1.1.0 - 2022.08.23

Changed

• moldrug.utils.Individual now is a hashable object.

• moldrug.utils.GA.SawIndividuals now is a set instead of a list

• moldrug.utils.update_reactant_zone sets the keywords matchValences and ringMatchesRingOnly to True on rdFMCS.FindMCS. This prevent undesired effects. E.g:

  from moldrug.utils import update_reactant_zone
  from rdkit import Chem
  mol1 = Chem.MolFromSmiles('c1ccccc1')
  mol2 = Chem.MolFromSmiles('CCCN')
  update_reactant_zone(parent=mol1,offspring=mol2 parent_replace_ids = [0,2])

  Before the results was: ([0, 2, 3], []). Now it is: ([0, 1, 2, 3], []). The behavior was because the ringMatchesRingOnly is set to False by default inside RDKit.

Removed

• moldrug.utils.timeit. No longer needed.

Added

• Documentation.

1.0.2 - 2022.08.08

Removed

• Popen option in moldrug.utils.run.

Changed

• RuntimeError by warnings.warn when vina run fails and save every error as idx_error.pbz2. Where idx is the index of the failed individual.
• Print format when mutate fails inside of moldrug.utils.GA

Added

• Print MolDrug's version when the command line is used.

1.0.0 - 2022.07.30

Fixed

• Hidden RuntimeError in moldrug.fitness module.
• Bug printing number of generations in final info.

Removed

• Unused code in moldrug.home module.
• 3D conformation in the mol attribute of the Individual during initialization.
• The use of grow_mol in the initialization of the populating when get_similar = True. Now the population is initialized with mutate_mol and the same set of crem parameters used during the searching.
• The automatic addition of Hs in the case where min_size and/or max_size were equal to zero. Now if your intention is work with the hydrogens, you must provided a SMILES with the explicit Hs. In the future the input could be just a RDKit mol. Now you must specify if you would like to add explicit Hs to the molecule withe keyword AddHs; default is False and is used for both moldrug.utils.GA and moldrug.utils.Local.

Added

• Handling vina RuntimeError and keeping track for debug. This feature is used to identify what is the error. In the future will be removed.
• Two new fitness functions: moldrug.fitness.CostOnlyVina and moldrug.fitness.CostMultiReceptorsOnlyVina. They only use the information of vina scoring function. See the docs for more info about it.
• Tracking of atom indexes during generations in order to use protected_ids and replace_ids options of mutate_mol function of CReM. Before it was not possible; the use of these features generate undesired solutions because the indexes are not static over generations. Even so, there are still some problems for symmetric molecules. We are working on it.

Changed

• The whole MolDrug works base on the RDKit mol instead of the SMILES string:
  1. moldrug.utils.Individual is now initialized mol instead smiles. Now the SMILES string is generated internally, it still used as identifying for the instance.
  2. moldrug.utils.GA changed smiles for seed_mol and it is not needed the mol variable any more.
  3. moldrug.utils.Local changed mol for seed_mol in the initialization variables..
  4. moldrug.utils.confgen changed smiles for mol variable.

0.1.0 - 2022.07.25

Fixed

• Minor cleaning code.
• Bug during the import of user custom cost function.

Added

• outdir option for the command line.
• User custom desirability.

0.0.4

1. Fix minor compatibility issue with Python 3.8 (issue #4).
2. Add Local class to the command line.
3. Fix supplied custom fitness function from the command line.
4. Better code covering during testing.
5. Cleaning code.

0.0.3

1. Fix issue with molecules that does not produce a correct pdbqt file

0.0.2

1. Follow jobs on the command line
2. Fixing make_sdf for multiple receptors
3. Much better documentation

0.0.1.beta.9

1. Comand line available
2. Adding sdf export
3. Eliminate OpenBabel dependency
4. Adding meeko for working with pdbqt

0.0.1.beta7

Here we add the automatic detection based on mutate_crem_kwargs on GA of the need of add explicit hydrogens inside of the mutate function

0.0.1.beta6

change version name