- Required data is extracted from MMCifInfoBioUnit object and fed into a
  custom BUInfo object
- BUInfo object can be serialized into bytestring
- Chain names in biounit have following naming scheme:
  <idx>.<asu_chain_name>. idx -> n-th copy of ASU chain with n
  starting from 1.
- Properties are copied from whatever is stored in assymetric unit
- Chain properties copied: ChainType (CHAINTYPE_POLY etc.)
- Residue properties copied: OLC, sec struct, ChemClass (PEPTIDE_LINKING etc.),
  ChemType(AMINOACIDS etc.), IsProtein flag, IsLigand flag
- Atom properties copied: name, element, occupancy, bfactor, ishetatm

The intention of this breaking change is to strictly concentrate
on coordinates. It makes limited sense to provide biounit support
without also providing SEQRES etc. So this should all go to some
external annotation system.

enable contact scores and separate dockq and per-interface qs scores

- Use bond/angle parameter of PeptideBuilder
  (https://github.com/clauswilke/PeptideBuilder/)
  RMSDs get lower as compared to the previously used parameters
  that were extracted from Charmm36 IC data.
- Introduce internal coordinates to reprent backbone positions.
  Problem with peptide chains is error propagation. An error
  introduced at one backbone dihedral propagates along the whole
  subsequent chain. We thus compress short fragments (tri-peptides)
  that are independent of each other.

Test behavior when no results and of invalid "by"

This replaces the awkward FindCompound(... by=) which would return
a single arbitrary compound.

This allows preferentially picking a non-obsolete compound
in FindCompound by="smiles/inchi"

Move long description to separate properties

Computes Interface Contact Similarity (ICS) score which is relevant
for CASP.

- Results slightly differ from predictioncenter website, need
  to check with Andriy where this comes from
- Should be a good basis to also add Interface Patch Similarity (IPS)
  score.

This allows IsValid() to work as intended without throwing an exception.
Because the validity of the underlying entity is now checked in CheckValidity,
it should no longer be a problem.

In both cases we check if we have a valid implementation AND if the
underlying entity is still there.