def checkArray(self,val,shape,dtype,raiseShape=0):
if type(val)==type([]):
val=numpy.array(val)
elif type(val)==type(""):
if os.path.exists(val):
print "Loading %s"%val
val=FITS.Read(val)[1]
else:
raise Exception("File not found %s"%val)
if type(val)==numpy.ndarray:
if dtype is not None:
val=val.astype(dtype)
if shape is not None:
if type(shape)!=type(()):
shape=(shape,)
if val.shape!=shape:
print "Warning - shape not quite right (expecting %s, got %s)?"%(str(shape),str(val.shape))
if raiseShape:
raise Exception("checkArray shape")
try:
val.shape=shape
except:
print val.shape,shape
raise
else:
raise Exception("checkArray")
return val
def DarcInit(self, data):
"""data can be a FITS filename, or the contents of the FITS file"""
if os.path.exists(data):
data = FITS.Read(data)[1]
nbytes = data.size * data.itemsize
self.sock.send(
numpy.array(
[0x55555555, DARCINIT, nbytes & 0xffffffff,
nbytes >> 32]).astype(numpy.int32))
self.sock.sendall(data)
self.checkReply()
def __init__(self, name):
"""For extracting data from a large FITS file."""
self.bpdict = {
8: numpy.uint8,
16: numpy.int16,
32: numpy.int32,
-32: numpy.float32,
-64: numpy.float64,
-16: numpy.uint16
}
self.name = name
self.fd = open(self.name, "r")
self.HDUoffset = 0
self.hdr = FITS.ReadHeader(self.fd)["parsed"]
self.nd = int(self.hdr["NAXIS"])
dims = []
for i in range(self.nd):
dims.append(int(self.hdr["NAXIS%d" % (i + 1)]))
dims.reverse()
self.dims = numpy.array(dims)
self.bitpix = int(self.hdr["BITPIX"])
self.dataOffset = self.fd.tell()
dsize = self.getDataSize(self.hdr)
#Now get the frame list - move to the frame list HDU
self.fd.seek(dsize, 1)
try:
self.frameno = FITS.Read(self.fd, allHDU=0)[1]
except:
print "Unable to read frame numbers"
traceback.print_exc()
try:
self.timestamp = FITS.Read(self.fd, allHDU=0)[1]
except:
print "Unable to read timestamps"
traceback.print_exc()
self.nextHDUoffset = self.fd.tell()
def checkNoneOrArray(self,val,size,dtype):
if type(val)==type([]):
val=numpy.array(val)
elif type(val)==type(""):
if os.path.exists(val):
print "Loading %s"%val
val=FITS.Read(val)[1]
else:
print "File %s not found"%val
raise Exception("File %s not found"%val)
if val is None:
pass
elif type(val)==numpy.ndarray:
val=val.astype(dtype)
if size is not None:
val.shape=size,
else:
raise Exception("checkNoneOrArray size requested %d"%size)
return val
def loadSineData(self,fname,stdThresh=0.25):
"""
fname is the name of a file that has been created by e.g. doSinePokeGTC()
stdThresh is the threshold above which slopes are ignored if their std is higher than the max std * thresh. This allows us to automatically cut out slopes the don't have light."""
sList=[]
vList=[]
data=FITS.Read(fname)
for i in range(len(data)//12):
s=data[1+i*12]#the slopes
vmes=data[7+i*12]#the actuators (sinusoidal)
if stdThresh>0:
#define a mask of valid subaps based on those with low rms.
sstd=s.std(0)#get standard deviation of each subap
maxstd=sstd.max()#find the max...
valid=numpy.where(sstd<maxstd*stdThresh,1,0)
s*=valid
sList.append(s)
vList.append(vmes)
return vList,sList
import sys
import numpy
import darc
import FITS
rcond=float(sys.argv[1])
pmx=FITS.Read("pmx.fits")[1]#This contains high order DM and TT.
rmx=-numpy.linalg.pinv(pmx,rcond).T#separating out the TT from rest would be better - but I'm too lazy...
