Allow bytestring arguments

[mahmoud]

Hi there!

I noticed that align() expects text arguments and immediately encodes them into bytestrings. In many cases, developers already have bytestrings available, and decoding them just to encode them again is pretty inefficient

I would propose making the lines linked conditional, or moving them to a separate wrapper around an also-public bytestring function.

Other issues/questions:

• I don't think calling .encode() with no arguments and relying on sys's default encoding is a very safe pattern. Is there a reason not to do utf8 explicitly? Or maybe allow the encoding to be passed in?
• I may have skimmed this wrong, but realistically, isn't the result including offsets that are based on the bytestrings, making the result difficult to interpret in the context of decoded unicode anyways?

Excited to have found your library, thanks for all the great work! If I can help with a PR, let me know!

[Martinsos]

Hi @mahmoud, thank you for reaching out :)! Those are very good suggestions, I think I actually wasn't sure at that moment about that part so I decided to leave it default, but it is certainly better to get this fixed with 'utf-8' and reduce possibility of bugs and something going wrong.

Thanks for the offer, PR would be awesome! Just after you @kurtbrose also opened issue #100 with same stuff you mentioned plus some other things, so I thought maybe let him do the PR since it is a superset of yours and then you also check out the PR and help?

[Martinsos]

This got implemented in the meantime, thanks for help!