[MiroDev] Scraping thumbnails and spacer gifs.

[will]

I thought we used the feed thumbnail and then when the user downloads 
the movie, we use a thumbnail we generate from the movie.

When do we scrape the description?

/will


Ben Dean-Kawamura wrote:
> Currenly our wired science feed
> (http://feeds.pbs.org/pbs/kcet/wiredscience-video) has blank
> thumbnails.
> 
> This happens because WIRED doesn't provide us with a thumbnail for the
> items, so we try to scrape them from the description.  The
> descriptions contain 1x1 spacer GIFs
> (http://feeds.pbs.org/~r/pbs/kcet/wiredscience-video/~4/203608185)
> that we then happily use for our thumbnail.
> 
> What's the solution for this?  Maybe we should try to detect images
> under a certain size (200B?) and not use them as scraped images if so.
>  This would be a little bit of work, but not too much.  Maybe an hour
> or 2.  Is the problem happening with other feeds?  Is it worth fixing?
> 
> Ben