Apify

Apify Discord Mirror

Updated 2 weeks ago

How to implement persistent login with crawlee-js/playwright?

oosenvosem
·
I need to scrape content on multiple pages in one social network (x.com) that requires auth. Where to implement the login mechanism in order to it happened before following urls and persisted to use it until it is valid?
2
!
C
O
10 comments
!! Spooky
await page.context().storageState({ path: authFilePath })
Look up storageState() on the playwright docs
CCrafty
that gets the cookie and localstorage state but how do you load it into a new session. im facing the same dilema
OOleg V.
You can store you cookies in named KV store and then modify Your session with function:
https://crawlee.dev/api/next/core/interface/SessionPoolOptions#createSessionFunction

or You can do the same (update your requests with those cookies) in preNavigationHooks (https://crawlee.dev/api/next/browser-crawler/interface/BrowserCrawlerOptions#preNavigationHooks)
CCrafty
what about local storage? im supprised there seems to be no easy way to seed a session with local data
CCrafty
and also, setCookie wants a raw cookie string and a url rather than the format the getState() gives you
CCrafty
@osenvosem I have found a solution though it isnt great. If when creating your scraper, you set useIncognitoPages to true, you can modify the pageOptions to set your cookies and localStorage inside of a prePageCreateHook


Plain Text
const authStorage = await KeyValueStore.open('auth');

const crawler = new PlaywrightCrawler({
    launchContext: {
      launcher: chromium,
      useIncognitoPages: true
    },
    requestHandler: router,
    browserPoolOptions: {
      prePageCreateHooks: [
        async (pageId, browserController, pageOptions) => {
          if (!pageOptions) { // pageOptions is only exposed in incognito
            throw new Error("no page options")
          }
          pageOptions.storageState = await authStorage.getValue("state") ?? undefined
        }
      ]
    }
});


unfortunately this seems to add a fair chunk of overhead though.
CCrafty
crawl with incognito

INFO PlaywrightCrawler: Final request statistics:
Plain Text
{"requestsFinished":37,"requestsFailed":0,"retryHistogram":[37],"requestAvgFailedDurationMillis":null,"requestAvgFinishedDurationMillis":3514,"requestsFinishedPerMinute":53,"requestsFailedPerMinute":0,"requestTotalDurationMillis":130001,"requestsTotal":37,"crawlerRuntimeMillis":42227}


without incognito (and the hook)

INFO PlaywrightCrawler: Final request statistics:
Plain Text
{"requestsFinished":37,"requestsFailed":0,"retryHistogram":[37],"requestAvgFailedDurationMillis":null,"requestAvgFinishedDurationMillis":2238,"requestsFinishedPerMinute":67,"requestsFailedPerMinute":0,"requestTotalDurationMillis":82813,"requestsTotal":37,"crawlerRuntimeMillis":32955}
AApifyBot
@Crafty just advanced to level 4! Thanks for your contributions! 🎉
CCrafty
in the future, it may be possibe to do the same using experimentalContainers instead of useIncognitoPages though it doesnt seem to work yet
ВВан\
You can implement the login in the handlePageFunction of Crawlee’s PlaywrightCrawler. Before navigating to any URLs, you'll first need to handle the login, and then persist the cookies or local storage to reuse the session across subsequent requests.
Add a reply
Sign up and join the conversation on Discord