store data in objects rather than arrays?

So basically, I think it might make the interface a bit less janky if we had reads stored in an object with various attributes instead of a single array. The observation which really motivated me to consider this was the weird requirement we have to provide a chrom to write_wig. (ideally this information would just be collected during the call to `read_wig` in the first place, and then passed to `write_wig`, and the user wouldn't need to capture and pass this information....)

But one could imagine that there would be other benefits as well - for example calculating the z_scores/peaks/steps could now be done by adding a new attribute, or updating an attribute of reads.

Could also then build in an option to convert this data object into a `pd.DataFrame` object which could help downstream analysis. :)

Anyway - this is a potentially large refactor, so would love to hear if people had thoughts/suggestions/concerns before I embark :)