Automatically remove non-working videos resources

[benoit74]

I'm not sure it is feasible, but we know that many videos are not working currently. The first seconds (at least) of the video are however crawled and pushed into the ZIM, even if not usable in the end.

We hence end-up with big ZIMs for nothing.

Manual inspection of the ZIM can help to identify URLs which must be ignored, and openzim/zimit#353 can help to excude these resources from the crawl / warc.

It would be super cool if we could do it in a more automated way inside warc2zim.

This is somehow a companion issue of #396