func _crawlAmebloPost(req *wcg.Request, post *hplink.AmebloPost, members []hplink.Member) error {
crawler := crawlers.NewAmebloPostCrawler(cacheutil.NewURLCacheAwareClient(req, entities.URLCache))
req.Logger.Infof("[Task.Crawlers.AmebloPosts] crawling URL: %s", post.URL)
crawled, err := crawler.RunOnPostURL(post.URL)
if err != nil {
return err
}
post.NumLikes = crawled.NumLikes
post.NumComments = crawled.NumComments
post.NumReblogs = crawled.NumReblogs
post.PostAt = crawled.PostAt
post.Title = crawled.Title
post.Theme = crawled.Theme
post.IsContentsCrawled = true
post.Images = make([]models.Image, len(crawled.ImageURLs))
if cacher, err := cacheutil.NewImageCacher(req, entities.ImageCache); err == nil {
iterator.ParallelSlice(crawled.ImageURLs, func(i int, url string) error {
req.Logger.Infof("[Task.Crawlers.AmebloPosts] caching image URL %s (on %s)", url, post.URL)
const retries = 5
var err error
var cache *models.ImageCache
for j := 0; j < retries; j++ {
if j != 0 {
req.Logger.Infof("[Task.Crawlers.AmebloPosts] Retry image URL caching: %s (on %s) (%d)", url, post.URL, j)
}
cache, err = cacher.Cache(url)
if err == nil {
post.Images[i] = *cache.ToImage()
return nil
}
lib.WaitAndEnsureAfter(lib.Now(), 3*time.Second)
}
req.Logger.Warnf("[Task.Crawlers.AmebloPosts] Failed to cache image URL %s (on %s) - %v", url, post.URL, err)
return err
})
} else {
for j, url := range crawled.ImageURLs {
req.Logger.Infof("[Task.Crawlers.AmebloPosts] skip caching image URL %s (on %s)", url, post.URL)
post.Images[j] = models.Image{
URL: url,
}
}
}
// Update MemberKey only if not set.
if post.MemberKey == "" {
if m := _guessMember(req, post, members); m != nil {
post.MemberKey = m.Key
}
}
req.Logger.Infof("[Task.Crawlers.AmebloPosts] finished crawling")
return nil
}
func runTasksCrawlersAmebloPosts(req *wcg.Request, task *models.AsyncAPITask) (*models.AsyncAPITaskProgress, error) {
// Tuning conditions to set the proper `limit`` and `concurrency`` value:
//
// - The task should be done in 10 minutes.
// - We'll invoke the number of goroutines defined in `concurrency`.
// - Each goroutine can access one URL per `urlCrawlWait` seconds (fixed value - 5 seconds)
//
// So we finally gets N (< 120) URLs updated
//
concurrency := configs.GetIntValue(req, "hplink.ameblo_crawler_post_concurrency", 3, 1, 10)
perThread := configs.GetIntValue(req, "hplink.ameblo_crawler_num_posts_per_thread", 30, 1, 100)
wait := configs.GetIntValue(req, "hplink.ameblo_crawler_url_wait", 2, 0, 10)
selected, err := _selectAmebloPosts(req, perThread*concurrency)
if err != nil {
return nil, err
}
// prepare members for MemberKey detection
p, err := entities.Member.Query().Execute(req)
if err != nil {
return nil, err
}
members := p.Data.([]hplink.Member)
req.Logger.Infof("[Task.Crawlers.AmebloPosts] Crawling %d URLs (concurrency: %d)", len(selected), concurrency)
iterator.ParallelSliceWithMaxConcurrency(selected, concurrency, func(i int, post *hplink.AmebloPost) error {
startTime := lib.Now()
if err := _crawlAmebloPost(req, post, members); err != nil {
req.Logger.Errorf("Crawler Failure: %v", err)
return err
}
selected[i] = post
lib.WaitAndEnsureAfter(startTime, time.Duration(wait)*time.Second)
return nil
})
req.Logger.Infof("[Task.Crawlers.AmebloPosts] Updating datastore.")
