def testSubStructureSearchFromIndexSelected(self):
matchOpts = self.__myKwargs.get("matchOpts", "sub-struct-graph-relaxed")
numProc = self.__numProcSearch
oemp = OeSearchMoleculeProvider(**self.__myKwargs)
ok = oemp.testCache()
self.assertTrue(ok)
oesU = OeSubStructSearchUtils(oemp)
#
ccIdxP = ChemCompSearchIndexProvider(**self.__myKwargs)
ok = ccIdxP.testCache(minCount=self.__minCount)
self.assertTrue(ok)
ccIdxD = ccIdxP.getIndex()
ky = next(iter(ccIdxD))
oeMol = oemp.getMol(ky)
#
for ccId in ["BNZ", "ALA"]:
# ----
startTime = time.time()
oeMol = oemp.getMol(ccId)
#
ccIdL = oesU.prefilterIndex(oeMol, ccIdxP, matchOpts=matchOpts)
logger.info("%s search length %d in (%.4f seconds)", ccId, len(ccIdL), time.time() - startTime)
#
retStatus, mL = oesU.searchSubStructure(oeMol, ccIdList=ccIdL, matchOpts=matchOpts, numProc=numProc)
logger.info("%s status %r result length %d in (%.4f seconds)", ccId, retStatus, len(mL), time.time() - startTime)
self.assertTrue(retStatus)
self.assertTrue(self.__resultContains(ccId, mL))
def updateSearchIndex(self, useCache=False):
"""Rebuild the search index from source chemical component and BIRD definitions.
Update the internal state of this index in the current object instance.
Resource requirements 771 secs 6 proc macbook pro 7GB memory.
Args:
useCache (bool): False to rebuild search index and True to reload
Returns:
bool: True for success or false otherwise
"""
ok = False
try:
kwargs = copy.deepcopy(
self.__configD["ccsiKwargs"]
) if "ccsiKwargs" in self.__configD else None
if kwargs:
kwargs["useCache"] = useCache
siIdxP = ChemCompSearchIndexProvider(**kwargs)
ok = siIdxP.testCache()
self.__siIdxP = siIdxP if siIdxP else None
self.__siIdx = siIdxP.getIndex() if siIdxP and ok else {}
logger.info("Search index status %r index len %d", ok,
len(self.__siIdx) if self.__siIdx else 0)
except Exception as e:
logger.exception("Failing with %s", str(e))
return ok
def __reload(self, **kwargs):
"""Reload the dictionary of OE molecules and related data artifacts for chemical component definitions.
Args:
limitPerceptions(bool): process input descriptors in essentially verbatim mode (default: True)
fpTypeList (list): fingerprint type (TREE,PATH,MACCS,CIRCULAR,LINGO)
screenTypeList (list): fast sub search screen type (MOLECULE, SMARTS, MDL, ... )
useCache (bool, optional): flag to use cached files. Defaults to True.
cachePath (str): path to the top cache directory. Defaults to '.'.
numProc (int): number processors to engage in screen substructure search database generation.
suppressHydrogens (bool, optional): flag to suppress explicit hydrogens in the OE data store.
molLimit (int):
Returns:
(bool) : True for success or False othewise
"""
try:
useCache = kwargs.get("useCache", True)
cachePath = kwargs.get("cachePath", ".")
