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 testBuildCifFromOE(self):
"""Build chemical component definitions from OE Mol object"""
try:
ccMolD = self.__getChemCompDefs()
oemf = OeMoleculeFactory()
#
for ccId, ccObj in list(ccMolD.items())[:10]:
# ----
tId = oemf.setChemCompDef(ccObj)
self.assertEqual(tId, ccId)
ok = oemf.build(molBuildType="model-xyz")
self.assertTrue(ok)
fp = os.path.join(self.__ccCifPath, ccId + "-gen.cif")
oeMol = oemf.getMol()
oeccU = OeChemCompUtils()
ok = oeccU.addOeMol(ccId,
oeMol,
missingModelXyz=False,
writeIdealXyz=False)
self.assertTrue(ok)
ok = oeccU.write(fp)
# ----
except Exception as e:
logger.exception("Failing with %s", str(e))
self.fail()
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 __displayAlignedDescriptorPair(self,
ccId,
descrRef,
buildTypeRef,
descrFit,
buildTypeFit,
title=None,
limitPerceptions=True):
oemfRef = OeMoleculeFactory()
oemfRef.setDescriptor(descrRef, buildTypeRef, ccId)
oemfRef.build(molBuildType=buildTypeRef,
limitPerceptions=limitPerceptions)
oeMolRef = oemfRef.getMol()
#
oemfFit = OeMoleculeFactory()
oemfFit.setDescriptor(descrFit, buildTypeFit, ccId)
oemfFit.build(molBuildType=buildTypeFit,
limitPerceptions=limitPerceptions)
oeMolFit = oemfFit.getMol()
#
oed = OeDepictMCSAlignPage()
oed.setSearchType(sType="graph-relaxed", minAtomMatchFraction=0.50)
oed.setDisplayOptions(labelAtomName=True,
labelAtomCIPStereo=True,
labelAtomIndex=False,
labelBondIndex=False,
highlightStyleFit="ballAndStickInverse",
bondDisplayWidth=0.5)
oed.setRefMol(oeMolRef, ccId)
oed.setFitMol(oeMolFit, ccId)
myTitle = title if title else buildTypeRef + "-" + buildTypeFit
imgPath = os.path.join(self.__workPath, myTitle + "-" + ccId + ".svg")
logger.info("Using image path %r", imgPath)
aML = oed.alignPair(imagePath=imgPath)
if aML:
logger.info("%s aligned image path %r", ccId, imgPath)
for (rCC, rAt, tCC, tAt) in aML:
logger.debug("%5s %-5s %5s %-5s", rCC, rAt, tCC, tAt)
def write(self, filePath, oeMol, constantMol=False, addSdTags=True):
"""Write an oeMol with format type inferred from the filePath extension (e.g. .mol)
Args:
filePath (str): filepath with a chemical type extension
constantMol (bool, optional): copies molecule before performing format specific perceptions
Returns:
bool: True for success or False otherwise
"""
try:
molId = os.path.splitext(os.path.basename(filePath))[0]
fmt = os.path.splitext(os.path.basename(filePath))[1][1:].lower()
#
if addSdTags:
oemf = OeMoleculeFactory()
oemf.setOeMol(oeMol, molId)
oemf.addSdTags()
oeMol = oemf.getMol()
#
self.__mU.mkdir(os.path.dirname(filePath))
ofs = oechem.oemolostream()
ofs.open(filePath)
logger.debug("Writing (fmt=%s) molId %s path %s title %s", fmt,
molId, filePath, oeMol.GetTitle())
#
if constantMol:
oechem.OEWriteConstMolecule(ofs, oeMol)
else:
oechem.OEWriteMolecule(ofs, oeMol)
#
# If this is a mol2 file, we need to replace the resname
if fmt.startswith("mol2"):
# If this is a mol2/mol2h substitute the default substructure id
with open(filePath, "r", encoding="utf-8") as ifh:
lines = ifh.readlines()
lines = [line.replace("<0>", molId) for line in lines]
with open(filePath, "w", encoding="utf-8") as ofh:
ofh.writelines(lines)
return True
except Exception as e:
logger.exception("Failing for %s with %s", filePath, str(e))
return False
def chemCompToMol(self,
ccdFilePath,
molBuildType="model-xyz",
quietFlag=False):
retMolL = []
try:
rdCcObjL = self.__mU.doImport(ccdFilePath, fmt="mmcif")
logger.info("Read %s with %d definitions", ccdFilePath,
len(rdCcObjL))
oemf = OeMoleculeFactory()
if quietFlag:
oemf.setQuiet()
for ccObj in rdCcObjL:
ccId = oemf.setChemCompDef(ccObj)
if ccId:
ok = oemf.build(molBuildType=molBuildType)
if ok:
oeMol = oemf.getMol()
retMolL.append(oeMol)
except Exception as e:
logger.exception("Loading %s failing with %s", ccdFilePath, str(e))
return retMolL
def buildOeBinaryMolCacheFromIndex(self,
filePath,
ccIdxD,
quietFlag=False,
fpTypeList=None,
limitPerceptions=False,
suppressHydrogens=False):
"""Build cache of OEGraphMol() objects from the input chemical component search index.
