def main():
if not oefastrocs.OEFastROCSIsGPUReady():
oechem.OEThrow.Info("No supported GPU available!")
return 0
args = getargs()
dbname = args.d
# read in database
ifs = oechem.oemolistream()
if not ifs.open(dbname):
oechem.OEThrow.Fatal("Unable to open '%s'" % dbname)
print("Opening database file %s ..." % dbname)
timer = oechem.OEWallTimer()
opts = oefastrocs.OEShapeDatabaseOptions()
opts.SetLimit(1)
dbase = oefastrocs.OEShapeDatabase()
moldb = oechem.OEMolDatabase()
if not moldb.Open(ifs):
oechem.OEThrow.Fatal("Unable to open '%s'" % dbname)
dots = oechem.OEThreadedDots(10000, 200, "conformers")
if not dbase.Open(moldb, dots):
oechem.OEThrow.Fatal("Unable to initialize OEShapeDatabase on '%s'" %
dbname)
dots.Total()
print("%f seconds to load database" % timer.Elapsed())
df = pd.read_csv(args.i)
res = []
for smile in tqdm(df.loc[:, 'smiles'].tolist()):
resn = len(res)
try:
q = FromString(smile)[0]
for score in dbase.GetSortedScores(q, 1):
res.append(score.GetTanimotoCombo())
break
except KeyboardInterrupt:
print("caught")
exit()
except:
res.append(np.nan)
if len(res) == resn:
res.append(np.nan)
df['fastroc'] = res
print(df.head)
df.to_csv(args.o, sep=',', index=False)
return 0
def __init__(self, dbname, cutoff, shapeOnly):
self.cutoff = cutoff
# set up and options and database based upon shapeOnly
self.defaultOptions = oefastrocs.OEShapeDatabaseOptions()
dbtype = oefastrocs.OEShapeDatabaseType_Default
if shapeOnly:
dbtype = oefastrocs.OEShapeDatabaseType_Shape
self.defaultOptions.SetScoreType(dbtype)
self.shapedb = oefastrocs.OEShapeDatabase(dbtype)
self.dbmols = []
volumes = []
# read in database
ifs = oechem.oemolistream()
if not ifs.open(dbname):
oechem.OEThrow.Fatal("Unable to open '%s'" % dbname)
count = 0
for mol in ifs.GetOEGraphMols():
title = mol.GetTitle()
if not title:
title = "Untitled" + str(count)
mol.SetTitle(title)
count += 1
idx = self.shapedb.AddMol(oechem.OEMol(mol))
volume = oeshape.OEGetCachedSelfShape(mol)
if volume == 0.0:
volume = oeshape.OESelfShape(mol)
volumes.append((volume, idx))
dbmol = oechem.OEGraphMol(mol, oechem.OEMolBaseType_OEDBMol)
dbmol.Compress()
self.dbmols.append(dbmol)
numMols = len(volumes)
# find the molecule with the median volume as our first query
volumes.sort()
medianVolume, medianIdx = volumes[numMols // 2]
self.nextClusterHeadIdx = medianIdx
self.remainingMolecules = numMols
self.tanimotos = [0.0] * numMols
self.scoreGetter = GetScoreGetter(shapeOnly)
def __init__(self, shapedb, querymolstr, nhits, iformat, oformat,
errorLevel, **kwargs):
""" Create a new thread to perform a query. The query doesn't
execute until start is called.
shapedb - database to run the query against
See MCMolShapeDatabase.GetBestOverlays for a description of
the querymolstr and nhits arguments.
