Bad handling of fragments in CIP code

[greglandrum]

Describe the bug

The code tries to kekulize the molecule before doing its work and fails if the kekulization fails.
This is perfectly sensible for most molecules, but can cause problems if trying to do CIP assignment on fragments:

In [44]: m2  = Chem.MolFromSmarts('F[C@H](CNC)CCc(:c):c')

In [45]: m2.UpdatePropertyCache()

In [46]: Chem.AssignCIPLabels(m2)
[17:27:52] non-ring atom 7 marked aromatic
---------------------------------------------------------------------------
AtomKekulizeException                     Traceback (most recent call last)
<ipython-input-46-7ee1ccbf16f8> in <module>
----> 1 Chem.AssignCIPLabels(m2)

AtomKekulizeException: non-ring atom 7 marked aromatic

In this case, the CIP assignment is clearly defined and we should be able to generate it.

Things get more interesting with a molecule like this:

In [42]: m2  = Chem.MolFromSmarts('F[C@](Cl)c(:c):c')

In [43]: Chem.AssignCIPLabels(m2)
[16:36:35] non-ring atom 3 marked aromatic
---------------------------------------------------------------------------
AtomKekulizeException                     Traceback (most recent call last)
<ipython-input-43-7ee1ccbf16f8> in <module>
----> 1 Chem.AssignCIPLabels(m2)

AtomKekulizeException: non-ring atom 3 marked aromatic

Here we'll have to decide how to treat the aromatic bonds (since the CIP code only supports integral bond orders).

Configuration (please complete the following information):

• RDKit version: master and current release
• OS: all