numProc = kwargs.get("numProc", 2)
molLimit = kwargs.get("molLimit", None)
fpTypeList = kwargs.get(
"fpTypeList", ["TREE", "PATH", "MACCS", "CIRCULAR", "LINGO"])
# screenTypeList = kwargs.get("screenTypeList", ["SMARTS"])
screenTypeList = kwargs.get("screenTypeList", None)
limitPerceptions = kwargs.get("limitPerceptions", False)
suppressHydrogens = kwargs.get("suppressHydrogens", False)
quietFlag = kwargs.get("quietFlag", True)
logSizes = kwargs.get("logSizes", False)
fpDbType = "STANDARD"
buildScreenedDb = True
#
oeCount = 0
errCount = 0
failIdList = []
oeIo = OeIoUtils(quietFlag=quietFlag)
# --------
oeSearchMolFilePath = os.path.join(self.__dirPath,
self.__getOeSearchMolFileName())
if not useCache or (useCache
and not self.__mU.exists(oeSearchMolFilePath)):
cmpKwargs = {
k: v
for k, v in kwargs.items()
if k not in ["cachePath", "useCache", "molLimit"]
}
ccsiP = ChemCompSearchIndexProvider(cachePath=cachePath,
useCache=True,
molLimit=molLimit,
**cmpKwargs)
ok = ccsiP.testCache(minCount=molLimit, logSizes=logSizes)
# ----
ccIdxD = ccsiP.getIndex() if ok else {}
idxCount = len(ccIdxD)
# ------- JDW OE mol construction here -----
startTime = time.time()
oeCount, errCount, failIdList = oeIo.buildOeBinaryMolCacheFromIndex(
oeSearchMolFilePath,
ccIdxD,
quietFlag=quietFlag,
fpTypeList=fpTypeList,
limitPerceptions=limitPerceptions,
suppressHydrogens=suppressHydrogens)
if failIdList:
logger.info("failures %r", failIdList)
endTime = time.time()
logger.info(
"Constructed %d/%d cached oeMols (unconverted %d) (%.4f seconds)",
oeCount, idxCount, errCount, endTime - startTime)
# --------
oeMolDbFilePath = os.path.join(self.__dirPath,
self.__getOeMolDbFileName())
if not useCache or (useCache
and not self.__mU.exists(oeMolDbFilePath)):
startTime = time.time()
molCount = oeIo.createOeBinaryDatabaseAndIndex(
oeSearchMolFilePath, oeMolDbFilePath)
endTime = time.time()
logger.info(
"Created and stored %d indexed oeMols in OE database format (%.4f seconds)",
molCount, endTime - startTime)
# --------
if fpDbType == "FAST":
for fpType in fpTypeList:
startTime = time.time()
# Fast FP search database file names
fpPath = os.path.join(self.__dirPath,
self.__getFastFpDbFileName(fpType))
if not useCache or (useCache
and not self.__mU.exists(fpPath)):
ok = oeIo.createOeFingerPrintDatabase(oeMolDbFilePath,
fpPath,
fpType=fpType)
endTime = time.time()
logger.info(
"Created and stored %s fingerprint database (%.4f seconds)",
fpType, endTime - startTime)
# --------
if buildScreenedDb and screenTypeList:
for screenType in screenTypeList:
startTime = time.time()
fp = os.path.join(self.__dirPath,
self.__getSubSearchFileName(screenType))
if not useCache or (useCache and not self.__mU.exists(fp)):
ok = oeIo.createOeSubSearchDatabase(
oeSearchMolFilePath,
fp,
screenType=screenType,
numProc=numProc)
endTime = time.time()
logger.info(
"Constructed screened substructure database (status %r) with screenType %s (%.4f seconds)",
ok, screenType, endTime - startTime)
# ---------
ssDb = oeIo.loadOeSubSearchDatabase(
fp, screenType=screenType, numProc=numProc)
ok = ssDb.NumMolecules() == oeCount
# ----------
#
return True
except Exception as e:
logger.exception("Failing with %s", str(e))
return False
def buildSearchFiles(self, **kwargs):
"""Build cif, sdf (optional), and mol2 files for components in the chemical component search index.
Exclude ions or other extraneous molecules lacking bonds.
