def getCCDefObj(self, dataContainer, molBuildType="model-xyz", suppressHydrogens=False):
"""Build OE molecule from the input chemical component definition object."""
#
oemf = OeMoleculeFactory()
if not self.__verbose:
oemf.setQuiet()
ccId = oemf.setChemCompDef(dataContainer)
oemf.build(molBuildType=molBuildType)
if self.__verbose:
logger.info(" CCId = %s", ccId)
logger.info(" Title = %s", oemf.getTitle())
logger.info(" SMILES = %s", oemf.getCanSMILES())
logger.info(" SMILES (stereo) = %s", oemf.getIsoSMILES())
logger.info(" Formula (Hill) = %s", oemf.getFormula())
logger.info(" InChI key = %s", oemf.getInChIKey())
logger.info(" InChI = %s", oemf.getInChI())
fD = {}
fD = {"Formula": oemf.getFormula(), "SMILES": oemf.getCanSMILES(), "SMILES_STEREO": oemf.getIsoSMILES(), "InChI": oemf.getInChI(), "InChIKey": oemf.getInChIKey()}
if suppressHydrogens:
tMol = oemf.getGraphMolSuppressH()
else:
tMol = oemf.getMol()
fD["OEMOL"] = tMol
fD["xyz"] = oemf.getAtomDetails(xyzType="model")
return (ccId, tMol, fD)
def getCCDefFile(self, ccFilePath, molBuildType="model-xyz", suppressHydrogens=False):
"""Fetch the molecule definition (ccPath) and build OE molecules
for comparison.
"""
#
mU = MarshalUtil(workPath=self.__workPath)
rdCcObjL = mU.doImport(ccFilePath, fmt="mmcif")
oemf = OeMoleculeFactory()
if not self.__verbose:
oemf.setQuiet()
ccId = oemf.setChemCompDef(rdCcObjL[0])
oemf.build(molBuildType=molBuildType)
if self.__verbose:
logger.info(" CCId = %s", ccId)
logger.info(" Title = %s", oemf.getTitle())
logger.info(" SMILES = %s", oemf.getCanSMILES())
logger.info(" SMILES (stereo) = %s", oemf.getIsoSMILES())
logger.info(" Formula (Hill) = %s", oemf.getFormula())
logger.info(" InChI key = %s", oemf.getInChIKey())
logger.info(" InChI = %s", oemf.getInChI())
fD = {}
fD = {"Formula": oemf.getFormula(), "SMILES": oemf.getCanSMILES(), "SMILES_STEREO": oemf.getIsoSMILES(), "InChI": oemf.getInChI(), "InChIKey": oemf.getInChIKey()}
if suppressHydrogens:
tMol = oemf.getGraphMolSuppressH()
else:
tMol = oemf.getMol()
fD["OEMOL"] = tMol
fD["xyz"] = oemf.getAtomDetails(xyzType="model")
return (ccId, tMol, fD)
def __testReproduceDescriptors(self, molBuildType, limitPerceptions=True):
#
ccMolD, ccIdxD = self.__getChemCompDefs()
oemf = OeMoleculeFactory()
countD = defaultdict(int)
for ccId, ccDef in ccMolD.items():
tId = oemf.setChemCompDef(ccDef)
if ccId != tId:
continue
oemf.build(molBuildType=molBuildType,
limitPerceptions=limitPerceptions)
oeMol = oemf.getMol()
#
countD["total components"] += 1
if ccId not in ccIdxD:
logger.info("Missing ccIndex entry for %s", ccId)
continue
ccdD = ccIdxD[ccId]
if ccdD["ambiguous"]:
countD["ambiguous component"] += 1
continue
#
countD["total molecules"] += 1
nativeCanIsoSmiles = oechem.OECreateIsoSmiString(oeMol)
canIsoSmiles = oechem.OEMolToSmiles(oeMol)
isoSmiles = oemf.getIsoSMILES()
canSmiles = oemf.getCanSMILES()
# check interal consistency
if nativeCanIsoSmiles != isoSmiles:
logger.error("%s stored and calculated OE smiles differ %s %s",
ccId, nativeCanIsoSmiles, isoSmiles)
if canIsoSmiles != isoSmiles:
logger.error(
"%s calculated OE ISO and canonical smiles differ %s %s",
ccId, isoSmiles, canIsoSmiles)
# compare with archived values
if isoSmiles != ccdD["oe-iso-smiles"]:
logger.info("%s ISO SMILES differ \nccd: %r \nOE: %r", ccId,
ccdD["oe-iso-smiles"], isoSmiles)
countD["iso_smiles_diff"] += 1
# ----------
if canSmiles != ccdD["oe-smiles"]:
logger.info("%s CAN SMILES differ \nccd: %r \nOE: %r", ccId,
ccdD["oe-smiles"], canSmiles)
countD["smiles_diff"] += 1
formula = oemf.getFormula()
if formula.upper() != ccdD["formula"].upper():
logger.debug("%s formulas differ \nccd: %r \nOE: %r", ccId,
ccdD["formula"], formula)
countD["formula_diff"] += 1
# ---------
inchiKey = oemf.getInChIKey()
if inchiKey != ccdD["inchikey"]:
logger.debug("%s InChI keys differ \nccd: %r \nOE: %r", ccId,
ccdD["inchikey"], inchiKey)
countD["inchikey_diff"] += 1
#
inchi = oemf.getInChI()
if inchi != ccdD["inchi"]:
logger.debug("%s InChIs differ \nccd: %r \nOE: %r", ccId,
ccdD["inchi"], inchi)
countD["inchi_diff"] += 1
#
#
for ky, vl in countD.items():
logger.info("%-12s %6d", ky, vl)
def testSssWithFingerPrintFromDescriptor(self):
oemp = OeMoleculeProvider(**self.__myKwargs)
ok = oemp.testCache()
ccmP = ChemCompIndexProvider(**self.__myKwargs)
ccIdxD = ccmP.getIndex()
ok = ccmP.testCache(minCount=self.__minCount)
self.assertTrue(ok)
limitPerceptions = False
# minFpScore = 0.5
maxFpResults = 50
matchOpts = "graph-relaxed"
numMols = 20
oeioU = OeIoUtils()
oesU = OeSearchUtils(oemp, fpTypeList=self.__fpTypeList)
missTupL = []
missedD = {}
missedFpD = {}
# ----
startTime = time.time()
for ccId, ccD in list(ccIdxD.items())[:numMols]:
for buildType in [
"oe-iso-smiles", "oe-smiles", "acdlabs-smiles",
"cactvs-iso-smiles", "cactvs-smiles", "inchi"
]:
if buildType in ccD:
logger.debug("Search %s %r", ccId, ccD[buildType])
if buildType in ["inchi"]:
oemf = OeMoleculeFactory()
oemf.setDescriptor(ccD["inchi"], "inchi", ccId)
ok = oemf.build(molBuildType="inchi",
limitPerceptions=limitPerceptions)
if not ok:
logger.info("%s build failed with InChI %r", ccId,
ccD["inchi"])
else:
oeMol = oemf.getMol()
if oemf.getInChI() != ccD["inchi"]:
logger.info(
"%s regenerated InChI differs\n%r\n%s",
ccId, ccD["inchi"], oemf.getInChI())
else:
oeMol = oeioU.smilesToMol(
ccD[buildType], limitPerceptions=limitPerceptions)
if not oeMol:
continue
maxHits = 0
minHits = maxFpResults
selfHit = False
for fpType, minFpScore in self.__fpTypeCuttoffList:
retStatus, mL = oesU.searchSubStructureWithFingerPrint(
oeMol,
fpType,
minFpScore,
maxFpResults,
matchOpts=matchOpts)
self.assertTrue(retStatus)
logger.debug("%s fpType %r hits %d", ccId, fpType,
len(mL))
maxHits = max(maxHits, len(mL))
minHits = min(minHits, len(mL))
matchedSelf = self.__resultContains(ccId, mL)
selfHit = selfHit or matchedSelf
