def __init__(self, **kwargs):
""" Module entry point for chemical component definition 2D image generation
Args:
ccUrlTarget (str): URL or path for concatenated chemical component dictionary (default: public wwPDB ftp)
birdUrlTarget (str): URL or path for concatenated BIRD dictionary (default: public wwPDB ftp)
licenseFilePath (str) = path to OpenEye license text file
fileDirPath(str): directory containing generated image tree
cachePath(str): cache directory for temporary files
molBuildType(str): build type for constructing OE moleclues ('ideal-xyz', 'model-xyz' default: 'ideal-xyz')
"""
#
ccUrlTarget = kwargs.get("ccUrlTarget", None)
birdUrlTarget = kwargs.get("birdUrlTarget", None)
self.__licensePath = kwargs.get("licenseFilePath", "oe_license.txt")
self.__fileDirPath = kwargs.get("fileDirPath", ".")
self.__molBuildType = kwargs.get("molBuildType", "ideal-xyz")
cachePath = kwargs.get("cachePath", ".")
cachePath = os.path.abspath(cachePath)
#
self.__oemp = OeMoleculeProvider(
ccUrlTarget=ccUrlTarget,
birdUrlTarget=birdUrlTarget,
ccFileNamePrefix="cc-full",
cachePath=cachePath,
molBuildType=self.__molBuildType,
useCache=True,
oeFileNamePrefix="oe-full",
)
self.__oeMolD = self.__oemp.getOeMolD()
def testSubStructureSearchBase(self):
matchOpts = self.__myKwargs.get("matchOpts", "sub-struct-graph-relaxed")
numProc = self.__numProcSearch
oemp = OeMoleculeProvider(**self.__myKwargs)
ok = oemp.testCache()
self.assertTrue(ok)
oesU = OeSubStructSearchUtils(oemp)
#
ccIdxP = ChemCompIndexProvider(**self.__myKwargs)
ok = ccIdxP.testCache(minCount=self.__minCount)
self.assertTrue(ok)
ccIdxD = ccIdxP.getIndex()
#
ky = next(iter(ccIdxD))
oeMol = oemp.getMol(ky)
#
numMols = 10
for ccId, _ in list(ccIdxD.items())[:numMols]:
# ----
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 result length %d in (%.4f seconds)", ccId, len(mL), time.time() - startTime)
self.assertTrue(retStatus)
self.assertTrue(self.__resultContains(ccId, mL))
def testSubStructureSearchScreened(self):
oeioU = OeIoUtils()
oemp = OeMoleculeProvider(**self.__myKwargs)
ok = oemp.testCache()
ccmP = ChemCompIndexProvider(**self.__myKwargs)
ccIdxD = ccmP.getIndex()
ok = ccmP.testCache(minCount=self.__minCount)
self.assertTrue(ok)
oesU = OeSearchUtils(oemp,
screenType=self.__screenType,
numProc=self.__numProc)
numMols = 20
missL = []
for ccId, ccD in list(ccIdxD.items())[:numMols]:
# ----
startTime = time.time()
if "oe-smiles" not in ccD:
continue
logger.info("Search %s %r", ccId, ccD["oe-smiles"])
oeQMol = oeioU.smartsToQmol(ccD["oe-smiles"])
retStatus, mL = oesU.searchSubStructureScreened(oeQMol,
maxMatches=100)
if retStatus:
logger.info("%s (status=%r) match length %d in (%.4f seconds)",
ccId, retStatus, len(mL),
time.time() - startTime)
if not self.__resultContains(ccId, mL):
missL.append(ccId)
#
# self.assertGreaterEqual(len(mL), 1)
# ----
logger.info("Missed searches (%d) %r", len(missL), missL)
def testSubStructureSearchWithFingerPrint(self):
oemp = OeMoleculeProvider(**self.__myKwargs)
#
ok = oemp.testCache()
ccmP = ChemCompIndexProvider(**self.__myKwargs)
ccIdxD = ccmP.getIndex()
ok = ccmP.testCache(minCount=self.__minCount)
self.assertTrue(ok)
minFpScore = 0.40
maxFpResults = 50
numMols = 20
matchOpts = "graph-relaxed"
oesU = OeSearchUtils(oemp, fpTypeList=self.__fpTypeList)
# ----
startTime = time.time()
for ccId, _ in list(ccIdxD.items())[:numMols]:
for fpType in self.__fpTypeList:
oeMol = oemp.getMol(ccId)
retStatus, mL = oesU.searchSubStructureWithFingerPrint(
oeMol,
fpType,
minFpScore,
maxFpResults,
matchOpts=matchOpts)
self.assertTrue(retStatus)
self.assertTrue(self.__resultContains(ccId, mL))
logger.info("%s fingerprints search on %d in (%.4f seconds)",
len(self.__fpTypeList), numMols,
time.time() - startTime)
def testFingerPrintSearch(self):
oemp = OeMoleculeProvider(**self.__myKwargs)
