sheepforce / spicy Goto Github PK

Molecules need to be saved to files. By Text implement writers for:

• Molden XYZ
• Tinker XYZ
• Sybyl MOL2
• PDB
• Internal representation

We need parsers for the following file formats to the internal representation:

• Molden XYZ
• Tinker XYZ
• Sybyl MOL2
• PDB
• Internal representation

We need wrappers to the major quantum chemistry and MM codes. A YAML based input format should enable carrying out the majority of calculations with high level of abstraction but still the possibility to access lower level settings of underlying programs.

• Define a "calculation type" data structure
• Define a YAML based input format
• Define YAML based format to abstract capabilities of underlying programs

Implement fragment detection to split molecules into subunits:

• Distance matrix
• Bond Bool matrix
• Fragment merging

For exotic residues, such as the vanadium atom in 4ndg, which have more than 4 bonds, the PDB specification states, that the CONECT record for a single atom might occur more than once, if it has more than 4 bonds. This is currently not handled by parsePDB and the resulting bonds can therefore sometimes be not strictly bidirectorial. The consequence is, that a Molecule created by parsePDB can be malformed and is not passing checkMolecule.

We need wrappers, that take a "calculation type data structure" and writes input to QC software, runs them, and parses their results.

• ORCA
• Psi4
• Tinker
• MRCC

Implement internal coordinates for a single layer of Molecule.

checkMolecule is not checking if bonds are correctly bidirectorial. If a key from IntMap maps to values in an IntSet, the values in the IntSet must also be keys, mapping to the original key. This check must be implemented.

checkMolecule is also not checking if the coordinate sequence of all atoms is strictly of length 3.

I noticed that the current .xyz-parser (parseXYZ) fails to parse the molecule, as long as there is nothing on the second line (usually used for comments).

I suppose this should not crash in the first place?
As a first measure, I added a water molecule to the goldentest folder (H2O_broken.xyz) for testing this.