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 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 __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 testRoundTrip(self):
"""Round trip smiles comparisons -"""
try:
ccMolD = self.__getChemCompDefs()
# useCache = True
# quietFlag = False
# molBuildTypeL = ["model-xyz", "ideal-xyz", None]
# molBuildTypeL = [None]
buildTypeRef = "oe-iso-smiles"
oemf1 = OeMoleculeFactory()
oemf2 = OeMoleculeFactory()
#
for ccId, ccObj in ccMolD.items():
# ----
ccIt = iter(PdbxChemCompIt(ccObj))
cc = next(ccIt)
formula = cc.getFormulaWithCharge()
# ccId = cc.getId()
ccName = cc.getName()
ifCharge = cc.getFormalChargeAsInt()
isAmbiguous = cc.getAmbiguousFlag() in ["Y", "y"]
isCurrent = cc.getReleaseStatus() in ["REL"]
logger.debug("%s name %r formula %r charge %d", ccId, ccName,
formula, ifCharge)
# ----
ccId = oemf1.setChemCompDef(ccObj)
ok = oemf1.build(molBuildType=buildTypeRef,
limitPerceptions=False)
if not ok:
logger.info(
"Build using %r failed for %s (ambiguous flag %r current %r)",
buildTypeRef, ccId, isAmbiguous, isCurrent)
#
isDiff = False
#
if isDiff:
genIsoSmi = oemf1.getCanSMILES()
oemf2 = OeMoleculeFactory()
oemf2.setDescriptor(genIsoSmi, "oe-iso-smiles", ccId)
oemf2.build(molBuildType="oe-iso-smiles",
limitPerceptions=False)
regenIsoSmi = oemf2.getIsoSMILES()
if genIsoSmi != regenIsoSmi:
logger.info(
"%s regenerated ISOSMILES differ \n -- INP: %s\n -- OUT: %s",
ccId, genIsoSmi, regenIsoSmi)
oed = OeDepictMCSAlignPage()
oed.setDisplayOptions(
labelAtomName=True,
labelAtomCIPStereo=True,
labelAtomIndex=False,
labelBondIndex=False,
labelBondCIPStereo=True,
highlightStyleFit="ballAndStickInverse",
highLightNotMatchColorRef="pink",
bondDisplayWidth=0.5,
)
oed.setRefMol(oemf1.getGraphMol(), ccId)
oed.setFitMol(oemf2.getGraphMol(), ccId)
imgPath = os.path.join(
self.__workPath,
"compare-assigned-" + ccId + "-calc-" + ccId + ".svg")
logger.info("Using image path %r", imgPath)
aML = oed.alignPair(imagePath=imgPath)
if aML:
for (rCC, rAt, tCC, tAt) in aML:
logger.info("%5s %-5s %5s %-5s", rCC, rAt, tCC,
tAt)
else:
logger.debug("%s matched all cases", ccId)
except Exception as e:
logger.exception("Failing with %s", str(e))
self.fail()
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