def GenAlignment(soi_seq, ts_seq, soi, ts,parameters):
lensoi = int(soi_seq.length)
lents = int(len(ts))
breakpoints = soi_seq.dataDepths
c_soi = clubgen_c.intArray(lensoi)
if breakpoints[0] == 0:
for i in range(lensoi):
c_soi[i] = soi[i]
else:
for i in range(lensoi):
c_soi[i] = soi[i]%2**breakpoints[0]
c_ts = clubgen_c.intArray(lents)
if breakpoints[0] == 0:
for i in range(lents):
c_ts[i] = ts[i]
else:
for i in range(lents):
c_ts[i] = ts[i]%2**breakpoints[0]
MAX_MATCH = 100
MAX_INLINE = 3
c_localStart = clubgen_c.intArray(MAX_MATCH)
c_localEnd = clubgen_c.intArray(MAX_MATCH)
c_localShift = clubgen_c.intArray(MAX_MATCH)
for i in range(MAX_MATCH):
c_localStart[i] = 0
c_localEnd[i] = 0
c_localShift[i] = 0
#print soiVecComp
#print tsVecComp
#print numMatches
#print 'lol?'
#print 'local window', parameters['LOCAL_WINDOW']
t0 = time.time()
a = clubgen_c.GenAlignment_c(c_soi,c_ts,lensoi,lents, parameters['GLOBAL_SIG_LEVEL'], parameters['LOCAL_SIG_LEVEL'], parameters['LOCAL_WINDOW'], MAX_MATCH, MAX_INLINE, parameters['LOCAL_BRIDGE_WIDTH'], parameters['COMB_SPEED'], c_localStart,c_localEnd,c_localShift)
TIMER = time.time() - t0
theAlignment = Alignment(soi_seq, ts_seq)
##############################
##### C does local ###########
##############################
if c_localStart[0] == c_localEnd[0]:
return theAlignment
cdfL = numpy.zeros(lensoi)
cdfR = numpy.zeros(lensoi)
if breakpoints[0] == 0:
colScore = numpy.zeros(lensoi)
else:
colScore = numpy.zeros((len(breakpoints),lensoi))
locWin = parameters['LOCAL_WINDOW']
i = 0
while(i < MAX_MATCH):
if c_localStart[i] != c_localEnd[i]:
fullScoreVec = numpy.zeros(lensoi)
if soi_seq.dataDepths[0] != 0:
colScore = numpy.zeros((len(soi_seq.dataDepths),lensoi))
else:
colScore = numpy.zeros(lensoi)
SOIrange = []
TSrange = []
while(c_localShift[i] == c_localShift[i+1] and c_localStart[i] != c_localEnd[i]):
#t1,t2,t3,t4,t5,t6 = cToPyDataFixer(c_localStart[i], c_localEnd[i], c_localShift[i], lensoi, lents, locWin, theAlignment)
#colScore += t1
#fullScoreVec += t2
t3,t4 = cFindRanges(c_localStart[i], c_localEnd[i], c_localShift[i], lensoi, lents, locWin)
SOIrange.append(t3)
TSrange.append(t4)
i += 1
#t1,t2,t3,t4,cdfL,cdfR = cToPyDataFixer(c_localStart[i], c_localEnd[i], c_localShift[i], lensoi, lents, locWin, theAlignment)
#colScore += t1
#fullScoreVec += t2
t3,t4 = cFindRanges(c_localStart[i], c_localEnd[i], c_localShift[i], lensoi, lents, locWin)
SOIrange.append(t3)
TSrange.append(t4)
soiStart = max(0,c_localShift[i]-lents+locWin)
soiEnd = min(lensoi-1,c_localShift[i]+locWin-1)
tsStart = max(0,lents-locWin-c_localShift[i])
tsEnd = lents-1 - max(0,locWin+c_localShift[i]-lensoi)
