPsn)
	}

	if err := processImports(); err != nil {
		log.Printf("processImports failed: %v", err)
	}

	syncDatabases()
	log.Println("crawler stopped...")
}
Exemple #3
0
func crawlEntriesLoop() {
	httpClient := gcse.GenHttpClient("")

	for time.Now().Before(AppStopTime) {
		checkImports()

		if gcse.CrawlByGodocApi {
			processGodoc(httpClient)
		}

		didSomething := false
		var wg sync.WaitGroup

		pkgGroups := listPackagesByHost(5, 50)
		if len(pkgGroups) > 0 {
			didSomething = true

			log.Printf("Crawling packages of %d groups", len(pkgGroups))

			wg.Add(len(pkgGroups))

			for host, ents := range pkgGroups {
				go func(host string, ents []EntryInfo) {
					failCount := 0
					for _, ent := range ents {
						if time.Now().After(AppStopTime) {
							break
						}
						runtime.GC()
						p, err := gcse.CrawlPackage(httpClient, ent.ID, ent.Etag)
						if err != nil && err != gcse.ErrPackageNotModifed {
							log.Printf("Crawling pkg %s failed: %v", ent.ID, err)

							if gcse.IsBadPackage(err) {
								// a wrong path
								deletePackage(ent.ID)
								log.Printf("Remove wrong package %s", ent.ID)
							} else {
								failCount++

								schedulePackage(ent.ID, time.Now().Add(
									12*time.Hour), ent.Etag)

								if failCount >= 10 {
									durToSleep := 10 * time.Minute
									if time.Now().Add(durToSleep).After(AppStopTime) {
										break
									}

									log.Printf("Last ten crawling %s packages failed, sleep for a while...",
										host)
									time.Sleep(durToSleep)
									failCount = 0
								}
							}
							continue
						}

						failCount = 0
						if err == gcse.ErrPackageNotModifed {
							log.Printf("Package %s unchanged!", ent.ID)
							schedulePackageNextCrawl(ent.ID, ent.Etag)
							continue
						}

						log.Printf("Crawled package %s success!", ent.ID)

						pushPackage(p)
						log.Printf("Package %s saved!", ent.ID)
					}

					wg.Done()
				}(host, ents)
			}
		}

		personGroups := listPersonsByHost(5, 100)
		if len(personGroups) > 0 {
			didSomething = true

			log.Printf("Crawling persons of %d groups", len(personGroups))

			wg.Add(len(personGroups))

			for host, ents := range personGroups {
				go func(host string, ents []EntryInfo) {
					failCount := 0
					for _, ent := range ents {
						if time.Now().After(AppStopTime) {
							break
						}

						p, err := gcse.CrawlPerson(httpClient, ent.ID)
						if err != nil {
							failCount++
							log.Printf("Crawling person %s failed: %v", ent.ID, err)

							schedulePerson(ent.ID, time.Now().Add(12*time.Hour))

							if failCount >= 10 {
								durToSleep := 10 * time.Minute
								if time.Now().Add(durToSleep).After(AppStopTime) {
									break
								}

								log.Printf("Last ten crawling %s persons failed, sleep for a while...",
									host)
								time.Sleep(durToSleep)
								failCount = 0
							}
							continue
						}

						log.Printf("Crawled person %s success!", ent.ID)
						pushPerson(p)
						log.Printf("Push person %s success", ent.ID)
						failCount = 0
					}

					wg.Done()
				}(host, ents)
			}
		}
		wg.Wait()

		syncDatabases()

		if gcse.CrawlGithubUpdate {
			if touchByGithubUpdates() {
				didSomething = true
			}
		}

		if !didSomething {
			log.Printf("Nothing to crawl sleep for a while...")
			time.Sleep(2 * time.Minute)
		}
	}
}
Exemple #4
0
func main() {
	log.Println("Running tocrawl tool, to generate crawling list")
	log.Println("NonCrawlHosts: ", configs.NonCrawlHosts)
	log.Println("CrawlGithubUpdate: ", configs.CrawlGithubUpdate)
	log.Println("CrawlByGodocApi: ", configs.CrawlByGodocApi)

	log.Printf("Using personal: %v", configs.CrawlerGithubPersonal)
	gcse.GithubSpider = github.NewSpiderWithToken(configs.CrawlerGithubPersonal)

	// Load CrawlerDB
	cDB = gcse.LoadCrawlerDB()

	if configs.CrawlGithubUpdate || configs.CrawlByGodocApi {
		// load pkgUTs
		pkgUTs, err := loadPackageUpdateTimes(
			sophie.LocalFsPath(configs.DocsDBPath().S()))
		if err != nil {
			log.Fatalf("loadPackageUpdateTimes failed: %v", err)
		}