FITS.Write(rmx,"rmx%g.fits"%rcond)
d=darc.Control("main")
d.Set("rmx",rmx)
print "Saved rmx%g.fits and set in darc"%rcond
def valid(self,label,val,buf):
"""Checks a value is valid. buf is the buffer that contains all the other parameters"""
#buf=self.guibuf
if label=="reconstructMode":
if(val not in ["simple","truth","open","offset"]):
raise Exception(label)
elif label=="windowMode":
if val not in ["basic","adaptive","global"]:
raise Exception(label)
elif label in ["cameraName","mirrorName","comment","slopeName","figureName","version","configfile"]:
if type(val)!=type(""):
raise Exception(label)
elif label in ["reconName","calibrateName","bufferName"]:
if type(val) not in [type(""),type(None)]:
raise Exception(label)
elif label=="centroidMode":
if type(val)==numpy.ndarray:
nsubaps=buf.get("nsub").sum()
try:
val=self.checkArray(val,nsubaps,"i")
except:
print "centroidMode array wrong"
traceback.print_exc()
raise
elif val not in ["WPU","CoG","Gaussian","CorrelationCoG","CorrelationGaussian",0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15]:
print "centroidMode not correct (%s)"%str(val)
raise Exception(label)
elif label in ["cameraParams","mirrorParams","slopeParams","figureParams","reconParams","calibrateParams","bufferParams"]:
if type(val)==type(""):
l=4-len(val)%4
if l<4:
val+="\0"*l
val=numpy.fromstring(val,dtype=numpy.int32)
if type(val)==type([]):
val=numpy.array(val)
if type(val)!=type(None) and type(val)!=numpy.ndarray:
print "ERROR in val for %s: %s"%(label,str(val))
raise Exception(label)
elif label in ["closeLoop","nacts","thresholdAlgo","delay","maxClipped","camerasFraming","camerasOpen","mirrorOpen","clearErrors","frameno","corrThreshType","nsubapsTogether","nsteps","addActuators","recordCents","averageImg","averageCent","kalmanPhaseSize","figureOpen","printUnused","reconlibOpen","currentErrors","xenicsExposure","calibrateOpen","iterSource","bufferOpen","bufferUseSeq","subapLocType","noPrePostThread","asyncReset","openLoopIfClip","threadAffElSize","mirrorStep","mirrorUpdate","mirrorReset","mirrorGetPos","mirrorDoMidRange","lqgPhaseSize","lqgActSize"]:
val=int(val)
elif label in ["dmDescription"]:
if val.dtype.char!="h":
raise Exception("dmDescription type not int16")
elif label in ["actuators"]:
val=self.checkNoneOrArray(val,None,"f")
if val is not None:
if val.size%buf.get("nacts")==0:
# shape okay... multiple of nacts.
pass
else:
raise Exception("actuators should be array (nacts,) or (X,nacts)")
elif label in ["actMax","actMin"]:
nact=buf.get("nacts")
try:
val=self.checkArray(val,nact,None)
except:
print "WARNING (Check.py) - actMax/actMin as int now depreciated... this may not work (depending on which version of RTC you're running). The error was"
#traceback.print_exc()
#val=int(val)
print "Continuing... using %s"%str(val)
#elif label in["lastActs"]:
# nact=buf.get("nacts")
# val=self.checkNoneOrArray(val,nact,"H")
elif label in ["actInit"]:
val=self.checkNoneOrArray(val,None,"H")
elif label in ["actMapping"]:
val=self.checkNoneOrArray(val,None,"i")
elif label in ["figureActSource"]:
val=self.checkNoneOrArray(val,None,"i")
elif label in ["figureActScale","figureActOffset","actScale","actOffset"]:
val=self.checkNoneOrArray(val,None,"f")
elif label in ["actuatorMask"]:
val=self.checkNoneOrArray(val,None,"f")
if val is not None:
if val.size==buf.get("nacts"):