#locAlign = LocalAlignment(SOIrange,TSrange,colScore,cdfL,cdfR,fullScoreVec)
locAlign = LocalAlignment(SOIrange,TSrange,[soiStart,soiEnd],[tsStart,tsEnd])
theAlignment.AddLocalAlign(locAlign) # adds the local alignment to the total alignment
logging.debug('Added align! ' + str(SOIrange))
i += 1
else:
break
i = MAX_MATCH
return theAlignment
def GenClubsC(seqOfInt, testSeqs, parameters):
import clubgen_c
# build up sequence of interest
soiPartTotal = seqOfInt.length/1000+1
soiParts = clubgen_c.int_array_1000(soiPartTotal)
if seqOfInt.readingFrames > 1:
soiseq = seqOfInt.seq[0]
else:
soiseq = seqOfInt.seq
for i in range(soiPartTotal):
if seqOfInt.sequenceType == 'AA':
for j in range(min(1000,seqOfInt.length-1000*i)):
clubgen_c.a_set(i,j,numpy.int(soiseq[j+1000*i]),soiParts)
for j in range(min(1000,seqOfInt.length-1000*i),1000):
clubgen_c.a_set(i,j,0,soiParts)
else:
for j in range(min(1000,seqOfInt.length-1000*i)):
clubgen_c.a_set(i,j,numpy.int(soiseq[j+1000*i]%2**seqOfInt.dataDepths[0]),soiParts)
for j in range(min(1000,seqOfInt.length-1000*i),1000):
clubgen_c.a_set(i,j,0,soiParts)
tsPartTotal = 0
tsPartInfo = []
tsPartNum = 0
if seqOfInt.readingFrames == 1:
for j in range(len(testSeqs)):
tmpPart = testSeqs[j].length/1000+1
tsPartTotal += tmpPart
for i in range(tmpPart):
tsPartInfo.append(tsPartNum)
tsPartNum += 1
else:
for k in range(seqOfInt.readingFrames):
for j in range(len(testSeqs)):
tmpPart = len(testSeqs[j].seq[k])/1000+1
tsPartTotal += tmpPart
for i in range(tmpPart):
tsPartInfo.append(tsPartNum)
tsPartNum += 1
#print tsPartInfo
thePart = -1
counter = 0
tsPartNum = 0
tsParts = clubgen_c.int_array_1000(tsPartTotal)
if seqOfInt.readingFrames == 1:
if seqOfInt.sequenceType == 'AA':
for i in range(tsPartTotal):
if thePart == tsPartInfo[i]:
counter += 1
else:
thePart = tsPartInfo[i]
counter = 0
for j in range(min(1000,testSeqs[tsPartInfo[i]].length-1000*counter)):
clubgen_c.a_set(i,j,numpy.int(testSeqs[tsPartInfo[i]].seq[j+1000*counter]),tsParts)
for j in range(min(1000,testSeqs[tsPartInfo[i]].length-1000*counter),1000):
clubgen_c.a_set(i,j,0,tsParts)
else:
for i in range(tsPartTotal):
if thePart == tsPartInfo[i]:
counter += 1
else:
thePart = tsPartInfo[i]
counter = 0
for j in range(min(1000,testSeqs[tsPartInfo[i]].length-1000*counter)):
clubgen_c.a_set(i,j,numpy.int(testSeqs[tsPartInfo[i]].seq[j+1000*counter]%2**seqOfInt.dataDepths[0]),tsParts)
for j in range(min(1000,testSeqs[tsPartInfo[i]].length-1000*counter),1000):
clubgen_c.a_set(i,j,0,tsParts)
else:
#ts = numpy.zeros(tsPartTotal*1000,dtype = numpy.uint8)
for k in range(seqOfInt.readingFrames):
for i in range(tsPartTotal/seqOfInt.readingFrames):