if not matchedSelf:
missedFpD.setdefault(ccId, []).append(
(buildType, fpType, len(mL)))
if not selfHit:
missedD.setdefault(ccId, []).append(buildType)
logger.info("%s (%r) buildType %r min hits %d max hits %d",
ccId, selfHit, buildType, minHits, maxHits)
else:
logger.info("%s missing descriptor %r", ccId, buildType)
#
for ccId, missL in missedD.items():
logger.info("%s missed list %r", ccId, missL)
if ccId in missedFpD:
logger.info("%s unmatched for fpTypes %r", ccId,
missedFpD[ccId])
# ----
doDepict = False
if doDepict:
mD = {}
for missTup in missTupL:
mD.setdefault(missTup[0], []).append(missTup[1])
for ccId, buildTypeL in mD.items():
idxD = ccIdxD[ccId]
if "oe-iso-smiles" in idxD:
for buildType in buildTypeL:
self.__displayAlignedDescriptorPair(
ccId,
idxD["oe-iso-smiles"],
"oe-iso-smiles",
idxD[buildType],
buildType,
title=None,
limitPerceptions=True)
logger.info("%s fingerprints search on %d in (%.4f seconds)",
len(self.__fpTypeList), numMols,
time.time() - startTime)
def __getMiscFile(self, filePath, suppressHydrogens=False, importType="2D", title=None, largestPart=False):
"""Fetch a miscellaneous chemical file (ccPath) and build OE molecules
for comparison.
"""
try:
oeioU = OeIoUtils()
oeMolL = oeioU.fileToMols(filePath, use3D=importType == "3D", largestPart=largestPart)
logger.info("Read (%d) from %s ", len(oeMolL), filePath)
oeMol = oeMolL[0]
ccId = title if title else oeMol.GetTitle()
if title:
oeMol.SetTitle(ccId)
#
oemf = OeMoleculeFactory()
if not self.__verbose:
oemf.setQuiet()
oemf.setOeMol(oeMol, ccId)
#
fD = oemf.getOeMoleculeFeatures()
if self.__verbose:
logger.info(" Title = %s", title)
logger.info(" Title OEMF = %s", oemf.getTitle())
logger.info(" SMILES = %s", oemf.getCanSMILES())
logger.info(" SMILES (stereo) = %s", oemf.getIsoSMILES())
logger.info(" Formula (Hill) = %s", oemf.getFormula())
logger.info(" InChI key = %s", oemf.getInChIKey())
logger.info(" InChI = %s", oemf.getInChI())
#
ccId = oemf.getTitle()
if suppressHydrogens:
tMol = oemf.getGraphMolSuppressH()
else:
tMol = oemf.getMol()
molXyzL = []
if importType == "3D":
for atm in tMol.GetAtoms():
xyzL = oechem.OEFloatArray(3)
tMol.GetCoords(atm, xyzL)
molXyzL.append(
ComponentAtomDetails(
atIdx=atm.GetIdx(),
atNo=atm.GetAtomicNum(),
atName=atm.GetName(),
atType=atm.GetType(),
x=xyzL[0],
y=xyzL[1],
z=xyzL[2],
atFormalCharge=atm.GetFormalCharge(),
)
)
fD = {}
fD = {
"Formula": oemf.getFormula(),
"SMILES": oemf.getCanSMILES(),
"SMILES_STEREO": oemf.getIsoSMILES(),
"InChI": oemf.getInChI(),
"InChIKey": oemf.getInChIKey(),
"xyz": molXyzL,
}
for atm in tMol.GetAtoms():
xyzL = oechem.OEFloatArray(3)
tMol.GetCoords(atm, xyzL)
if self.__verbose:
logger.debug("atom %s %s %s %s %r", atm.GetIdx(), atm.GetAtomicNum(), atm.GetName(), atm.GetType(), xyzL)
fD["OEMOL"] = tMol
return (ccId, tMol, fD)
except Exception as e:
logger.exception("Failing with %s", str(e))
return None, None, None