"""
Thread.__init__(self)
self.shapeOnly = kwargs.pop('shapeOnly', False)
self.tversky = kwargs.pop('tversky', False)
self.altStarts = kwargs.pop('altStarts', False)
self.randStarts = kwargs.pop('randStarts', False)
self.shapedb = shapedb
self.querymolstr = querymolstr
self.iformat = iformat
self.oformat = oformat
self.scoretype = GetDatabaseType(self.shapeOnly)
self.simFuncType = GetSimFuncType(self.tversky)
numHistBins = 200
if self.shapeOnly:
numHistBins = 100
self.tracer = oefastrocs.OEDBTracer(numHistBins)
self.options = oefastrocs.OEShapeDatabaseOptions()
self.options.SetTracer(self.tracer)
self.options.SetLimit(nhits)
self.options.SetScoreType(self.scoretype)
self.options.SetSimFunc(self.simFuncType)
if self.altStarts:
self.options.SetInitialOrientation(GetStartType(self.altStarts))
if self.randStarts:
self.options.SetNumRandomStarts(self.randStarts)
self.lock = Lock()
self.errorLevel = errorLevel
def GetCluster(self, query):
options = oefastrocs.OEShapeDatabaseOptions(self.defaultOptions)
dots = oechem.OEDots(10000, 200, "molecules searched")
minTani = sys.float_info.max
minIdx = None
for score in self.shapedb.GetScores(query, options):
idx = score.GetMolIdx()
# check if already in a cluster
if self.dbmols[idx] is None:
continue
if self.cutoff < self.scoreGetter(score):
yield self._removeMolecule(idx), score
else:
self.tanimotos[idx] = max(self.tanimotos[idx], self.scoreGetter(score))
minTani, minIdx = min((minTani, minIdx), (self.tanimotos[idx], idx))
dots.Update()
dots.Total()
self.nextClusterHeadIdx = minIdx
def main(argv=[__name__]):
itf = oechem.OEInterface(InterfaceData, argv)
ifs = oechem.oemolistream()
if not ifs.open(itf.GetString("-dbase")):
oechem.OEThrow.Fatal("Unable to open %s for reading" %
itf.GetString("-dbase"))
colname = "TanimotoCombo"
getter = oefastrocs.OEShapeDatabaseScore.GetTanimotoCombo
dbtype = oefastrocs.OEShapeDatabaseType_Default
if itf.GetBool("-shapeOnly"):
colname = "ShapeTanimoto"
getter = oefastrocs.OEShapeDatabaseScore.GetShapeTanimoto
dbtype = oefastrocs.OEShapeDatabaseType_Shape
csvwriter = csv.writer(open(itf.GetString("-matrix"), 'w'))
csvwriter.writerow(["Title1", "Title2", colname])
shapedb = oefastrocs.OEShapeDatabase(dbtype)
options = oefastrocs.OEShapeDatabaseOptions()
options.SetScoreType(dbtype)
lmat = [[]]
titles = []
for mol in ifs.GetOEMols():
if titles:
bestscores = [0.0] * len(titles)
for conf in mol.GetConfs():
for score in shapedb.GetScores(conf, options):
midx = score.GetMolIdx()
bestscores[midx] = max(bestscores[midx], getter(score))
lmat.append(bestscores)
shapedb.AddMol(mol)
title = mol.GetTitle()
if not title:
title = str(len(titles) + 1)
titles.append(title)
# write csv file
csvwriter = csv.writer(open(itf.GetString("-matrix"), 'w'))
csvwriter.writerow(titles)
nrows = len(titles)
for i in range(nrows):
row = [i + 1]
for j in range(nrows):
val = 2.0
if itf.GetBool("-shapeOnly"):
val = 1.0
if j > i:
val -= lmat[j][i]
elif j < i:
val -= lmat[i][j]
elif j == i:
val = 0.0
row.append("%.3f" % val)
csvwriter.writerow(row)
return 0
def main(argv=[__name__]):
parser = argparse.ArgumentParser()
# positional arguments retaining backward compatibility
parser.add_argument(
'database',
help='File containing the database molecules to be search \
(format not restricted to *.oeb).')
parser.add_argument(
'query',
default=[],
nargs='+',
help='File containing the query molecule(s) to be search \
(format not restricted to *.oeb).')
parser.add_argument(
'--nHits',
dest='nHits',
type=int,
default=100,
help='Number of hits to return (default = number of database mols).')
parser.add_argument('--cutoff',
dest='cutoff',
type=float,
default=argparse.SUPPRESS,
help='Specify a cutoff criteria for scores.')
parser.add_argument(
'--tversky',
dest='tversky',
action='store_true',
default=argparse.SUPPRESS,
help='Switch to Tversky similarity scoring (default = Tanimoto).')
args = parser.parse_args()
dbname = args.database
if not oefastrocs.OEFastROCSIsGPUReady():
oechem.OEThrow.Info("No supported GPU available!")
return 0
# set options
opts = oefastrocs.OEShapeDatabaseOptions()
opts.SetLimit(args.nHits)
print("Number of hits set to %u" % opts.GetLimit())
if hasattr(args, 'cutoff') is not False:
opts.SetCutoff(args.cutoff)
print("Cutoff set to %f" % args.cutoff)
if hasattr(args, 'tversky') is not False:
opts.SetSimFunc(args.tversky)
print("Tversky similarity scoring set.")
