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 __buildChemCompSearchIndex(self, procName, ccIdList, descrD, limitPerceptions=False, quietFlag=False):
"""Internal method return a dictionary of extracted chemical component descriptors and formula."""
rL = []
fL = []
try:
for ccId in ccIdList:
if ccId not in self.__ccObjD:
logger.error("%s missing chemical definition for %s", procName, ccId)
fL.append(ccId)
continue
dataContainer = self.__ccObjD[ccId]
# ----
oemf = OeMoleculeFactory()
if quietFlag:
oemf.setQuiet()
tId = oemf.setChemCompDef(dataContainer)
if tId != ccId:
logger.error("%s %s chemical component definition import error", procName, ccId)
fL.append(ccId)
continue
# ----
oemf.clearExternalDescriptors()
for smi in descrD[ccId] if ccId in descrD else []:
oemf.addExternalDescriptor("smiles", smi, "chemaxon-smiles")
# ----
relD = oemf.buildRelated(limitPerceptions=limitPerceptions)
logger.debug("%s %s related molecular forms %d", procName, ccId, len(relD))
if relD:
rL.extend([relD[v] for v in relD])
else:
fL.append(ccId)
except Exception as e:
logger.exception("%s failing with %s", procName, str(e))
return rL, fL
def __buildChemCompSearchIndex(self, ccObjD, descrD, limitPerceptions=False, molLimit=None):
"""Internal method return a dictionary of extracted chemical component descriptors and formula."""
rD = {}
try:
for ii, ccId in enumerate(ccObjD, 1):
if molLimit and ii > molLimit:
break
# ----
oemf = OeMoleculeFactory()
oemf.setQuiet()
tId = oemf.setChemCompDef(ccObjD[ccId])
if tId != ccId:
logger.error("%s chemical component definition import error", ccId)
# ----
oemf.clearExternalDescriptors()
for smi in descrD[ccId] if ccId in descrD else []:
oemf.addExternalDescriptor("smiles", smi, "chemaxon-smiles")
# ----
smiD = oemf.buildRelated(limitPerceptions=limitPerceptions)
logger.debug("%s related molecular forms %d", ccId, len(smiD))
rD.update(smiD)
except Exception as e:
logger.exception("Failing with %s", str(e))
return rD
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 testBuilders(self):
try:
ccMolD = self.__getChemCompDefs()
quietFlag = False
molBuildTypeL = ["model-xyz", "ideal-xyz"]
for molBuildType in molBuildTypeL:
oemf = OeMoleculeFactory()
if quietFlag:
oemf.setQuiet()
#
eCount = 0
for tId, ccObj in ccMolD.items():
ccId = oemf.setChemCompDef(ccObj)
self.assertEqual(tId, ccId)
logger.debug("Building %s using molBuildType %r", ccId,
molBuildType)
if ccId:
ok = oemf.build(molBuildType=molBuildType)
logger.debug(
"Comparing built component %s using molBuildType %r",
ccId, molBuildType)
# ok = oemf.compare()
ok = True
if not ok:
logger.info("Failing on %s molBuildType %r", ccId,
molBuildType)
eCount += 1
# self.assertTrue(ok)
else:
logger.error("Cannot process %r", ccObj.getName())
logger.info(
"Processed %d components molBuildType %r errors %d",
len(ccMolD), molBuildType, eCount)
#
except Exception as e:
logger.exception("Failing with %s", str(e))
self.fail()
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 testCompareByBuildType(self):
"""Compare depictions constructed molecules with various builds from chemical defintions -
all build types 8769 (all)
connect - smiles 6743
model vs iso smiles 5937
ideal va iso smiles 7047
"""
doDepict = False
ccResultD = {}
genResultD = {}
smilesByBuildTypeD = {}
try:
ccMolD, ccIdxD = self.__getChemCompDefs()
#
limitPerceptions = True
# molBuildTypeL = ["model-xyz", "ideal-xyz", "connection-table", "oe-iso-smiles"]
molBuildTypeL = ["ideal-xyz", "oe-iso-smiles"]
#
startTime = time.time()
oefm = OeMoleculeFactory()
oefm.setQuiet()
for molBuildType in molBuildTypeL:
for ccId, idxD in ccIdxD.items():
ccObj = ccMolD[ccId]
# ----
ccIsoSmiles = idxD["oe-iso-smiles"]
ccSmiles = idxD["oe-smiles"]
# ----
tId = oefm.setChemCompDef(ccObj)
if not tId:
logger.info("Skipping bad component %r", ccId)
continue
self.assertEqual(tId, ccId)
ok = oefm.build(molBuildType=molBuildType,
limitPerceptions=limitPerceptions)
if not ok:
logger.info("Build using %r failed for %s",