# This will reload the oe binary cache.
oeMol = oemp.getMol("004")
self.assertGreaterEqual(len(list(oeMol.GetAtoms())), 12)
#
ok = oemp.testCache()
ccmP = ChemCompIndexProvider(**self.__myKwargs)
ccIdxD = ccmP.getIndex()
ok = ccmP.testCache(minCount=self.__minCount)
self.assertTrue(ok)
minScore = 0.50
maxResults = 50
numMols = 50
oesU = OeSearchUtils(oemp, fpTypeList=self.__fpTypeList)
# ----
startTime = time.time()
for ccId, _ in list(ccIdxD.items())[:numMols]:
for fpType in self.__fpTypeList:
oeMol = oemp.getMol(ccId)
retStatus, mL = oesU.searchFingerPrints(
oeMol,
fpType=fpType,
minFpScore=minScore,
maxFpResults=maxResults)
self.assertTrue(retStatus)
self.assertTrue(self.__resultContains(ccId, mL))
# self.assertGreaterEqual(len(mL), 1)
logger.info("%s fingerprints search on %d in (%.4f seconds)",
len(self.__fpTypeList), numMols,
time.time() - startTime)
def __getCache(self, molBuildType="model-xyz", useCache=True):
oemp = OeMoleculeProvider(
ccUrlTarget=self.__ccUrlTarget,
birdUrlTarget=self.__birdUrlTarget,
ccFileNamePrefix="cc-abbrev",
cachePath=self.__cachePath,
molBuildType=molBuildType,
useCache=useCache,
)
ok = oemp.testCache()
self.assertTrue(ok)
return oemp.getOeMolD()
def testSubStructureSearch(self):
oemp = OeMoleculeProvider(**self.__myKwargs)
ok = oemp.testCache()
ccmP = ChemCompIndexProvider(**self.__myKwargs)
ccIdxD = ccmP.getIndex()
ok = ccmP.testCache(minCount=self.__minCount)
self.assertTrue(ok)
oesU = OeSearchUtils(oemp, fpTypeList=self.__fpTypeList)
numMols = 10
for ccId, _ in list(ccIdxD.items())[:numMols]:
# ----
startTime = time.time()
oeMol = oemp.getMol(ccId)
retStatus, mL = oesU.searchSubStructure(oeMol, matchOpts="relaxed")
logger.info("%s match length %d in (%.4f seconds)", ccId, len(mL),
time.time() - startTime)
self.assertTrue(retStatus)
self.assertTrue(self.__resultContains(ccId, mL))
def testSubStructureSearchScreenedFiltered(self):
myKwargs = {
"cachePath": self.__cachePath,
"useCache": True,
"fpTypeList": self.__fpTypeList,
"ccFileNamePrefix": "cc-filtered",
"oeFileNamePrefix": "oe-filtered",
"molBuildType": "oe-iso-smiles",
"limitPerceptions": False,
}
oeioU = OeIoUtils()
oemp = OeMoleculeProvider(**myKwargs)
ok = oemp.testCache()
ccmP = ChemCompIndexProvider(**myKwargs)
ccIdxD = ccmP.getIndex()
ok = ccmP.testCache(minCount=self.__minCount)
self.assertTrue(ok)
oesU = OeSearchUtils(oemp,
screenType=self.__screenType,
numProc=self.__numProc)
numMols = 5000
missL = []
for ccId, ccD in list(ccIdxD.items())[:numMols]:
# ----
startTime = time.time()
if "oe-smiles" not in ccD:
continue
logger.info("Search %s %r", ccId, ccD["oe-smiles"])
oeQMol = oeioU.smartsToQmol(ccD["oe-smiles"])
retStatus, mL = oesU.searchSubStructureScreened(oeQMol,
maxMatches=100)
logger.info("%s (status=%r)match length %d in (%.4f seconds)",
ccId, retStatus, len(mL),
time.time() - startTime)
if not self.__resultContains(ccId, mL):
missL.append(ccId)
# self.assertGreaterEqual(len(mL), 1)
# ----
logger.info("Missed searches (%d) %r", len(missL), missL)
class ChemCompFileWorkflow(object):
def __init__(self, **kwargs):
""" Module entry point for chemical component definition 2D image generation
Args:
ccUrlTarget (str): URL or path for concatenated chemical component dictionary (default: public wwPDB ftp)
birdUrlTarget (str): URL or path for concatenated BIRD dictionary (default: public wwPDB ftp)
licenseFilePath (str) = path to OpenEye license text file
fileDirPath(str): directory containing generated image tree
cachePath(str): cache directory for temporary files
molBuildType(str): build type for constructing OE moleclues ('ideal-xyz', 'model-xyz' default: 'ideal-xyz')
"""
#
ccUrlTarget = kwargs.get("ccUrlTarget", None)
birdUrlTarget = kwargs.get("birdUrlTarget", None)
self.__licensePath = kwargs.get("licenseFilePath", "oe_license.txt")
self.__fileDirPath = kwargs.get("fileDirPath", ".")