infoLen = len(testSeqs[tsPartInfo[i]%len(testSeqs)].seq[k])
#print thePart,tsPartInfo[i+k*tsPartTotal/seqOfInt.readingFrames]
if thePart == tsPartInfo[i+k*tsPartTotal/seqOfInt.readingFrames]:
counter += 1
else:
thePart = tsPartInfo[i+k*tsPartTotal/seqOfInt.readingFrames]
counter = 0
#ts[1000*i+k*tsPartTotal/seqOfInt.readingFrames*1000:1000*i+k*tsPartTotal/seqOfInt.readingFrames*1000+min(1000,infoLen-1000*counter)] = testSeqs[tsPartInfo[i]%len(testSeqs)].seq[k][1000*counter:1000*counter+min(1000,infoLen-1000*counter)]%2**seqOfInt.dataDepths[0]
#ts[1000*i+k*tsPartTotal/seqOfInt.readingFrames*1000+min(1000,infoLen-1000*counter):1000*i+k*tsPartTotal/seqOfInt.readingFrames*1000+1000] = numpy.zeros(1000-min(1000,infoLen-1000*counter),dtype = numpy.uint8)
#print counter
#print infoLen-1000*counter
for j in range(min(1000,infoLen-1000*counter)):
clubgen_c.a_set(i+k*tsPartTotal/seqOfInt.readingFrames,j,numpy.int(testSeqs[tsPartInfo[i]%len(testSeqs)].seq[k][j+1000*counter]%2**seqOfInt.dataDepths[0]),tsParts)
#print 'set', i+k*tsPartTotal/seqOfInt.readingFrames,j,numpy.int(testSeqs[tsPartInfo[i]%len(testSeqs)].seq[k][j+1000*counter]%2**seqOfInt.dataDepths[0])
#ts[j+1000*i+k*tsPartTotal/seqOfInt.readingFrames*1000] = testSeqs[tsPartInfo[i]%len(testSeqs)].seq[k][j+1000*counter]%2**seqOfInt.dataDepths[0]
for j in range(min(1000,infoLen-1000*counter),1000):
clubgen_c.a_set(i+k*tsPartTotal/seqOfInt.readingFrames,j,0,tsParts)
#ts[j+1000*i+k*tsPartTotal/seqOfInt.readingFrames*1000] = 0
cTsPartInfo = clubgen_c.intArray(tsPartTotal)
for i in range(tsPartTotal):
cTsPartInfo[i] = tsPartInfo[i]
# make vector that contains all of the info
MAX_MATCH = 20
MAX_INLINE = 5
alignmentInfo = clubgen_c.intArray(MAX_MATCH*5*len(testSeqs)*seqOfInt.readingFrames)
for i in range(MAX_MATCH*5*len(testSeqs)*seqOfInt.readingFrames):
alignmentInfo[i] = 0;
if seqOfInt.readingFrames == 1:
tsLens = clubgen_c.intArray(len(testSeqs))
for i in range(len(testSeqs)):
tsLens[i] = testSeqs[i].length
else:
tsLens = clubgen_c.intArray(len(testSeqs)*seqOfInt.readingFrames)
for k in range(seqOfInt.readingFrames):
for i in range(len(testSeqs)):
tsLens[i+k*len(testSeqs)] = len(testSeqs[i].seq[k])
#tsLengths = clubgen_c.intArray(len(testSeqs))
#for i in range(len(testSeqs)):
#tsLengths[i] = testSeqs[i].length
#print 'about to go into c'
clubgen_c.TesterAlign_c(soiPartTotal, soiParts, tsPartTotal, tsParts, cTsPartInfo, alignmentInfo, seqOfInt.length, tsLens, parameters['GLOBAL_SIG_LEVEL'], parameters['LOCAL_SIG_LEVEL'], parameters['LOCAL_WINDOW'], MAX_MATCH, MAX_INLINE, parameters['LOCAL_BRIDGE_WIDTH'], parameters['COMB_SPEED'])
alignments = SetOfAlignments(seqOfInt)
i = 0
seqOn = 0