# read in database
ifs = oechem.oemolistream()
if not ifs.open(dbname):
oechem.OEThrow.Fatal("Unable to open '%s'" % dbname)
print("\nOpening database file %s ..." % dbname)
timer = oechem.OEWallTimer()
dbase = oefastrocs.OEShapeDatabase()
moldb = oechem.OEMolDatabase()
if not moldb.Open(ifs):
oechem.OEThrow.Fatal("Unable to open '%s'" % dbname)
dots = oechem.OEThreadedDots(10000, 200, "conformers")
if not dbase.Open(moldb, dots):
oechem.OEThrow.Fatal("Unable to initialize OEShapeDatabase on '%s'" %
dbname)
dots.Total()
print("%f seconds to load database\n" % timer.Elapsed())
for qfname in args.query:
# read in query
qfs = oechem.oemolistream()
if not qfs.open(qfname):
oechem.OEThrow.Fatal("Unable to open '%s'" % qfname)
mcmol = oechem.OEMol()
if not oechem.OEReadMolecule(qfs, mcmol):
oechem.OEThrow.Fatal("Unable to read query from '%s'" % qfname)
qfs.rewind()
ext = oechem.OEGetFileExtension(qfname)
qmolidx = 0
while oechem.OEReadMolecule(qfs, mcmol):
# write out to file name based on molecule title
ofs = oechem.oemolostream()
moltitle = mcmol.GetTitle()
if len(moltitle) == 0:
moltitle = str(qmolidx)
ofname = moltitle + "_results." + ext
if not ofs.open(ofname):
oechem.OEThrow.Fatal("Unable to open '%s'" % argv[4])
print("Searching for %s of %s (%s conformers)" %
(moltitle, qfname, mcmol.NumConfs()))
qconfidx = 0
for conf in mcmol.GetConfs():
for score in dbase.GetSortedScores(conf, opts):
dbmol = oechem.OEMol()
dbmolidx = score.GetMolIdx()
if not moldb.GetMolecule(dbmol, dbmolidx):
print(
"Unable to retrieve molecule '%u' from the database"
% dbmolidx)
continue
mol = oechem.OEGraphMol(
dbmol.GetConf(oechem.OEHasConfIdx(score.GetConfIdx())))
oechem.OESetSDData(mol, "QueryConfidx", "%s" % qconfidx)
oechem.OESetSDData(mol, "ShapeTanimoto",
"%.4f" % score.GetShapeTanimoto())
oechem.OESetSDData(mol, "ColorTanimoto",
"%.4f" % score.GetColorTanimoto())
oechem.OESetSDData(mol, "TanimotoCombo",
"%.4f" % score.GetTanimotoCombo())
score.Transform(mol)
oechem.OEWriteMolecule(ofs, mol)
qconfidx += 1
print("%s conformers processed" % qconfidx)
print("Wrote results to %s\n" % ofname)
qmolidx += 1
return 0
def main(argv=[__name__]):
if len(argv) < 3:
oechem.OEThrow.Usage("%s <database> [<queries> ... ]" % argv[0])
return 0
# check system
if not oefastrocs.OEFastROCSIsGPUReady():
oechem.OEThrow.Info("No supported GPU available!")