# shape okay... multiple of nacts.
pass
else:
raise Exception("actuatorMask should be array (nacts,)")
elif label in ["actSequence"]:
actuators=buf.get("actuators")
if actuators is None:
size=None
else:
size=actuators.size/buf.get("nacts")
val=self.checkNoneOrArray(val,size,"i")
elif label in ["bgImage","flatField","darkNoise","pxlWeight","calmult","calsub","calthr"]:
val=self.checkNoneOrArray(val,(buf.get("npxlx")*buf.get("npxly")).sum(),"f")
elif label in ["thresholdValue"]:
if type(val)==type(""):
if os.path.exists(val):
val=FITS.Read(val)[1]
else:
val=eval(val)
if type(val)==numpy.ndarray:
npxls=(buf.get("npxlx")*buf.get("npxly")).sum()
if val.size==npxls:
val=self.checkArray(val,(npxls,),"f")
else:
val=self.checkArray(val,(buf.get("nsub").sum(),),"f")
else:
try:
val=float(val)
except:
print "thresholdValue: %s"%str(type(val))
print "thresholdValue should be float or array of floats"
raise
elif label in ["useBrightest"]:
if type(val)==type(""):
if os.path.exists(val):
val=FITS.Read(val)[1]
else:
val=eval(val)
if type(val)==numpy.ndarray:
val=self.checkArray(val,(buf.get("nsub").sum(),),"i")
else:
try:
val=int(val)
except:
print "%s: %s"%(label,str(type(val)))
print "%s should be int or array of ints size equal to total number of subapertures (valid and invalid)"%label
raise
elif label in ["fluxThreshold"]:
if type(val)==type(""):
if os.path.exists(val):
val=FITS.Read(val)[1]
else:
val=eval(val)
if type(val)==numpy.ndarray:
val=self.checkArray(val,buf.get("subapFlag").sum(),"f")
else:
try:
val=float(val)
except:
print "%s: %s"%(label,str(type(val)))
print "%s should be float or array of floats size equal to tutal number of used subapertures"
raise
elif label in ["maxAdapOffset"]:
if type(val)==type(""):
if os.path.exists(val):
val=FITS.Read(val)[1]
else:
val=eval(val)
if type(val)==numpy.ndarray:
val=self.checkArray(val,buf.get("subapFlag").sum(),"i")
else:
try:
val=int(val)
except:
print "maxAdapOffset",val
print "maxAdapOffset should be int or array of ints of size equal to number of valid subaps %s"%str(type(val))
raise
elif label in ["powerFactor","adaptiveWinGain","corrThresh","figureGain","uEyeFrameRate","uEyeExpTime"]:
val=float(val)
elif label=="bleedGain":
if type(val)==numpy.ndarray:
if val.dtype.char!='f':
val=val.astype('f')
else:
val=float(val)
elif label=="bleedGroups":
val=self.checkNoneOrArray(val,buf.get("nacts"),"i")
elif label in ["switchTime"]:
val=self.checkDouble(val)
elif label in ["fakeCCDImage"]:
val=self.checkNoneOrArray(val,(buf.get("npxlx")*buf.get("npxly")).sum(),"f")
elif label in ["centroidWeight"]:
val=self.checkNoneOrFloat(val)
elif label in ["gainE","E"]:
if val is None:
pass
else:
if type(val)==numpy.ndarray:
if val.dtype!=numpy.float32:
val=val.astype(numpy.float32)
if val.shape!=(buf.get("nacts"),buf.get("nacts")) and val.shape!=(buf.get("subapFlag").sum()*2,buf.get("nacts")):
raise Exception("E should be shape nacts,nacts or nslopes,nacts")
else:#lazy - this doesn't check for nslopes,nacts...