_, err = entities.AmebloPost.PutMulti().Update(req, selected)
if err != nil {
return nil, err
}
return nil, nil
}
func runTasksCrawlersAmebloEntryLists(req *wcg.Request, task *models.AsyncAPITask) (*models.AsyncAPITaskProgress, error) {
const FollowLinkKey = "fl"
const SettingsKey = "s"
const URLKey = "u"
var query = req.HTTPRequest().URL.Query()
var settingsList []*hplink.CrawlerSettings
var urlList []string
if settingsKeys, ok := query[SettingsKey]; ok {
_, _list := entities.CrawlerSettings.GetMulti().Keys(settingsKeys...).MustList(req)
settingsList = _list.([]*hplink.CrawlerSettings)
} else {
query := entities.CrawlerSettings.Query().Filter("Type=", hplink.CrawlerSettingsTypeAmeblo)
if pagination := query.MustExecute(req); pagination.Length() > 0 {
list := pagination.Data.([]hplink.CrawlerSettings)
settingsList = make([]*hplink.CrawlerSettings, len(list))
for i := range list {
settingsList[i] = &list[i]
}
}
}
var numList = len(settingsList)
urlList = make([]string, numList)
if urls, ok := query[URLKey]; ok {
if numList != len(urls) {
return nil, fmt.Errorf("List mismatch - found %d settings but %d urls are specified", numList, len(urls))
}
urlList = query[URLKey]
} else {
for i := range settingsList {
urlList[i] = (*hplink.AmebloCrawlerSettings)(settingsList[i]).GetEntryListURL()
}
}
startTime := lib.Now()
nextParamSettingsKeys := make([]string, numList)
nextParamURLs := make([]string, numList)
err := iterator.ParallelSlice(settingsList, func(i int, v *hplink.CrawlerSettings) error {
next, err := _crawlAmebloEntryList(req, v, urlList[i])
if err != nil {
settingsList[i].Error = []byte(fmt.Sprintf("%v", err))
settingsList[i].Status = hplink.CrawlerStatusFailure
settingsList[i].LastRun = lib.Now()
return err
}
settingsList[i].Error = nil
settingsList[i].Status = hplink.CrawlerStatusSuccess
settingsList[i].LastRun = lib.Now()
if next != "" {
nextParamSettingsKeys[i] = v.URL
nextParamURLs[i] = next
}
return nil
})
entities.CrawlerSettings.PutMulti().MustUpdate(req, settingsList)
if err != nil {
return nil, err
}
if req.Query(FollowLinkKey) != "true" {
return nil, err
}
// fl=true make a recursive call to follow next links
// reduce empty urls from nextParam* and return it for recursive call
var fixedNextParamSettingsKeys []string
var fixedNextParamURLs []string
var hasNext = false
for i := range nextParamURLs {
if nextParamURLs[i] != "" {
hasNext = true
fixedNextParamSettingsKeys = append(fixedNextParamSettingsKeys, nextParamSettingsKeys[i])
fixedNextParamURLs = append(fixedNextParamURLs, nextParamURLs[i])
}
}
var progress models.AsyncAPITaskProgress
var lastProgress = task.LastProgress()
if lastProgress == nil {
progress.Current = len(urlList)
progress.Total = 0
} else {
progress.Current = lastProgress.Current + len(urlList)
}
if hasNext {
progress.Next = url.Values{
FollowLinkKey: []string{"true"},
SettingsKey: fixedNextParamSettingsKeys,
URLKey: fixedNextParamURLs,
}
wait := configs.GetIntValue(req, "hplink.ameblo_crawler_url_wait", 2, 0, 10)
lib.WaitAndEnsureAfter(startTime, time.Duration(wait)*time.Second)
}
req.Logger.Infof("No more URL needs to be crawled.")
return &progress, nil
}
