def computeFP(self, FPtype):
# set SMI after cleanning
prSMIclean = self.prDesc + "SMIclean/"
pathFolder.createFolder(prSMIclean)
self.prSMIclean = prSMIclean
dFP = {}
i = 1
for pSMI in listdir(
self.prSMI
): # !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
# for pSMI in ["/home/borrela2/interference/spDataAnalysis/Desc/SMIclean/1212-72-2.smi"]: # to verify for one chem
cas = pSMI.split("/")[-1].split(".")[0]
print cas, i, len(listdir(self.prSMI))
i += 1
psmiles = self.prSMI + cas + ".smi"
if path.exists(self.prSMI + cas + ".smi"):
fsmiles = open(psmiles, "r")
smiles = fsmiles.readlines()[0].strip()
fsmiles.close()
# chemical
chem = chemical.chemical(cas, smiles)
chem.prepareChem(prSMIclean)
error = chem.computeFP(FPtype)
if error == 1:
print "ERROR FP"
continue
else:
dFP[cas] = chem.FP
self.dFP = dFP
def search_release (term):
sFktname = "search_release"
term = term.replace(" ","+")
ePrint(1, sFktname, "search the web")
# search on chemical
query_chemical = chemical.chemical()
results = query_chemical.search(term)
# case if nothing was found
if len(results) == 0:
ePrint(1, sFktname, "nothing found on chemical")
return None
# case, if more then one was found
if len(results) > 1:
ePrint(1, sFktname, "Multiple Links found. Make a Choice\n")
k = 0
for i in results:
print("{0:3d}:\t{1}".format(k , results[k][1]))
k = k + 1
print("\n")
choice = int(input(" <-- "))
results = [ [ results[choice][0], results[choice][1] ] ]
# create an release instance and feed it
rel = release()
rel.shortinfo = results[0][1]
rel.infopage = results[0][0]
rel = query_chemical.getReleaseInfo(rel)
return rel
def computeDesc(self, opera=0, RDkitPhysico=1, pOperaDesc=""):
pdesc1D2D = self.prDesc + "tableDesc1D2D"
self.pdesc1D2D = pdesc1D2D
if opera == 1:
pdesc1D2D = pdesc1D2D + "Opera"
self.pdesc1D2D = pdesc1D2D
if RDkitPhysico == 0:
pdesc1D2D = pdesc1D2D + "NoRDKITPhyChem"
self.pdesc1D2D = pdesc1D2D
prSMIclean = self.prDesc + "SMIclean/"
pathFolder.createFolder(prSMIclean)
self.prSMIclean = prSMIclean
prDescbyCAS = self.prDesc + "DESCbyCAS/"
pathFolder.createFolder(prDescbyCAS)
self.prDescByCAS = prDescbyCAS
if path.exists(pdesc1D2D) and path.getsize(pdesc1D2D) > 100:
return pdesc1D2D
else:
plog = self.prDesc + "log.log"
flog = open(plog, "w")
print pdesc1D2D, "No found"
fdesc1D2D = open(pdesc1D2D, "w")
if opera == 0:
ldesc = chemical.getLdesc("1D2D", RDkitPhysico)
fdesc1D2D.write("CAS\t" + "\t".join(ldesc) + "\n")
else:
ldesc = chemical.getLdesc("1D2D", RDkitPhysico)
ldesc = ldesc + chemical.getLdesc("Opera", RDkitPhysico)
doperaDesc = loadAllOperaDesc(pOperaDesc)
fdesc1D2D.write("CAS\t" + "\t".join(ldesc) + "\n")
#print ldesc
for pSMI in listdir(self.prSMI):
#for pSMI in ["/home/borrela2/interference/spDataAnalysis/Desc/SMIclean/1212-72-2.smi"]: # to verify for one chem
cas = pSMI.split("/")[-1].split(".")[0]
#print cas
psmiles = self.prSMI + cas + ".smi"
if path.exists(self.prSMI + cas + ".smi"):
fsmiles = open(psmiles, "r")
smiles = fsmiles.readlines()[0].strip()
fsmiles.close()
# chemical
chem = chemical.chemical(cas, smiles)
chem.prepareChem(prSMIclean)
chem.compute1D2DDesc(prDescbyCAS)
if opera == 1:
chem.loadOperaDesc(doperaDesc, flog)
err = chem.writeTablesDescCAS(prDescbyCAS) #
if err == 1: chem.writelog(self.prlog)
