Throw informative error for "chain name too long"

[phraenquex]

Error described in this slack thread.

I was trying to align some structures in lb32627-268 but I encountered this error: chain name too long for the pdb format: A10
Have you seen this before? Is there an error in one of my downloaded PDBs?

Files are available in that thread.

[tdudgeon]

As @mwinokan points out in slack, the main missing info seems to be the name of the offending crystal (pdb file).
Other than that the error message chain name too long for the pdb format: A10 is not too unreasonable.

But the problem here is:

1. the error is thrown from LNA so it's probably best for @ConorFWild to look at this first (I will add him to the assignees)
2. it's an uncaught error from gemmi, so is not going to be easy (e.g hacky) to handle it as a special case, might be better if it was handled by some earlier validation step, but again Conor is best placed to assess this
3. I recall some weirdness about handling of waters (re-writing them to a different chain) that had the inbuilt assumption that there would be no more than 9 chains, but again Conor is best placed to assess this

[mwinokan]

@tdudgeon please see if you can do any of this without @ConorFWild, maybe just adding the PDB code by catching gemmi error message

@mwinokan suggests a short-term fix to name the water chains using an alphanumeric rather than integer index, e.g. after A9 go to AA and so on

[tdudgeon]

PR containing what I see as the simplest (maybe not the best) way to report the offending structure.
ConorFWild/ligand_neighbourhood_alignment#4