val=self.checkArray(val,(buf.get("nacts"),buf.get("nacts")),"f")
elif label in ["gainReconmxT"]:
val=self.checkArray(val,(buf.get("subapFlag").sum()*2,buf.get("nacts")),"f")
elif label in ["kalmanAtur"]:
val=self.checkArray(val,(buf.get("kalmanPhaseSize"),buf.get("kalmanPhaseSize")),"f")
elif label in ["kalmanInvN"]:
val=self.checkNoneOrArray(val,buf.get("nacts")*buf.get("kalmanPhaseSize"),"f")
if val is not None:#now check shape
val=self.checkArray(val,(buf.get("nacts"),buf.get("kalmanPhaseSize")),"f")
elif label in ["kalmanHinfDM"]:
val=self.checkArray(val,(buf.get("kalmanPhaseSize")*3,buf.get("kalmanPhaseSize")),"f")
elif label in ["kalmanHinfT"]:
val=self.checkArray(val,(buf.get("subapFlag").sum()*2,buf.get("kalmanPhaseSize")*3),"f")
elif label in ["kalmanReset","kalmanUsed","printTime","usingDMC","go","pause","switchRequested","startCamerasFraming","stopCamerasFraming","openCameras","closeCameras","centFraming","slopeOpen"]:
val=self.checkFlag(val)
elif label in ["nsub","ncamThreads","npxlx","npxly"]:
val=self.checkArray(val,buf.get("ncam"),"i")
elif label in ["pxlCnt","subapFlag"]:
val=self.checkArray(val,buf.get("nsub").sum(),"i")
elif label in ["refCentroids"]:
val=self.checkNoneOrArray(val,buf.get("subapFlag").sum()*2,"f")
elif label in ["centCalBounds"]:
val=self.checkNoneOrArray(val,buf.get("subapFlag").sum()*2*2,"f")
elif label in ["centCalSteps","centCalData"]:
ncents=buf.get("subapFlag").sum()*2
val=self.checkNoneOrArray(val,None,"f")
if val is not None:
nsteps=val.size/ncents
if val.size!=nsteps*ncents:
raise Exception("%s wrong shape - should be multiple of %d, is %d"%(label,ncents,val.size))
elif label in ["subapLocation"]:
slt=buf.get("subapLocType")
if slt==0:
val=self.checkArray(val,(buf.get("nsub").sum(),6),"i")
else:
val=self.checkArray(val,None,"i")
n=val.size//buf.get("nsub").sum()#get the size and test again
val=self.checkArray(val,(buf.get("nsub").sum(),n),"i")
elif label in ["subapAllocation"]:
val=self.checkNoneOrArray(val,buf.get("nsub").sum(),"i")
elif label in ["gain"]:
val=self.checkArray(val,buf.get("nacts"),"f")
elif label in ["v0"]:
val=self.checkNoneOrArray(val,buf.get("nacts"),"f")
elif label in ["asyncInitState","asyncScales","asyncOffsets"]:
val=self.checkNoneOrArray(val,buf.get("nacts"),"f")
elif label in ["asyncCombines","asyncUpdates","asyncStarts","asyncTypes"]:
val=self.checkNoneOrArray(val,None,"i")
elif label in ["decayFactor"]:
val=self.checkNoneOrArray(val,buf.get("nacts"),"f")
elif label in ["rmx"]:
if val is None and buf.get("reconName") not in ["libreconpcg.so","libreconneural.so","libreconLQG.so","libreconcure.so"]:
raise Exception("rmx is None")
elif val is not None:
val=self.checkArray(val,(buf.get("nacts"),buf.get("subapFlag").sum()*2),"f",raiseShape=1)
elif label in ["slopeSumMatrix"]:
val=self.checkNoneOrArray(val,None,"f")
if (val is not None) and val.size%buf.get("nacts")!=0:
raise Exception("slopeSumMatrix wrong size")
elif label in ["slopeSumGroup"]:
val=self.checkNoneOrArray(val,buf.get("subapFlag").sum()*2,"i")
if (val is not None) and numpy.max(val)+1!=(buf.get("slopeSumMatrix").size/buf.get("nacts")):
raise Exception("Groupings in slopeSumGroup not consistent with size of slopeSumMatrix")
elif label in ["ncam"]:
val=int(val)
if val<1:
raise Exception("Illegal ncam")
elif label in ["threadAffinity"]:
if val is None:
pass
elif type(val)==numpy.ndarray:
if val.dtype!="i":
val=val.astype("i")
if val.size%(buf.get("ncamThreads").sum()+1)!=0:
raise Exception("threadAffinity error (size not multiple of %d)"%(buf.get("ncamThreads").sum()+1))
else:
raise Exception("threadAffinity error (should be an array, or None)")
elif label in ["threadPriority"]:
val=self.checkNoneOrArray(val,buf.get("ncamThreads").sum()+1,"i")
elif label in ["corrFFTPattern","corrPSF"]:
if type(val)==numpy.ndarray:
val=val.astype(numpy.float32)
elif val is not None:
raise Exception("corrFFTPattern error")
#val=self.checkNoneOrArray(val,(buf.get("npxlx")*buf.get("npxly")).sum(),"f")
elif label in ["adaptiveGroup"]:
val=self.checkNoneOrArray(val,buf.get("subapFlag").sum(),"i")
elif label in ["asyncNames"]:
pass#no checking needed...