#Write in the table
chem.writeDesc(ldesc, fdesc1D2D)
fdesc1D2D.close()
flog.close()
def search_clever (tr):
sFktname = "search_clever"
# build search-term
if tr.artist != "" and tr.title != "":
term = tr.artist + "+" + tr.title
term = term.replace("&","")
term = term.replace("(","")
term = term.replace(")","")
term = term.replace(" ","+")
# first try to find the track in the release
# cache
if len(cache.rels_found) > 0:
for rel in cache.rels_found:
ePrint(1, sFktname, "search in cache")
# search track in release
res = rel.search_track(tr)
# nothing found
if res == None:
continue
# more than one item found
if len(res) > 1:
ePrint(1, sFktname, "Multiple tracks found in release, make a choice")
k = 0
for t in res:
print("{0:3d}:\t{1}".format(k , t))
k = k + 1
print("\n")
choice = int(input(" <-- "))
res = res[choice]
return res
return res[0]
ePrint(1, sFktname, "search the web")
# search on chemical
query_chemical = chemical.chemical()
results = query_chemical.search(term)
# case if nothing was found
if len(results) == 0:
ePrint(1, sFktname, "nothing found on chemical")
return None
# case, if more then one was found
if len(results) > 1:
ePrint(1, sFktname, "Multiple Links found. Make a Choice\n")
k = 0
for i in results:
print("{0:3d}:\t{1}".format(k , results[k][1]))
k = k + 1
print("\n")
choice = int(input(" <-- "))
results = [ [ results[choice][0], results[choice][1] ] ]
# create an release instance and feed it
rel = release()
rel.shortinfo = results[0][1]
rel.infopage = results[0][0]
rel = query_chemical.getReleaseInfo(rel)
# append the release to the cache
cache.rels_found.append(rel)
# search track in release
res = rel.search_track(tr)
# nothing found
if res == None:
ePrint(1, sFktname, "Track not found in release page")
return None
# more than one item found
if len(res) > 1:
ePrint(1, sFktname, "Multible tracks found in release, make a choice")
k = 0
for t in res:
print("{0:3d}:\t{1}".format(k , t))
k = k + 1
print("\n")
choice = int(input(" <-- "))
res = res[choice]
return res[0]
def computeDesc(passay, PRDESC, PRSMI, prout, nbfile=1, update=0):
# by pass
pdescout = prout + "descMat"
paff = prout + "aff.txt"
if path.exists(pdescout) and update == 0 and nbfile == 1:
return pdescout
elif path.exists(pdescout) and update == 0 and nbfile == 2 and path.exists(
paff):
return [pdescout, paff]
dchem = toolbox.loadMatrix(passay)
lchemID = dchem.keys()
try:
lchemID.remove("RESULT_UNIT")
except:
pass
shuffle(lchemID)
i = 0
nbi = len(lchemID)
while i < nbi:
if search("error", dchem[lchemID[i]]
["SMILES"].lower()): # case of the table is computed before
del dchem[lchemID[i]]
del lchemID[i]
nbi = nbi - 1
continue
if dchem[lchemID[i]]["Active"] == "Inconclusive" or search(
"Error", dchem[lchemID[i]]["SMILES"]):
del dchem[lchemID[i]]
del lchemID[i]
nbi = nbi - 1
continue
# compute descriptors
cchem = chemical.chemical(lchemID[i], dchem[lchemID[i]]["SMILES"])
cchem.prepareChem(PRSMI)
if search("error", cchem.log.lower()):
del dchem[lchemID[i]]
del lchemID[i]
nbi = nbi - 1
continue
cchem.compute1D2DDesc(PRDESC)
if search("error", cchem.log.lower()):
del dchem[lchemID[i]]
del lchemID[i]
nbi = nbi - 1
continue
cchem.computeOpera(update=update)
if search("error", cchem.log.lower()):
del dchem[lchemID[i]]
del lchemID[i]
nbi = nbi - 1
