def prefilterIndex(self,
oeQueryMol,
idxP,
matchOpts="relaxed",
skipFeatures=False):
"""Filter the full search index base on minimum chemical formula an feature criteria.
Args:
oeQueryMol (object): search target moleculed (OEMol)
idxP (object): instance ChemCompSearchIndexProvider()
matchOpts (str, optional): search criteria options. Defaults to "default".
skipFeatures (bool, optional): skip feature filters. Defaults to False.
Returns:
(list): list of chemical component identifiers in the filtered search space
"""
startTime = time.time()
oemf = OeMoleculeFactory()
oemf.setOeMol(oeQueryMol, "queryTarget")
typeCountD = oemf.getElementCounts(useSymbol=True)
# ccIdL1 = idxP.filterMinimumMolecularFormula(typeCountD)
#
featureCountD = oemf.getFeatureCounts() if not skipFeatures else {}
# Adjust filter according to search options
if matchOpts in matchOpts in [
"relaxed", "graph-relaxed", "simple",
"sub-struct-graph-relaxed"
]:
for ky in ["rings_ar", "at_ar", "at_ch"]:
featureCountD.pop(ky, None)
elif matchOpts in [
"relaxed-stereo", "graph-relaxed-stereo",
"sub-struct-graph-relaxed-stereo", "graph-relaxed-stereo-sdeq",
"sub-struct-graph-relaxed-stereo-sdeq"
]:
for ky in ["rings_ar", "at_ar"]:
featureCountD.pop(ky, None)
elif matchOpts in [
"default", "strict", "graph-strict", "graph-default",
"sub-struct-graph-strict"
]:
pass
ccIdL = idxP.filterMinimumFormulaAndFeatures(typeCountD, featureCountD)
logger.info(
"Pre-filtering results for formula+feature %d (%.4f seconds)",
len(ccIdL),
time.time() - startTime)
return ccIdL
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