return 0
# read in database
dbname = argv[1]
print("Opening database file %s ..." % dbname)
dbase = oefastrocs.OEShapeDatabase()
moldb = oechem.OEMolDatabase()
if not moldb.Open(dbname):
oechem.OEThrow.Fatal("Unable to open '%s'" % dbname)
dots = oechem.OEThreadedDots(10000, 200, "conformers")
if not dbase.Open(moldb, dots):
oechem.OEThrow.Fatal("Unable to initialize OEShapeDatabase on '%s'" %
dbname)
# customize search options
opts = oefastrocs.OEShapeDatabaseOptions()
opts.SetLimit(5)
for qfname in argv[2:]:
# read in query
qfs = oechem.oemolistream()
if not qfs.open(qfname):
oechem.OEThrow.Fatal("Unable to open '%s'" % qfname)
query = oechem.OEGraphMol()
if not oechem.OEReadMolecule(qfs, query):
oechem.OEThrow.Fatal("Unable to read query from '%s'" % qfname)
ext = oechem.OEGetFileExtension(qfname)
base = qfname[:-(len(ext) + 1)]
# write out everthing to a similary named file
ofs = oechem.oemolostream()
ofname = base + "_results." + ext
if not ofs.open(ofname):
oechem.OEThrow.Fatal("Unable to open '%s'" % argv[4])
oechem.OEWriteMolecule(ofs, query)
print("Searching for %s" % qfname)
for score in dbase.GetSortedScores(query, opts):
print("Score for mol %u(conf %u) %f shape %f color" %
(score.GetMolIdx(), score.GetConfIdx(),
score.GetShapeTanimoto(), score.GetColorTanimoto()))
dbmol = oechem.OEMol()
molidx = score.GetMolIdx()
if not moldb.GetMolecule(dbmol, molidx):
print("Unable to retrieve molecule '%u' from the database" %
molidx)
continue
mol = oechem.OEGraphMol(
dbmol.GetConf(oechem.OEHasConfIdx(score.GetConfIdx())))
oechem.OESetSDData(mol, "ShapeTanimoto",
"%.4f" % score.GetShapeTanimoto())
oechem.OESetSDData(mol, "ColorTanimoto",
"%.4f" % score.GetColorTanimoto())
oechem.OESetSDData(mol, "TanimotoCombo",
"%.4f" % score.GetTanimotoCombo())
score.Transform(mol)
oechem.OEWriteMolecule(ofs, mol)
print("Wrote results to %s" % ofname)
return 0
def main(argv=[__name__]):
if len(argv) < 3:
oechem.OEThrow.Usage("%s <database> [<queries> ... ]" % argv[0])
return 0
# check system
if not oefastrocs.OEFastROCSIsGPUReady():
oechem.OEThrow.Info("No supported GPU available!")
return 0
# read in database
dbname = argv[1]
print("Opening database file %s ..." % dbname)
dbase = oefastrocs.OEShapeDatabase()
moldb = oechem.OEMolDatabase()
if not moldb.Open(dbname):
oechem.OEThrow.Fatal("Unable to open '%s'" % dbname)
dots = oechem.OEThreadedDots(10000, 200, "conformers")
if not dbase.Open(moldb, dots):
oechem.OEThrow.Fatal("Unable to initialize OEShapeDatabase on '%s'" %
dbname)
# customize search options
opts = oefastrocs.OEShapeDatabaseOptions()
opts.SetInitialOrientation(
oefastrocs.OEFastROCSOrientation_UserInertialStarts)
opts.SetLimit(5)
for qfname in argv[2:]:
# read in query
qfs = oechem.oemolistream()
if not qfs.open(qfname):
oechem.OEThrow.Fatal("Unable to open '%s'" % qfname)
query = oechem.OEGraphMol()
if not oechem.OEReadMolecule(qfs, query):
oechem.OEThrow.Fatal("Unable to read query from '%s'" % qfname)
ext = oechem.OEGetFileExtension(qfname)
base = qfname[:-(len(ext) + 1)]
# write out everthing to a similary named file
ofs = oechem.oemolostream()
ofname = base + "_user_results." + ext
if not ofs.open(ofname):
oechem.OEThrow.Fatal("Unable to open '%s'" % argv[4])
oechem.OEWriteMolecule(ofs, query)
startsCoords = oechem.OEFloatVector()
atomIdx = 1
xyz = query.GetCoords()[atomIdx]
for x in xyz:
