Feature/gridded data

[cparcerisas]

Changing the way some things are computed so we can have "gridded" data. This means that the seconds are at zero for all the bins, and that pypam will look for the next file in line to join the end and the beginning of the two files.
This is currently done only if there is a gap of less than 1 second between timestamp of the second file and end of the first one.

For this, I added a class acoustic_chunk so it can represent each bin.
In the future all the binsize and bin_overlap parameters should be changed to chunksize and chunk_overlap for clarity.

Still some things to be polished so code is more elegant and things are not repeated.

[cparcerisas]

@carueda Do you want to give it a go to testing if the gridded implementation works?
Right now it is reading the files in order (this might be a drawback already) and if files are less than 1 second apart in timestamps considering next file to be continuous.
This does not solve the "time-drifting" problem that @ryjombari wants to fix, but it's a start.