self.__molBuildType = kwargs.get("molBuildType", "ideal-xyz")
cachePath = kwargs.get("cachePath", ".")
cachePath = os.path.abspath(cachePath)
#
self.__oemp = OeMoleculeProvider(
ccUrlTarget=ccUrlTarget,
birdUrlTarget=birdUrlTarget,
ccFileNamePrefix="cc-full",
cachePath=cachePath,
molBuildType=self.__molBuildType,
useCache=True,
oeFileNamePrefix="oe-full",
)
self.__oeMolD = self.__oemp.getOeMolD()
def __setLicense(self, licensePath):
ok = False
try:
if os.environ.get("OE_LICENSE") and os.access(os.environ["OE_LICENSE"], os.R_OK):
logger.info("Using license from environment %r", os.environ["OE_LICENSE"])
ok = True
elif os.access(licensePath, os.R_OK):
os.environ["OE_LICENSE"] = licensePath
logger.info("Setting environmenal variable OE_LICENSE to %r", os.environ["OE_LICENSE"])
ok = True
except Exception as e:
logger.error("Setting license file %r failing %s", licensePath, str(e))
return ok
def testCache(self):
return self.__oemp.testCache() if self.__oemp else False
def makeFiles(self, fmt="sdf"):
""" Create files (mol, mol2) for all public chemical components.
"""
try:
if fmt not in ["mol", "mol2", "mol2h", "sdf"]:
return False
if not self.__setLicense(self.__licensePath):
logger.error("Invalid license details - exiting")
return False
for ccId, oeMol in self.__oeMolD.items():
if self.__molBuildType == "ideal-xyz":
filePath = os.path.join(self.__fileDirPath, fmt, ccId[0], ccId + "_ideal." + fmt)
oeioU = OeIoUtils()
oeioU.write(filePath, oeMol, constantMol=True)
else:
filePath = os.path.join(self.__fileDirPath, fmt, ccId[0], ccId + "_model." + fmt)
oeioU = OeIoUtils()
oeioU.write(filePath, oeMol, constantMol=True)
return True
except Exception as e:
logger.exception("Failing with %s", str(e))
return False
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)
class ChemCompImageWorkflow(object):
def __init__(self, **kwargs):
""" Module entry point for chemical component definition 2D image generation
Args:
ccUrlTarget (str): URL or path for concatenated chemical component dictionary (default: public wwPDB ftp)
birdUrlTarget (str): URL or path for concatenated BIRD dictionary (default: public wwPDB ftp)
licenseFilePath (str) = path to OpenEye license text file
imagePath(str): directory containing generated image tree
cachePath(str): cache directory for temporary files
"""
#
ccUrlTarget = kwargs.get("ccUrlTarget", None)
birdUrlTarget = kwargs.get("birdUrlTarget", None)
self.__licensePath = kwargs.get("licenseFilePath", "oe_license.txt")
self.__imagePath = kwargs.get("imagePath", ".")
cachePath = kwargs.get("cachePath", ".")
cachePath = os.path.abspath(cachePath)
#
self.__oemp = OeMoleculeProvider(
ccUrlTarget=ccUrlTarget,
birdUrlTarget=birdUrlTarget,
ccFileNamePrefix="cc-full",
cachePath=cachePath,
molBuildType="model-xyz",
useCache=True,
oeFileNamePrefix="oe-full",
)
self.__oeMolD = self.__oemp.getOeMolD()
def __setLicense(self, licensePath):
ok = False
try:
if os.environ.get("OE_LICENSE") and os.access(
os.environ["OE_LICENSE"], os.R_OK):
logger.info("Using license from environment %r",
os.environ["OE_LICENSE"])
ok = True
elif os.access(licensePath, os.R_OK):
os.environ["OE_LICENSE"] = licensePath
logger.info("Setting environmenal variable OE_LICENSE to %r",
os.environ["OE_LICENSE"])
ok = True
except Exception as e:
logger.error("Setting license file %r failing %s", licensePath,
str(e))
return ok
def testCache(self):
return self.__oemp.testCache() if self.__oemp else False
def makeImages(self):
""" Create images for all public chemical components with and without atom labels.
