Take length of node content into account

[psolbach]

I saw menu structures of <ul> nodes pop up alongside the desired text body in libextract's results. This could be easily mitigated through defining a minimum threshold of node content length. Something like 15 characters possibly. Will do a PR if this is a wanted improvement. Not sure if the project is still alive.

[eugene-eeo]

Hi, sure; please do write a PR if you have the time and also attach the differences in output, i.e. the output given by code prior to the PR and after the PR.