elif label in ["adapWinShiftCnt"]:
val=self.checkNoneOrArray(val,buf.get("nsub").sum()*2,"i")
elif label in ["centIndexArray"]:
if type(val)==type([]):
val=numpy.array(val)
elif type(val)==type(""):
if os.path.exists(val):
print "Loading %s"%val
val=FITS.Read(val)[1]
else:
print "File %s not found"%val
raise Exception("File %s not found"%val)
if val is None:
pass
elif type(val)==numpy.ndarray:
val=val.astype("f")
fft=buf.get("corrFFTPattern",None)
if fft is None:
npxls=(buf.get("npxlx")*buf.get("npxly")).sum()
else:
npxls=fft.size
if val.size not in [npxls,npxls*2,npxls*3,npxls*4]:
raise Exception("centIndexArray wrong size")
else:
raise Exception("centIndexArray")
elif label=="actsToSend":
val=self.checkNoneOrArray(val,None,"i")
elif label in ["mirrorSteps","mirrorMidRange"]:#used for mirrorLLS.c
val=self.checkArray(val,buf.get("nacts"),"i")
elif label in ["lqgAHwfs"]:
val=self.checkArray(val,(buf.get("lqgPhaseSize")*2,buf.get("lqgPhaseSize")),"f",raiseShape=1)
elif label in ["lqgAtur"]:
val=self.checkArray(val,(buf.get("lqgPhaseSize"),buf.get("lqgPhaseSize")),"f",raiseShape=1)
elif label in ["lqgHT"]:
val=self.checkArray(val,(buf.get("subapFlag").sum()*2,2*buf.get("lqgPhaseSize")),"f",raiseShape=1)
elif label in ["lqgHdm"]:
try:
val=self.checkArray(val,(2*buf.get("lqgPhaseSize"),buf.get("nacts")),"f",raiseShape=1)
except:
val=self.checkArray(val,(2,2*buf.get("lqgPhaseSize"),buf.get("nacts")),"f",raiseShape=1)
elif label in ["lqgInvN"]:
try:
val=self.checkArray(val,(buf.get("nacts"),buf.get("lqgPhaseSize")),"f",raiseShape=1)
except:
val=self.checkArray(val,(2,buf.get("nacts"),buf.get("lqgPhaseSize")),"f",raiseShape=1)
elif label in ["lqgInvNHT"]:
val=self.checkArray(val,(buf.get("subapFlag").sum()*2,buf.get("nacts")),"f",raiseShape=1)
elif CustomCheck is not None:
val=CustomCheck.valid(label,val,buf)
else:
print "Unchecked parameter %s"%label
return val
def loadBuf(fname, hdu=0):
data = FITS.Read(fname)[hdu * 2 + 1]
b = Buffer(None, size=data.size)
b.assign(data)
return b
import controlCorba
import FITS
import sys
import time
file="tmp.fits"
if len(sys.argv)>1:
file=sys.argv[1]
#Create the corba client
c=controlCorba.controlClient()
def set(name,val,com="",swap=1,check=1):
c.obj.Set(controlCorba.sdata(name),controlCorba.encode([val]),controlCorba.sdata(com),swap,check)
#time.sleep(0.01)
set("fakeCCDImage",FITS.Read("shimage.fits")[1][0])
FITS.Write(controlCorba.decode(c.obj.AverageCentroids(1)),file)
set("thresholdAlgorithm",2)
FITS.Write(controlCorba.decode(c.obj.AverageCentroids(1)),file,writeMode="a")
set("thresholdAlgorithm",0)
FITS.Write(controlCorba.decode(c.obj.AverageCentroids(1)),file,writeMode="a")
set("thresholdValue",0)
FITS.Write(controlCorba.decode(c.obj.AverageCentroids(1)),file,writeMode="a")
set("thresholdAlgorithm",1)
FITS.Write(controlCorba.decode(c.obj.AverageCentroids(1)),file,writeMode="a")
set("thresholdAlgorithm",2)
"nacts":nacts,
"ncam":ncam,
"nsub":nsub,
"npxly":npxly,
"npxlx":npxlx,
"ncamThreads":ncamThreads,
"pxlCnt":pxlCnt,
"subapLocation":subapLocation,
"bgImage":bgImage,
"darkNoise":darkNoise,
"closeLoop":1,
"flatField":flatField,#numpy.random.random((npxls,)).astype("f"),
"thresholdValue":1.0,
"powerFactor":1.,#raise pixel values to this power.
