Convert the input and output data into integrated datasets in `pd.DataFrame` and `xr.Dataset`

[tztsai]

The intention is to convert input and output data files into a couple of integrated files with a more structured data format, like pandas.DataFrame or xarray.Dataset, replacing the current multitude of txt files.

• [x] Combine the txt files produced by the ML task into a single DataFrame and store it into a CSV file.
• [x] Combine packdata and auxil into a single xarray Dataset and store it into a NetCDF file.
• [x] Update the code in the clustering and ML steps in accordance with the update to the data formats.
• [ ] Consider adding a test to check equivalence between auxil.npy and packdata.npy with packdata.nc.

[ma595]

This all looks good. I've cherry-picked the relevant commits to a new branch and force pushed. You will need to pull the changes again (probably best to do this to a new branch, as I have ignored the last commit for now (CNP2 check)).

I've also done a first pass of a review (still need to check Cluster.py and ML.py). I've also checked that there is correspondence in the outputs of packdata.npy and auxil.npy to packdata.nc and all looks good so far. I found that approximately 70% of the data are nans. I've written some code to do these checks which could be integrated into the testing PR. I will do a proper review tomorrow but I suspect it'll be mostly on style and the addition of comments.