Module MolExt

• nUnbondedElectrons
• nLonePairs
• nRadicals
• nEmptyOrbitals

OEChem has OEGetHybridization method, and this works similarly.

Return value is the p character of the hybridization, for example 2 for sp2

Simple rule based now, that should work for atoms up to the second row at least. Easy to tell if this atom has more than one pi bond. Otherwise, assume is sp3 unless have an empty orbital then assume is sp2. Or, if have a lone pair in an aromatic system, must be sp2. In cases like oxygen of C=CO, is probably somewhere in between.

For the given molecule or intermediate, enumerate all of the reasonable resonance structures by shifting electrons around available p atomic orbitals and pi molecular orbitals. For efficiency, will always be the same molecule object yielded, just modified. Caller must make sure that it comes back in the same state.

Strategy: Look for any 2 adjacent atoms hybridized with p or pi orbitals and rearrange electrons in a chemically natural manner.

Simple wrapper around resonance structureIter method.  Before generating
the resonance structures however, will give a unique atom map label to
every atom in the molecule with the effect that symmetric resonance structures
will each be counted individually, instead of being skipped over as if identical.

undoLabels: Specifies whether to clear out these fake labels before
                each result is yielded

Helper to label the atom map attribute of all of the atoms of the molecule equal to their index position in the molecule, which ensures they will all receive a unique value.

Given just the p and pi orbitals for the given bond,
enumerate each natural way of rearranging the electrons within them.
Don't have to consider mirror cases, leave that as caller's responsibility
by calling this function twice, but swapping the begin and end atoms and pOrbs.

Cases of interest (not counting radicals for now):
    - Both atoms share a pi molecular orbital.  That is, the bond includes a pi bond.
        . Separate the pi molecular orbitals into one empty orbital (+ charge) and 
            one lone pair orbital (- charge).
            Then caller can recurse on this intermediate which should fit the above rules.
    - One atom has lone pair, other has empty p orbital (artificial case representing charge separation modeling)
        . Move lone pair to empty p orbital to form new pi bond here
    - One atom has pi molecular orbital other has empty p orbital
        . Move the pi electrons to form a new pi bond here
    - One atom has a lone pair in p orbital, other has a pi molecular orbital
        . Move the lone pair into a new pi bond
        . Move the pi electrons to a lone pair over the far (3rd) atom

Additional case where atoms each have a pi molecular orbital, but not the same one,
like the central bond in C=C-C=C.  Do nothing here, let another case separate
the pi orbital of one of the adjacent double bonds, and then it will look like
a case we can work with (lone pair or empty orbital next to pi orbital).

Note that this keeps yielding the same (but modified) molecule object.
Thus, in implementing, after each yield usage, the atom and bond states must 
be returned to how they were.

The real implementation method. Interface method is just a wrapper on this one with a filter to eliminate overcharged resonance structure proposals.

Figure out the total gross positive and negative charge on the molecule, which may be non-zero even if the net charge on the molecule is zero.

The molecule should represent a proposed resonance structure.
Decide whether it should be accepted as a valid / meaningful one.

For example:
    Don't accept charge separated structures if it leaves an empty orbital 
    (incomplete octet) on the formal positive atom.
    It's okay if the formal positive atom has a complete octet, like
    the nitrogen on nitro and azide groups.
    Also okay if lone positive charge, like carbocation with no anion available
    to "fill-in" the empty orbital.

Most general half-reaction, movement of electrons from one molecular orbital to another. Function should figure out all of the substeps necessary to properly represent the change, including consideration for whether the orbitals are atomic (non-bond) or molecular (bond).

Assume this is called in a sensible manner however. Cannot use any arbitary orbitals as the arguments. The source orbital (orb1) must contain electrons to move, no empty orbitals allowed. The target orbital (orb2) must have space for electrons to move in. In general this means it is either an empty orbital or a bond orbital which can displace the existing bond electrons to the neighbor atom. No lone pairs allowed.

Move electrons from atom1 to atom2. If these are bonded already, just change the bond order and adjust the formal charges. If they are not already bonded, then create a new bond and adjust the formal charges.

If moving single electrons (radical reactions) better provide halfBonds list. Used to keep track of which bonds have only been "half-formed" by a radical movement, and are waiting for a complementary one to complete or eliminate the bond. Note that the bond could be "1 1/2" or "2 1/2." The point is that it's just not a whole number bond yet.

Move electrons from the bond to the target atom. If the atom is already part of the bond, then just down shift the bond order (maybe deleting the bond altogether) and adjust atom formal charges.

Otherwise, down shift this bond order and create a new one to the atom, again adjusting atom formal charges. Note that this only works if the "pivotAtom" is specified. That is, the atom which appears in both the source and the target bond. Otherwise, without the polarity of the source bond specified, can't tell which end to remain connected to.

Given 2 molecules, that should be based on a common one, determine how many atoms in the molecule have been altered by electron rearrangement. One of the molecules should probably have been a copy of the other, because this method depends on the mol.GetAtoms iterator returning the atoms in the same order.

• Organometallics do not have configurationally stable stereochemistry

• Vinylic carbocations should not have stereochemistry, should be linear, sp hybridized

Reparse the molecule SMILES string to ensure any unusual properties are cleared out. Happens often when doing direct molecule graph manipulations or reaction processing.

Assume the molecule object represents a racemic mixture of otherwise identical isomers.  
Go through every unassigned stereocenter and generate all possible specific isomers
by assiging all possible stereo configurations.

allCisTrans: If true, will enumerate all cis-trans isomers for unspecified chiral bonds.
    Otherwise, will only generate one with the bulkiest (largest by molecular weight)
    groups trans to each other

Given a list of molecule objects, call enumerateRacemicMixture on each to yield a list of all possible isomer combinations of the molecules.

• All implicit and explicit hydrogen atoms are removed
• All bond orders are set to 1
• All formal charges are neutralized
• All isotopes are reset

• All atom mapping indexes are reset

  If carbonSkeleton is True
  • All non-carbon atoms that are not within a ring or chain are removed

ORBITAL_TYPES

Value:

Set(["sigma", "pi", "s", "sp0", "sp", "sp1", "sp2", "sp3", "p"])