i = i - 1
continue
cchem.writeTablesDesc(PRDESC, update=update)
i = i + 1
if nbfile == 1:
fildesc = open(pdescout, "w")
ldesc = chemical.getLdesc("1D2D", 1) + chemical.getLdesc("Opera", 0)
fildesc.write("ID," + ",".join(ldesc) + ",Aff" + "\n")
for chemID in lchemID:
print chemID
if dchem[chemID]["Active"] == "Active":
aff = 1
else:
aff = 0
pdesc = PRDESC + chemID + ".txt"
if path.exists(pdesc):
ddesc = toolbox.loadMatrix(pdesc)
lval = []
for desc in ldesc:
if not desc in ddesc[chemID].keys():
lval.append("NA")
else:
lval.append(str(ddesc[chemID][desc]))
fildesc.write(chemID + "," + ",".join(lval) + "," + str(aff) +
"\n")
fildesc.close()
return pdescout
else:
fildesc = open(pdescout, "w")
paff = prout + "aff.txt"
filaff = open(paff, "w")
ldesc = chemical.getLdesc("1D2D", 1) + chemical.getLdesc("Opera", 0)
fildesc.write("ID," + ",".join(ldesc) + "\n")
filaff.write("ID\tAff\n")
for chemID in lchemID:
print chemID
if dchem[chemID]["Active"] == "Active":
aff = 1
else:
aff = 0
pdesc = PRDESC + chemID + ".txt"
if path.exists(pdesc):
ddesc = toolbox.loadMatrix(pdesc)
lval = []
for desc in ldesc:
if not desc in ddesc[chemID].keys():
lval.append("NA")
else:
lval.append(str(ddesc[chemID][desc]))
fildesc.write(chemID + "," + ",".join(lval) + "," + str(aff) +
"\n")
filaff.write(chemID + "\t" + str(aff) + "\n")
fildesc.close()
filaff.close()
return [pdescout, paff]
def predictSMI(self, nameChemical, smiles, plot=0, verbose=0):
dpred = {}
prresult = pathFolder.createFolder(self.prout + nameChemical + "/")
ppred = prresult + "pred"
if path.exists(ppred):
dpred = toolbox.loadMatrix(ppred)
return dpred
chem = chemical.chemical(nameChemical, smiles)
chem.prepareChem(prresult)
chem.compute1D2DDesc(prresult)
chem.writeTablesDescCAS(prresult)
chem.computeFP(typeFP="All")
for channel in self.dcluster:
for cell in self.dcluster[channel].keys():
kpred = str(cell) + "_" + str(channel)
dpred[kpred] = {}
for typeDesc in self.dcluster[channel][cell].keys():
if verbose == 1:
print channel, cell, typeDesc
print self.dcluster[channel][cell].keys()
if search("Desc", typeDesc):
distMeth = typeDesc.split("-")[1]
aggMeth = typeDesc.split("-")[2]
enrichment = runExternalSoft.findCluster(
self.cDB.pdesc1D2Dclean, chem.pdesc,
self.dcluster[channel][cell][typeDesc]["files"][0],
self.dcluster[channel][cell][typeDesc]["files"][1],
distMeth, aggMeth)
else:
# generate FP
typeFP = typeDesc.split("-")[0]
metric = typeDesc.split("-")[-1].split("_")[0]
metricAgg = typeDesc.split("-")[-1]
if verbose == 1: print typeFP, metric
dFP = {}
for CASID in self.cDB.dFP.keys():
if verbose == 1:
print self.cDB.dFP[CASID]
print chem.FP[typeFP]
print metric
dFP[CASID] = float(
toolbox.computeSimilarityFP(
self.cDB.dFP[CASID][typeFP],
chem.FP[typeFP], metric))
maxSim = max(dFP.values())
i = 0
imax = len(dFP.keys())
lCAS = dFP.keys()
while i < imax:
if float(dFP[lCAS[i]] == maxSim):
CASclose = lCAS[i]
i += 1
if verbose == 1:
print CASclose
print channel, cell
print self.ChemClust[CASclose][channel][cell]
clusterfound = self.ChemClust[CASclose][channel][cell][
str(typeFP) + "-" + str(metricAgg)]
enrichment = self.dcluster[channel][cell][typeDesc][
clusterfound]['Enrichment']
dpred[kpred][typeDesc] = enrichment
if plot == 1:
self.writeResultBySMI(dpred, prresult)
return dpred