Args:
filePath (str): output cache file path
ccIdxD (dict): search index dictionary
quietFlag (bool, optional): suppress OE output. Defaults to False.
fpTypeList (list, optional): list of fingerprint types. Defaults to None.
limitPerceptions (bool, optional): suppress automatic chemical perceptions. Defaults to False.
suppressHydrogens (bool, optional): suppress explicit hydrogen count. Defaults to False.
Returns:
(int, int, list): chem comp success count, error count, chem comp identifier failure list
"""
failIdList = []
ccCount = 0
errCount = 0
startTime = time.time()
try:
ofs = oechem.oemolostream()
ofs.SetFormat(oechem.OEFormat_OEB)
if ofs.open(filePath):
oemf = OeMoleculeFactory()
if quietFlag:
oemf.setQuiet()
for searchCcId, ccIdx in ccIdxD.items():
oemf.setDescriptor(ccIdx["smiles"], "oe-iso-smiles",
searchCcId)
ok = oemf.build(molBuildType="oe-iso-smiles",
limitPerceptions=limitPerceptions)
if ok and fpTypeList:
fpOk = oemf.addFingerPrints(fpTypeList)
if not fpOk:
logger.info("Fingerprint generation fails for %r",
searchCcId)
if ok:
if not suppressHydrogens:
oemf.addExplicitHydrogens()
oemf.setSimpleAtomNames()
oeMol = oemf.getMol(
suppressHydrogens=suppressHydrogens)
oechem.OEWriteMolecule(ofs, oeMol)
ccCount += 1
if not ok:
# build failed incomplete component (e.g. missing atoms or bonds)
errCount += 1
failIdList.append(searchCcId)
else:
logger.error("Unable to open cache database %s", filePath)
errCount += 1
except Exception as e:
logger.exception("Failing with %s", str(e))
#
endTime = time.time()
logger.info("Completed operation at %s (%.4f seconds)",
time.strftime("%Y %m %d %H:%M:%S", time.localtime()),
endTime - startTime)
return ccCount, errCount, failIdList
def buildOeBinaryMolCache(self,
filePath,
ccObjD,
molBuildType="model-xyz",
quietFlag=False,
fpTypeList=None,
limitPerceptions=False,
suppressHydrogens=False):
"""Build cache of OEMol() objects from the input chemical component definition list.
Args:
filePath (str): output cache file path
ccObjD (dict): chemical component object dictionary
molBuildType (str, optional): [description]. Defaults to "model-xyz".
quietFlag (bool, optional): [description]. Defaults to False.
fpTypeList (list, optional): fingerprint type list. Defaults to None.
limitPerceptions (bool, optional): suppress automatic chemical perceptions. Defaults to False.
suppressHydrogens (bool, optional): suppress explicit hydrogen count. Defaults to False.
Returns:
(int, int, list): chem comp success count, error count, chem comp identifier failure list
"""
ok = False
startTime = time.time()
failIdList = []
ccCount = 0
errCount = 0
try:
ofs = oechem.oemolostream()
ofs.SetFormat(oechem.OEFormat_OEB)
if ofs.open(filePath):
oemf = OeMoleculeFactory()
if quietFlag:
oemf.setQuiet()
for ccId, ccObj in ccObjD.items():
tId = oemf.setChemCompDef(ccObj)
if tId and tId == ccId:
ok = oemf.build(molBuildType=molBuildType,
limitPerceptions=limitPerceptions)
if ok and fpTypeList:
fpOk = oemf.addFingerPrints(fpTypeList)
if not fpOk:
logger.info(
"Fingerprint generation fails for %r",
ccId)
if ok:
oeMol = oemf.getMol(
suppressHydrogens=suppressHydrogens)
oechem.OEWriteMolecule(ofs, oeMol)
ccCount += 1
if not ok or not tId:
# build failed incomplete component (e.g. missing atoms or bonds)
errCount += 1
failIdList.append(ccId)
else:
logger.error("Unable to open cache database %s", filePath)
errCount += 1
except Exception as e:
logger.exception("Failing with %s", str(e))
#
endTime = time.time()
logger.info("Completed operation at %s (%.4f seconds)",
time.strftime("%Y %m %d %H:%M:%S", time.localtime()),
endTime - startTime)
return ccCount, errCount, failIdList
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