startsCoords.append(x)
if not len(startsCoords) % 3 == 0:
oechem.OEThrow.Fatal(
"Something went wrong whilst reading in user-starts coordinates"
)
opts.SetUserStarts(oechem.OEFloatVector(startsCoords),
int(len(startsCoords) / 3))
opts.SetMaxOverlays(opts.GetNumInertialStarts() *
opts.GetNumUserStarts())
if opts.GetInitialOrientation(
) == oefastrocs.OEFastROCSOrientation_UserInertialStarts:
numStarts = opts.GetNumUserStarts()
print("This example will use %u starts" % numStarts)
print("Searching for %s" % qfname)
for score in dbase.GetSortedScores(query, opts):
print("Score for mol %u(conf %u) %f shape %f color" %
(score.GetMolIdx(), score.GetConfIdx(),
score.GetShapeTanimoto(), score.GetColorTanimoto()))
dbmol = oechem.OEMol()
molidx = score.GetMolIdx()
if not moldb.GetMolecule(dbmol, molidx):
print("Unable to retrieve molecule '%u' from the database" %
molidx)
continue
mol = oechem.OEGraphMol(
dbmol.GetConf(oechem.OEHasConfIdx(score.GetConfIdx())))
oechem.OESetSDData(mol, "ShapeTanimoto",
"%.4f" % score.GetShapeTanimoto())
oechem.OESetSDData(mol, "ColorTanimoto",
"%.4f" % score.GetColorTanimoto())
oechem.OESetSDData(mol, "TanimotoCombo",
"%.4f" % score.GetTanimotoCombo())
score.Transform(mol)
oechem.OEWriteMolecule(ofs, mol)
print("Wrote results to %s" % ofname)
return 0
def main(argv=[__name__]):
if len(argv) < 3:
oechem.OEThrow.Usage("%s <database> [<queries> ... ]" % argv[0])
dbname = argv[1]
# read in database
ifs = oechem.oemolistream()
if not ifs.open(dbname):
oechem.OEThrow.Fatal("Unable to open '%s'" % dbname)
print("Opening database file %s ..." % dbname)
timer = oechem.OEWallTimer()
dbase = oefastrocs.OEShapeDatabase()
moldb = oechem.OEMolDatabase()
if not moldb.Open(ifs):
oechem.OEThrow.Fatal("Unable to open '%s'" % dbname)
dots = oechem.OEThreadedDots(10000, 200, "conformers")
if not dbase.Open(moldb, dots):
oechem.OEThrow.Fatal("Unable to initialize OEShapeDatabase on '%s'" %
dbname)
dots.Total()
print("%s seconds to load database" % timer.Elapsed())
opts = oefastrocs.OEShapeDatabaseOptions()
opts.SetSimFunc(oefastrocs.OEShapeSimFuncType_Tversky)
numHits = moldb.NumMols()
opts.SetLimit(numHits)
for qfname in argv[2:]:
# read in query
qfs = oechem.oemolistream()
if not qfs.open(qfname):
oechem.OEThrow.Fatal("Unable to open '%s'" % argv[1])
query = oechem.OEGraphMol()
if not oechem.OEReadMolecule(qfs, query):
oechem.OEThrow.Fatal("Unable to read query from '%s'" % argv[1])
ext = oechem.OEGetFileExtension(qfname)
base = qfname[:-(len(ext) + 1)]
# write out everthing to a similary named file
ofs = oechem.oemolostream()
ofname = base + "_results." + ext
if not ofs.open(ofname):
oechem.OEThrow.Fatal("Unable to open '%s'" % argv[4])
print("Searching for %s" % qfname)
for score in dbase.GetSortedScores(query, opts):
dbmol = oechem.OEMol()
molidx = score.GetMolIdx()
if not moldb.GetMolecule(dbmol, molidx):
print("Unable to retrieve molecule '%u' from the database" %
molidx)
continue
mol = oechem.OEGraphMol(
dbmol.GetConf(oechem.OEHasConfIdx(score.GetConfIdx())))
oechem.OESetSDData(mol, "ShapeTversky",
"%.4f" % score.GetShapeTversky())
oechem.OESetSDData(mol, "ColorTversky",
"%.4f" % score.GetColorTversky())
oechem.OESetSDData(mol, "TverskyCombo",
"%.4f" % score.GetTverskyCombo())
score.Transform(mol)
oechem.OEWriteMolecule(ofs, mol)
print("Wrote results to %s" % ofname)
return 0