Improving xml2png.py

[MehtapIsik]

I was testing xml2png.py with Infinite results of different assay types. I will list here some suggestions:

1. Adding usage example with type argument will be useful here:
   https://github.com/choderalab/assaytools/blob/master/scripts/xml2png.py#L4
2. For single_96 type, the plots are too wide. We should adjust plot dimensions.
   There is empty space on x-axis for wells bigger than 12 on 96-well plot.
3. I think adding axis labels will be great. x-axis label: Well Index, y-axis label: Fluorescence
4. Plot title usually has None printed for 'Part of Plate' section.
   
   https://github.com/choderalab/assaytools/blob/master/scripts/xml2png.py#L111-L124
   Do we need "Part of Plate" information? Or can we assign a default "Full Plate" when xml doesn't have this information.
5. I don't have ant 384-well data, so I couldn't test singlet_384 type but I noticed it has hard coded ligand concentration(Lstated) for plotting. Can we remove that?
   https://github.com/choderalab/assaytools/blob/master/scripts/xml2png.py#L404-L418
6. The difference between types spectra and scan wasn't obvious to me.

[sonyahanson]

Thanks Mehtap!

[bas-rustenburg]

Could it also be made python 3 compatible?

[bas-rustenburg]

It may also be helpful if it prints the help string if called without any arguments.

[jchodera]

All good ideas!