"subapFlag":subapFlag,
"fakeCCDImage":FITS.Read("/home/ali/pyr.fits")[1],#None,#(numpy.random.random((npxls,))*20).astype("f"),
"printTime":0,#whether to print time/Hz
"rmx":rmx,#numpy.random.random((nacts,ncents)).astype("f"),
"gain":numpy.ones((nacts,),"f"),
"E":numpy.zeros((nacts,nacts),"f"),#E from the tomoalgo in openloop.
"threadAffinity":None,
"threadPriority":numpy.ones((ncamThreads.sum()+1,),numpy.int32)*10,
"delay":10000,#will usually be zero (except if you want to set your own frame rate, e.g. if loading images from a file)
"clearErrors":0,
"camerasOpen":0,
"cameraName":"libcamfile.so",#"libsl240Int32cam.so",#"camfile",
"cameraParams":cameraParams,
"mirrorName":"libmirrorSL240.so",
"mirrorParams":mirrorParams,
"mirrorOpen":0,
"frameno":0,
dmDescription[0] = 1 #1 DM
dmDescription[1] = 8 #1st DM has nacts linear actuators
tmp = dmDescription[2:]
tmp[:] = -1
tmp.shape = 8, 8
dmflag = tel.Pupil(8, 4, 0).fn.ravel()
numpy.put(tmp, dmflag.nonzero()[0], numpy.arange(52))
corrClip = 1
corrNStore = 20
corrThresh = 0.0
corrThreshType = 2 #CORR_FRAC_SUB
corrUpdateGain = 0.
centroidMode = "CorrelationCoG"
psf = FITS.Read(fname.strip("\0"))[1].mean(0).ravel() - bgImage
psf = numpy.where(psf < 0, 0, psf)
import correlation
corrdata = correlation.transformPSF(
psf, ncam, npxlx, npxly, nsub, subapLocation, subapFlag, pad=8
) #Note - to get identical slope estimates when doing auto correlation update and using the correlation result for shift and add (corrUpdateToCoG=0), probably need pad to be larger, to avoid wrapping. But in practice, the difference in results is small.
corrFFTPattern = corrdata["corrFFTPattern"]
corrSubapLoc = corrdata["corrSubapLoc"]
corrNpxlx = corrdata["corrNpxlx"]
corrNpxlCum = corrdata["corrNpxlCum"]
#Now populate the control structure - this is what gets used.
control = {
"switchRequested":
if rmx.shape[1] != gainmx.shape[0]:
transpose = 1
rmx = rmx.T
gainmx = numpy.dot(rmx, gainmx)
if transpose:
gainmx = gainmx.T
return gainmx
if __name__ == "__main__":
import sys
import FITS
if len(sys.argv) < 5:
print(
"Usage: %s globalGain [modal,gain,list] output.fits slopeToZernMx.fits optionalRmx"
% sys.argv[0])
sys.exit(0)
gain = sys.argv[1]
modalGainList = eval(sys.argv[2])
outMx = sys.argv[3]
slopeToZernMx = FITS.Read(sys.argv[4])[1]
if slopeToZernMx[0].std() == 0: #first row all zeros - probably piston
print("Stripping piston from %s" % sys.argv[4])
slopeToZernMx = slopeToZernMx[1:]
rmx = None
if len(sys.argv) > 5:
rmx = FITS.Read(sys.argv[5])[1]
gmx = computeModalGainMx(gain, modalGainList, slopeToZernMx, rmx)
FITS.Write(gmx, outMx)