rframe = 0
locWin = parameters['LOCAL_WINDOW']
theAlignment = Alignment(seqOfInt, testSeqs[seqOn])
while(i < MAX_MATCH*5):
if alignmentInfo[seqOn*MAX_MATCH*5+i] != alignmentInfo[seqOn*MAX_MATCH*5+i+1]:
#logging.debug('alignmentInfo: ' + str(alignmentInfo[seqOn*MAX_MATCH*5+i]) + ',' + str(alignmentInfo[seqOn*MAX_MATCH*5+i+1]) + ',' + str(alignmentInfo[seqOn*MAX_MATCH*5+i+2]) + ',' + str(alignmentInfo[seqOn*MAX_MATCH*5+i+3]) + ',' + str(alignmentInfo[seqOn*MAX_MATCH*5+i+4]))
lensoi = min(1000,seqOfInt.length-1000*alignmentInfo[seqOn*MAX_MATCH*5+i+3])
lents = min(1000,testSeqs[seqOn%len(testSeqs)].length-1000*alignmentInfo[seqOn*MAX_MATCH*5+i+4])
soiShift = 1000*alignmentInfo[seqOn*MAX_MATCH*5+i+3]
tsShift = 1000*alignmentInfo[seqOn*MAX_MATCH*5+i+4]
SOIrange = []
TSrange = []
while(alignmentInfo[seqOn*MAX_MATCH*5+i+2] == alignmentInfo[seqOn*MAX_MATCH*5+i+7] and alignmentInfo[seqOn*MAX_MATCH*5+i] != alignmentInfo[seqOn*MAX_MATCH*5+i+1]):
t3,t4 = cFindRanges(alignmentInfo[seqOn*MAX_MATCH*5+i], alignmentInfo[seqOn*MAX_MATCH*5+i+1], alignmentInfo[seqOn*MAX_MATCH*5+i+2], lensoi, lents, locWin, soiShift, tsShift)
SOIrange.append(t3)
TSrange.append(t4)
i += 5
logging.debug('alignmentInfo: ' + str(alignmentInfo[seqOn*MAX_MATCH*5+i]) + ',' + str(alignmentInfo[seqOn*MAX_MATCH*5+i+1]) + ',' + str(alignmentInfo[seqOn*MAX_MATCH*5+i+2]) + ',' + str(alignmentInfo[seqOn*MAX_MATCH*5+i+3]) + ',' + str(alignmentInfo[seqOn*MAX_MATCH*5+i+4]))
t3,t4 = cFindRanges(alignmentInfo[seqOn*MAX_MATCH*5+i], alignmentInfo[seqOn*MAX_MATCH*5+i+1], alignmentInfo[seqOn*MAX_MATCH*5+i+2], lensoi, lents, locWin, soiShift, tsShift)
SOIrange.append(t3)
TSrange.append(t4)
soiStart = max(0,alignmentInfo[seqOn*MAX_MATCH*5+i+2]-lents+locWin) + soiShift
soiEnd = min(lensoi-1,alignmentInfo[seqOn*MAX_MATCH*5+i+2]+locWin-1) + soiShift
tsStart = max(0,lents-locWin-alignmentInfo[seqOn*MAX_MATCH*5+i+2]) + tsShift
tsEnd = lents-1 - max(0,locWin+alignmentInfo[seqOn*MAX_MATCH*5+i+2]-lensoi) + tsShift
#locAlign = LocalAlignment(SOIrange,TSrange,colScore,cdfL,cdfR,fullScoreVec)
logging.debug('Added align! ' + str(SOIrange) + ' to ' + str(TSrange) + ' of ' + testSeqs[seqOn%len(testSeqs)].name + ' rf ' + str(rframe))
locAlign = LocalAlignment(SOIrange,TSrange,[soiStart,soiEnd],[tsStart,tsEnd],rframe)
theAlignment.AddLocalAlign(locAlign) # adds the local alignment to the total alignment
i += 5
else:
#print 'added the alignment'
alignments.AddAlign(theAlignment)
if seqOn < len(testSeqs)*seqOfInt.readingFrames - 1:
i = 0
seqOn += 1
if seqOn%len(testSeqs) == 0:
rframe += 1
theAlignment = Alignment(seqOfInt, testSeqs[seqOn%len(testSeqs)])
else:
i = MAX_MATCH*5
return alignments