molBuildType, ccId)
continue
# ------
oeMol = oefm.getGraphMol()
oeIsoSmiles = oefm.getIsoSMILES()
oeSmiles = oefm.getCanSMILES()
ccEq = oeIsoSmiles == ccIsoSmiles and oeSmiles == ccSmiles
#
oefmR = OeMoleculeFactory()
oefmR.setQuiet()
ccIdGen = ccId + "_gen"
oefmR.setDescriptor(oeIsoSmiles, "oe-iso-smiles", ccIdGen)
ok = oefmR.build(molBuildType="oe-iso-smiles",
limitPerceptions=limitPerceptions)
if not ok:
logger.info("Build using %r failed for %s",
molBuildType, ccIdGen)
continue
# ------
#
# oeMolGen = oefmR.getGraphMol()
oeIsoSmilesGen = oefmR.getIsoSMILES()
oeSmilesGen = oefmR.getCanSMILES()
genEq = oeIsoSmiles == oeIsoSmilesGen and oeSmiles == oeSmilesGen
smilesByBuildTypeD.setdefault(ccId, {}).setdefault(
molBuildType, []).append(oeIsoSmilesGen)
#
logger.debug("%s buildType %s ccEq %r genEq %r", ccId,
molBuildType, ccEq, genEq)
if not ccEq:
ccResultD.setdefault(molBuildType, []).append(ccId)
if not genEq:
genResultD.setdefault(molBuildType, []).append(ccId)
if doDepict:
pS = "-limited" if limitPerceptions else ""
imagePath = os.path.join(
self.__workPath,
ccId + "-%s%s.svg" % (molBuildType, pS))
oed = OeDepict()
title = ""
oed.setMolTitleList([(ccId, oeMol, title)])
oed.setDisplayOptions(labelAtomName=False,
labelAtomCIPStereo=True,
labelAtomIndex=False,
labelBondIndex=False,
cellBorders=False,
bondDisplayWidth=0.5)
oed.setGridOptions(rows=1, cols=1)
oed.prepare()
oed.write(imagePath)
logger.info(
"Completed comparing %d molecules in %d builds (%.4f seconds)",
len(ccIdxD), len(molBuildTypeL),
time.time() - startTime)
#
#
for molBuildType in molBuildTypeL:
if molBuildType in genResultD:
logger.info("GEN %s (%d) %r", molBuildType,
len(genResultD[molBuildType]),
genResultD[molBuildType])
numDiff = 0
for ccId, btD in smilesByBuildTypeD.items():
tS = set()
for molBuildType, sL in btD.items():
tS.add(sL[0])
if len(tS) > 1:
numDiff += 1
logger.debug("%s diff smiles (%d) %r", ccId, len(tS), tS)
logger.info("Components with inconsistent SMILES %d", numDiff)
#
except Exception as e:
logger.exception("Failing with %s", str(e))
self.fail()
def __buildChemCompIndex(self,
cD,
molBuildType="model-xyz",
doFeatures=True):
"""Internal method return a dictionary of extracted chemical component descriptors and formula."""
rD = {}
try:
quietFlag = True
for _, dataContainer in cD.items():
ccIt = iter(PdbxChemCompIt(dataContainer))
cc = next(ccIt, None)
ccId = cc.getId()
formula = str(cc.getFormula()).replace(" ", "")
ambiguousFlag = cc.getAmbiguousFlag().upper() in ["Y", "YES"]
tch = cc.getFormalCharge()
fcharge = int(tch) if tch and tch not in [".", "?"] else 0
#
logger.debug("ccId %r formula %r ambiguous %r fcharge %r",
ccId, formula, ambiguousFlag, fcharge)
if fcharge:
sign = "+" if fcharge > 0 else "-"
mag = str(abs(fcharge)) if abs(fcharge) > 1 else ""
formula = formula + sign + mag
#
atIt = PdbxChemCompAtomIt(dataContainer)
typeCounts = defaultdict(int)
for at in atIt:
aType = at.getType().upper()
typeCounts[aType] += 1
#
rD[ccId] = {
"formula": formula,
"type-counts": typeCounts,
"ambiguous": ambiguousFlag,
"feature-counts": {}
}
desIt = PdbxChemCompDescriptorIt(dataContainer)
for des in desIt:
desBuildType = des.getMolBuildType()
tS = des.getDescriptor()
descr = tS.strip() if tS else None
if not descr:
continue
if desBuildType in [
"oe-iso-smiles", "oe-smiles", "acdlabs-smiles",
"cactvs-iso-smiles", "cactvs-smiles", "inchi",
"inchikey"
]:
rD[ccId][desBuildType] = descr
else:
logger.error("%s unexpected descriptor build type %r",
ccId, desBuildType)
if doFeatures:
oemf = OeMoleculeFactory()
if quietFlag:
oemf.setQuiet()
tId = oemf.setChemCompDef(dataContainer)
if tId != ccId:
logger.error(
"%s chemical component definition import error",
ccId)
continue
ok = oemf.build(molBuildType=molBuildType)
if ok:
rD[ccId]["feature-counts"] = oemf.getFeatureCounts()
except Exception as e:
logger.exception("Failing with %s", str(e))
return rD
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