		if configs.CrawlGithubUpdate {
			touchByGithubUpdates(pkgUTs)
		}

		if configs.CrawlByGodocApi {
			httpClient := gcse.GenHttpClient("")
			pkgs, err := gcse.FetchAllPackagesInGodoc(httpClient)
			if err != nil {
				log.Fatalf("FetchAllPackagesInGodoc failed: %v", err)
			}
			gcse.AddBiValueAndProcess(bi.Max, "godoc.doc-count", len(pkgs))
			log.Printf("FetchAllPackagesInGodoc returns %d entries", len(pkgs))
			now := time.Now()
			for _, pkg := range pkgs {
				cDB.AppendPackage(pkg, func(pkg string) bool {
					_, ok := pkgUTs[pkg]
					return ok
				})
				site, path := utils.SplitPackage(pkg)
				if err := store.AppendPackageEvent(site, path, "godoc", now, sppb.HistoryEvent_Action_None); err != nil {
					log.Printf("UpdatePackageHistory %s %s failed: %v", site, path, err)
				}
			}
		}
		syncDatabases()
	}

	log.Printf("Package DB: %d entries", cDB.PackageDB.Count())
	log.Printf("Person DB: %d entries", cDB.PersonDB.Count())

	pathToCrawl := configs.DataRoot.Join(configs.FnToCrawl)

	kvPackage := kv.DirOutput(sophie.LocalFsPath(
		pathToCrawl.Join(configs.FnPackage).S()))
	kvPackage.Clean()
	if err := generateCrawlEntries(cDB.PackageDB, gcse.HostOfPackage, kvPackage); err != nil {
		log.Fatalf("generateCrawlEntries %v failed: %v", kvPackage.Path, err)
	}

	kvPerson := kv.DirOutput(sophie.LocalFsPath(
		pathToCrawl.Join(configs.FnPerson).S()))

	kvPerson.Clean()
	if err := generateCrawlEntries(cDB.PersonDB, func(id string) string {
		site, _ := gcse.ParsePersonId(id)
		return site
	}, kvPerson); err != nil {
		log.Fatalf("generateCrawlEntries %v failed: %v", kvPerson.Path, err)
	}
}
Exemple #5
0
func main() {
	runtime.GOMAXPROCS(2)

	log.Printf("Using personal: %v", configs.CrawlerGithubPersonal)
	gcse.GithubSpider = github.NewSpiderWithToken(configs.CrawlerGithubPersonal)

	if db, err := bh.Open(configs.DataRoot.Join("filecache.bolt").S(), 0644, nil); err == nil {
		log.Print("Using file cache!")
		gcse.GithubSpider.FileCache = spider.BoltFileCache{
			DB:         db,
			IncCounter: bi.Inc,
		}
	} else {
		log.Printf("Open file cache failed: %v", err)
	}

	cleanTempDir()
	defer cleanTempDir()

	singlePackage := flag.String("pkg", "", "Crawling a single package")
	singleETag := flag.String("etag", "", "ETag for the single package crawling")
	singlePerson := flag.String("person", "", "Crawling a single person")

	flag.Parse()

	httpClient := gcse.GenHttpClient("")

	if *singlePerson != "" {
		log.Printf("Crawling single person %s ...", *singlePerson)
		p, err := gcse.CrawlPerson(httpClient, *singlePerson)
		if err != nil {
			fmtp.Printfln("Crawling person %s failed: %v", *singlePerson, err)
		} else {
			fmtp.Printfln("Person %s: %+v", *singlePerson, p)
		}
	}
	if *singlePackage != "" {
		log.Printf("Crawling single package %s ...", *singlePackage)
		p, flds, err := gcse.CrawlPackage(httpClient, *singlePackage, *singleETag)
		if err != nil {
			fmtp.Printfln("Crawling package %s failed: %v, folders: %v", *singlePackage, err, flds)
		} else {
			fmtp.Printfln("Package %s: %+v, folders: %v", *singlePackage, p, flds)
		}
	}
	if *singlePackage != "" || *singlePerson != "" {
		return
	}

	log.Println("crawler started...")

	// Load CrawlerDB
	cDB = gcse.LoadCrawlerDB()

	fpDataRoot := sophie.FsPath{
		Fs:   sophie.LocalFS,
		Path: configs.DataRoot.S(),
	}

	fpDocs := fpDataRoot.Join(configs.FnDocs)
	if err := loadAllDocsPkgs(kv.DirInput(fpDocs)); err != nil {
		log.Fatalf("loadAllDocsPkgs: %v", err)
	}
	log.Printf("%d docs loaded!", len(allDocsPkgs))

	AppStopTime = time.Now().Add(configs.CrawlerDuePerRun)

	//pathToCrawl := gcse.DataRoot.Join(gcse.FnToCrawl)
	fpCrawler := fpDataRoot.Join(configs.FnCrawlerDB)
	fpToCrawl := fpDataRoot.Join(configs.FnToCrawl)

	fpNewDocs := fpCrawler.Join(configs.FnNewDocs)
	fpNewDocs.Remove()

	if err := processImports(); err != nil {
		log.Printf("processImports failed: %v", err)
	}

	pkgEnd := make(chan error, 1)
	go crawlPackages(httpClient, fpToCrawl.Join(configs.FnPackage), fpNewDocs, pkgEnd)

	psnEnd := make(chan error, 1)
	go crawlPersons(httpClient, fpToCrawl.Join(configs.FnPerson), psnEnd)

	errPkg, errPsn := <-pkgEnd, <-psnEnd
	bi.Flush()
	bi.Process()
	syncDatabases()
	if errPkg != nil || errPsn != nil {
		log.Fatalf("Some job may failed, package: %v, person: %v", errPkg, errPsn)
	}
	log.Println("crawler stopped...")
}