Args:
ccUrlTarget (str): locator for source chemical component dictionary (default: full public dictionary)
birdUrlTarget (str): locator for source BIRD dictionary (default: full public dictionary)
limitPerceptions (bool): restrict automatic perceptions in OE molecular build operations (default: False)
numProc (int): number of processors
useCache (bool): use existing resource file where possible (default: True)
molLimit (str): limit the number to ingested chemical compont (default: None)
quietFlag (bool): suppress output in OE library operations (default: True)
Returns:
(int): number molfiles generated
"""
cachePath = self.__cachePath
ccUrlTarget = kwargs.get("ccUrlTarget", None)
birdUrlTarget = kwargs.get("birdUrlTarget", None)
molLimit = kwargs.get("molLimit", None)
quietFlag = kwargs.get("quietFlag", True)
fpTypeList = kwargs.get("fpTypeList", [])
screenTypeList = kwargs.get("screenTypeList", [])
ccFileNamePrefix = "cc-%s" % self.__prefix if self.__prefix else "cc-full"
oeFileNamePrefix = "oe-%s" % self.__prefix if self.__prefix else "oe-cc-full"
numProc = kwargs.get("numProc", 2)
minCount = kwargs.get("minCount", 0)
useCache = kwargs.get("useCache", True)
useSdf = kwargs.get("useSdf", True)
useMol2 = kwargs.get("useMol2", False)
limitPerceptions = kwargs.get("limitPerceptions", False)
logSizes = False
#
startTime = time.time()
ccmP = ChemCompMoleculeProvider(cachePath=cachePath,
useCache=useCache,
ccFileNamePrefix=ccFileNamePrefix,
ccUrlTarget=ccUrlTarget,
birdUrlTarget=birdUrlTarget,
molLimit=molLimit)
ok = ccmP.testCache(minCount=minCount, logSizes=logSizes)
logger.info(
"Completed chemical component provider load %r (%.4f seconds)", ok,
time.time() - startTime)
#
startTime = time.time()
oesmp = OeSearchMoleculeProvider(
ccUrlTarget=ccUrlTarget,
birdUrlTarget=birdUrlTarget,
cachePath=cachePath,
ccFileNamePrefix=ccFileNamePrefix,
oeFileNamePrefix=oeFileNamePrefix,
useCache=useCache,
quietFlag=quietFlag,
fpTypeList=fpTypeList,
screenTypeList=screenTypeList,
numProc=numProc,
molLimit=molLimit,
limitPerceptions=limitPerceptions,
)
ok = oesmp.testCache()
logger.info("Completed OE molecule provider load %r (%.4f seconds)",
ok,
time.time() - startTime)
#
startTime = time.time()
ccSIdxP = ChemCompSearchIndexProvider(
cachePath=cachePath,
useCache=useCache,
ccFileNamePrefix=ccFileNamePrefix,
limitPerceptions=limitPerceptions,
numProc=numProc)
ok = ccSIdxP.testCache()
logger.info(
"Completed chemical component search index load %r (%.4f seconds)",
ok,
time.time() - startTime)
#
ccSIdx = ccSIdxP.getIndex() if ccSIdxP and ok else {}
logger.info("Search index status %r index length %d", ok, len(ccSIdx))
#
ccIdD = {}
mU = MarshalUtil()
oeU = OeIoUtils(dirPath=cachePath)
numMols = 0
searchFileDirPath = self.getSearchDirFilePath()
pathTupList = []
for sId in ccSIdx:
ccId = sId.split("|")[0]
# standard CIF definition
if ccId not in ccIdD:
cifPath = os.path.join(searchFileDirPath, ccId[0], ccId,
ccId + ".cif")
if not (useCache and mU.exists(cifPath)):
ccMol = ccmP.getMol(ccId)
if not self.__checkCif(ccMol):
continue
mU.doExport(cifPath, [ccMol], fmt="mmcif")
#
oeMol = oesmp.getMol(sId)
if not self.__checkOeMol(oeMol):
continue
#
# Sanity checks on the generated OE molecule
#
cifPath = os.path.join(searchFileDirPath, ccId[0], ccId,
sId + ".cif")
if sId != ccId and not (useCache and mU.exists(cifPath)):
oeccU = OeChemCompUtils()
ok = oeccU.addOeMol(sId,
oeMol,
missingModelXyz=True,
writeIdealXyz=False)
if ok:
oeccU.write(cifPath)
if useSdf:
molFilePath = os.path.join(searchFileDirPath, ccId[0], ccId,
sId + ".sdf")
if not (useCache and mU.exists(molFilePath)):
ok = oeU.write(molFilePath,
oeMol,
constantMol=False,
addSdTags=True)
if ok:
pathTupList.append((sId, molFilePath, "sdf"))
#
if useMol2:
mol2FilePath = os.path.join(searchFileDirPath, ccId[0], ccId,
sId + ".mol2")
if not (useCache and mU.exists(mol2FilePath)):
oeU.write(mol2FilePath,
oeMol,
constantMol=False,
addSdTags=True)
if ok:
pathTupList.append((sId, mol2FilePath, "mol2"))
numMols += 1
#
self.__storePathList(pathTupList)
return numMols