webrecorder / browsertrix-behaviors

Automated behaviors that run in browser to interact with complex sites automatically. Used by ArchiveWeb.page and Browsertrix Crawler.
GNU Affero General Public License v3.0
33 stars 18 forks source link

Behavior Bug: Facebook behavior not working #59

Open nvanderperren opened 1 year ago

nvanderperren commented 1 year ago

URL https://www.facebook.com/meemoo.be (or any other facebook URL)

Describe the bug When starting autopilot, it loads for some minutes and then autopilot stops. I think there's something wrong with the behavior.

What tool do you see this issue in?

Screenshots screenshot

Additional context see https://github.com/webrecorder/archiveweb.page/issues/185 for the autopilot issue.

Chickensoupwithrice commented 1 year ago

I am able to reproduce this issue. Loading my facebook timeline and then enabling autopilot it runs for a couple posts before being marked "complete".

Will continue to investigate.

Chickensoupwithrice commented 1 year ago

So there's 3 parts to the facebook behavior, and I checked all 3 of them, and none of them quite work right:

While the videos behavior does scroll through the videos, it didn't click on any of them and as a result didn't save the videos to the generated WACZ.

Timeline scrolling does mostly work, clicking through the cards, but occasionally it'll come up against a card it doesn't know how to close, and then be considered finished. The timeline behavior also is unsure how to scroll any pages that aren't the default user timeline.

While videos will be a fairly straightforward fix, the timeline behavior would need some more investigation for both: