lich-uct / molpher-lib Goto Github PK

looks great guys!

Windows build would be well appreciated!

For the time being I suggest to mention this in the docs in the install section (On windows the conda command will say the package is not available which at first is confusing).

Again this library looks interesting. While playing around and reading tutorial I wasn't able to figure out how to run ExplorationTree with custom operations but using different than classic algorithm.

Hi,
I created new environment of conda and install molpher-lib via conda command.
But I got following error when I call the package from python.

ModuleNotFoundError: No module named '_core'

How can I fix the problem?
Any comments or advices are greatly appreciated.
Thanks,

Changes introduced after v0.0.0b1 are not yet properly documented. They are part of the tutorial code examples, but the source code of these features generally lacks documentation so far. This should be fixed.

Something like this should be added to the readme:

conda env create -f environment.yml
. source_2_activate
python setup.py build_ext --inplace
cd doc/notebooks/
jupyter-notebook

after the compilation instructions.

When running the classic algorithm example code:

import os
import sys

from molpher.algorithms.classic.run import run
from molpher.algorithms.settings import Settings


def main(args):
    # our source and target molecules
    cocaine = 'CN1[C@H]2CC[C@@H]1[C@@H](C(=O)OC)[C@@H](OC(=O)c1ccccc1)C2'
    procaine = 'O=C(OCCN(CC)CC)c1ccc(N)cc1'

    # path to a directory where results will be stored (read from command line if possible)
    storage_dir = None
    if len(args) == 2:
        storage_dir = os.path.abspath(args[1])
    else:
        storage_dir = 'classic_data'

    # initialize the exploration settings
    settings = Settings(
        cocaine
        , procaine
        , storage_dir
        , max_threads=4
        , max_iters = 10
    )

    run(settings)

if __name__ == "__main__":
    exit(main(sys.argv))

the max_iters = 10 parameter is honored, but a confusing message is raised:

Error message:
Traceback (most recent call last):
  File "small.py", line 39, in <module>
    main(smiles[x],smiles[y], storage_dir='classical_data_'+start_id[x]+'_'+target_id[y]+'_run_'+str(runnum))
  File "small.py", line 32, in main
    run(settings)
  File "/home/isabel/anaconda2/envs/molpher_pseudo_random/lib/python3.6/site-packages/molpher-0.0.0b2-py3.6-linux-x86_64.egg/molpher/algorithms/classic/run.py", line 16, in run
    pathfinder()
  File "/home/isabel/anaconda2/envs/molpher_pseudo_random/lib/python3.6/site-packages/molpher-0.0.0b2-py3.6-linux-x86_64.egg/molpher/algorithms/pathfinders.py", line 46, in __call__
    self.path = find_path(self.tree, self.tree.params['target'])
  File "/home/isabel/anaconda2/envs/molpher_pseudo_random/lib/python3.6/site-packages/molpher-0.0.0b2-py3.6-linux-x86_64.egg/molpher/algorithms/functions.py", line 41, in find_path
    current = tree.fetchMol(end_mol)
  File "/home/isabel/anaconda2/envs/molpher_pseudo_random/lib/python3.6/site-packages/molpher-0.0.0b2-py3.6-linux-x86_64.egg/molpher/core/ExplorationTree.py", line 233, in fetchMol
    ret = super(ExplorationTree, self).fetchMol(canonSMILES)
  File "/home/isabel/anaconda2/envs/molpher_pseudo_random/lib/python3.6/site-packages/molpher-0.0.0b2-py3.6-linux-x86_64.egg/molpher/swig_wrappers/core.py", line 1918, in fetchMol
    return _core.ExplorationTree_fetchMol(self, canonSMILES)
RuntimeError: Molecule (COC1=CC=C(F)C=C1C(C)(C)CC(O)(CC1=CC=CC=C1)C(=O)NC1=CC=C2C(=O)OCC2=C1) is not present in the tree.

It should give a clear message about what went wrong.

Hello,

do we have any updates on a Mac release?

Many thanks

It would be nice if users had the option to store some data with the instances of MolpherMol and MolpherAtom (like it is possible in RDKit or other libraries). It could work like so:

mol = MolpherMol("CCO")
mol.setProp("Name", "ethanol")
for idx, atom in enumerate(mol.atoms):
    atom.setProp("AtomID", mol.getProp("Name") + "." + str(idx))

It would be nice to have this line:

sudo apt-get install git build-essential python3-dev python3-numpy cmake python3-setuptools

at the relevant place here and in the readme file.

RDKit 2018.03.2 probably changes the names of the binary library files so the linking code needs to be changed to reflect that.

TBB 2018.0.4 could cause a SEGFAULT for a yet unknown reason in some instances.