"""
try:
if not self.__setLicense(self.__licensePath):
logger.error("Invalid license details - exiting")
return False
for ccId, oeMol in self.__oeMolD.items():
imagePath = os.path.join(self.__imagePath, "image", ccId[0],
ccId + ".svg")
oed = OeDepict()
title = ""
oed.setMolTitleList([(ccId, oeMol, title)])
# ---
bondDisplayWidth = 10.0
numAtoms = oeMol.NumAtoms()
if numAtoms > 100 and numAtoms <= 200:
bondDisplayWidth = 6.0
elif numAtoms > 200:
bondDisplayWidth = 4.0
# ---
oed.setDisplayOptions(
labelAtomName=False,
labelAtomCIPStereo=True,
labelAtomIndex=False,
labelBondIndex=False,
labelBondCIPStereo=True,
cellBorders=False,
bondDisplayWidth=bondDisplayWidth,
)
oed.setGridOptions(rows=1, cols=1, cellBorders=False)
oed.prepare()
oed.write(imagePath)
for ccId, oeMol in self.__oeMolD.items():
imagePath = os.path.join(self.__imagePath, "image_labeled",
ccId[0], ccId + ".svg")
oed = OeDepict()
title = ""
oed.setMolTitleList([(ccId, oeMol, title)])
# ---
bondDisplayWidth = 10.0
numAtoms = oeMol.NumAtoms()
if numAtoms > 100 and numAtoms <= 200:
bondDisplayWidth = 6.0
elif numAtoms > 200:
bondDisplayWidth = 4.0
# ---
oed.setDisplayOptions(
labelAtomName=True,
labelAtomCIPStereo=True,
labelAtomIndex=False,
labelBondIndex=False,
labelBondCIPStereo=True,
cellBorders=False,
bondDisplayWidth=bondDisplayWidth,
)
oed.setGridOptions(rows=1, cols=1, cellBorders=False)
oed.prepare()
oed.write(imagePath)
return True
except Exception as e:
logger.exception("Failing with %s", str(e))
return False
def __testBuildMoleculeCacheFiles(self, **kwargs):
"""Test build OE cache files from full component dictionary"""
ccUrlTarget = kwargs.get("ccUrlTarget", None)
birdUrlTarget = kwargs.get("birdUrlTarget", None)
molLimit = kwargs.get("molLimit", 0)
quietFlag = kwargs.get("quietFlag", True)
molBuildType = kwargs.get("molBuildType", "ideal-xyz")
fpTypeList = kwargs.get("fpTypeList", ["TREE"])
screenTypeList = kwargs.get("screenTypeList", [])
ccFileNamePrefix = kwargs.get("ccFileNamePrefix", "cc")
oeFileNamePrefix = kwargs.get("oeFileNamePrefix", "oe")
#
startTime = time.time()
oemp = OeMoleculeProvider(
ccUrlTarget=ccUrlTarget,
birdUrlTarget=birdUrlTarget,
cachePath=self.__cachePath,
ccFileNamePrefix=ccFileNamePrefix,
oeFileNamePrefix=oeFileNamePrefix,
molBuildType=molBuildType,
useCache=False,
quietFlag=quietFlag,
fpTypeList=fpTypeList,
screenTypeList=screenTypeList,
)
ok = oemp.testCache()
self.assertTrue(ok)
oemp = OeMoleculeProvider(
ccUrlTarget=ccUrlTarget,
birdUrlTarget=birdUrlTarget,
cachePath=self.__cachePath,
ccFileNamePrefix=ccFileNamePrefix,
oeFileNamePrefix=oeFileNamePrefix,
molBuildType=molBuildType,
useCache=True,
)
endTime = time.time()
logger.info(">> Completed load molBuildType %r molLimit %r (%.4f seconds)", molBuildType, molLimit, endTime - startTime)
#
# ---
deltaMol = 2
minMol = minNumFp = molLimit - deltaMol if molLimit else 30000
for fpType in fpTypeList:
fpDb = oemp.getFingerPrintDb(fpType="TREE")
logger.debug("fpType %r length %d", fpType, fpDb.NumFingerPrints())
self.assertGreaterEqual(fpDb.NumFingerPrints(), minNumFp)
#
ccId = "004"
oeMol = oemp.getMol(ccId)
logger.debug("%s atom count %d", ccId, len(list(oeMol.GetAtoms())))
#
if molBuildType in ["oe-iso-smiles"]:
self.assertGreaterEqual(len(list(oeMol.GetAtoms())), 12)
else:
self.assertGreaterEqual(len(list(oeMol.GetAtoms())), 20)
#
oeDb, oeDbIdx = oemp.getOeMolDatabase()
logger.debug("Type db %r length %d type idx %r length %d", type(oeDb), oeDb.NumMols(), type(oeDbIdx), len(oeDbIdx))
self.assertGreaterEqual(oeDb.NumMols(), minMol)
self.assertGreaterEqual(len(oeDbIdx), minMol)
#
if molBuildType in ["oe-iso-smiles"] and screenTypeList:
ssDb = oemp.getSubSearchDb()
self.assertGreaterEqual(ssDb.NumMolecules(), minMol)
return True
