Include encoding information in the header

[dalonsoa]

I think encoding is the sort of thing that you need to know upfront to read the file properly. Reading it from the header might be an option, but requires assuming things about the header, already and I think the premise of PyCSVY is not to make any assumption. As Alex suggest, if we want to consider it, it will be a separate PR for sure.

I do agree with you, but just to add that people do put the encoding in actual documents sometimes (e.g. you see it in HTML etc.). It's kinda weird, but it works fine if a) the encoding is ASCII-compatible (like cp1252 or UTF-8) and b) you don't use any non-ASCII chars before the bit where you say what encoding the document uses. Definitely a job for another day though, if indeed we bother at all. (Who wouldn't want to use UTF-8 these days anyway?)

Originally posted by @alexdewar in #124 (comment)

[dalonsoa]

Maybe as a validation option, if the user asks for UTF-8 and the header says a different one, raising an error, but I'd not use it as the main source of information about how to load the file.