def selectScienceTargets(self):
'''
Based on configuration parameters select a good set of targets to run scheduler on a specified Julian Day.
'''
session = Session()
# [To be done] Reject objects that are close to the moon
for tbin, time in enumerate(self.obsTimeBins):
if self.obsTimeMask[tbin] < 1.0:
# Select objects from database that where not observed and where not scheduled yet
# In the future may include targets that where observed a number of nights ago.
# This is still incomplete. We should also consider the distance from the previous pointing to the next!
# Since a target can have a higher airmass but be farther away from a neaby target that will take less time
# to point.
# one way of selecting targets that are close together and have good airmass is to select regions that are close
# to the current location. it can start as searching an area with r1 ~ 10 x the FoV and, if there are no regions
# to to x2 that and then x4 that. If still there are no targets, than search for the higher in the sky.
targets = session.query(Targets).filter(
Targets.observed == False).filter(
Targets.scheduled == False).filter(
Targets.type == self.sciFlag)
lst = _skysub.lst(time, self.sitelong) #*360./24.
alt = np.array([
_skysub.altit(target.targetDec, lst - target.targetRa,
self.sitelat)[0] for target in targets
])
stg = alt.argmax()
log.info('Selecting %s' % (targets[stg]))
# Marking target as schedule
tst = session.query(Targets).filter(
Targets.id == targets[stg].id)
for t in tst:
t.scheduled = True
session.commit()
self.addObservation(t, time)
self.obsTimeMask[tbin] = 1.0
else:
log.debug(
'Bin %3i @mjd=%.3f already filled up with observations. Skipping...'
% (tbin, time - 2400000.5))
#print i
return 0 #targets
def setJD(self,jd=None):
'''
Configure time domain by specifing a julian day. It will use information on exposure time to build time bins that will be
filled when selecting targets.
'''
if not jd:
site = Site()
jd = np.floor(site.JD())+0.5
nightstart = _skysub.jd_sun_alt(self.sunMaxAlt, jd, self.sitelat, self.sitelong)
nightend = _skysub.jd_sun_alt(self.sunMaxAlt, jd+0.5, self.sitelat, self.sitelong)
self.log.debug('Nigh Start @JD= %.3f # Night End @JD = %.3f'%(nightstart,nightend))
tbin = np.max([np.max(self.sciExpTime),np.max(self.stdExpTime)])*self.nfilters/60./60./24.
self.obsTimeBins = np.arange(nightstart,nightend+tbin,tbin)
self.obsTimeMask = np.zeros(len(self.obsTimeBins))
self.obsTimeMask[-1] = 1.0
# Marking filled bins
session = Session()
scheduled = session.query(Program)
for target in scheduled:
tindex = np.abs(self.obsTimeBins - 2400000.5 - target.slewAt).argmin()
self.obsTimeMask[tindex] = 1.0
def restartAllPrograms(self):
session = Session()
programs = session.query(Program).all()
for program in programs:
program.finished = False
session.commit()
def restartAllPrograms(self):
session = Session()
programs = session.query(Program).all()
for program in programs:
program.finished = False
session.commit()
def selectScienceTargets(self):
'''
Based on configuration parameters select a good set of targets to run scheduler on a specified Julian Day.
'''
session = Session()
# [To be done] Reject objects that are close to the moon
for tbin,time in enumerate(self.obsTimeBins):
if self.obsTimeMask[tbin] < 1.0:
# Select objects from database that where not observed and where not scheduled yet
# In the future may include targets that where observed a number of nights ago.
# This is still incomplete. We should also consider the distance from the previous pointing to the next!
# Since a target can have a higher airmass but be farther away from a neaby target that will take less time
# to point.
# one way of selecting targets that are close together and have good airmass is to select regions that are close
# to the current location. it can start as searching an area with r1 ~ 10 x the FoV and, if there are no regions
# to to x2 that and then x4 that. If still there are no targets, than search for the higher in the sky.
targets = session.query(Targets).filter(Targets.observed == False).filter(Targets.scheduled == False).filter(Targets.type == self.sciFlag)
lst = _skysub.lst(time,self.sitelong) #*360./24.
alt = np.array([_skysub.altit(target.targetDec,lst - target.targetRa,self.sitelat)[0] for target in targets])
stg = alt.argmax()
self.log.info('Selecting %s'%(targets[stg]))
# Marking target as schedule
tst = session.query(Targets).filter(Targets.id == targets[stg].id)
for t in tst:
t.scheduled = True
session.commit()
self.addObservation(t,time)
self.obsTimeMask[tbin] = 1.0
else:
self.log.debug('Bin %3i @mjd=%.3f already filled up with observations. Skipping...'%(tbin,time-2400000.5))
#print i
return 0 #targets
def setJD(self,options):
'''
Configure time domain by specifing a julian day. It will use information on exposure time to build time bins that will be
filled when selecting targets.
'''
sysconfig = SystemConfig.fromFile(options.config)
manager = Manager()
site = manager.getProxy(sysconfig.sites[0])
jd = options.JD
if not jd:
jd = np.floor(site.JD())+0.5
lat = np.array([float(tt) / 60.**i for i,tt in enumerate(str(site['latitude']).split(':'))])
lat[1:] *= lat[0]/np.abs(lat[0])
sitelat = np.sum(lat)
long = np.array([float(tt) / 60.**i for i,tt in enumerate(str(site['longitude']).split(':'))])
long[1:] *= long[0]/np.abs(long[0])
sitelong = abs(np.sum(long)/360*24.)
print site['latitude'],'->',sitelat
print site['longitude'],'->',sitelong
nightstart = _skysub.jd_sun_alt(self.sunMaxAlt, jd, sitelat, sitelong)
nightend = _skysub.jd_sun_alt(self.sunMaxAlt, jd+0.5, sitelat, sitelong)
print('Nigh Start @JD= %.3f # Night End @JD = %.3f'%(nightstart,nightend))
# Creating a 1 minute time bin
tbin = self.tbin
self.obsTimeBins = np.arange(nightstart,nightend+tbin,tbin)
self.obsTimeMask = np.zeros(len(self.obsTimeBins))
self.obsTimeMask[-1] = 1.0
# Marking filled bins
session = Session()
scheduled = session.query(Program)
for target in scheduled:
tindex = np.abs(self.obsTimeBins - 2400000.5 - target.slewAt).argmin()
self.obsTimeMask[tindex] = 1.0
self.isJD = True
def next (self):
if self.rq.empty():
session = Session()
programs = session.query(Program).all()
for program in programs:
program.finished = False
session.commit()
self.reschedule(self.machine)
if not self.rq.empty():
return self.rq.get()
return None
def next(self):
if self.rq.empty():
session = Session()
programs = session.query(Program).all()
for program in programs:
program.finished = False
session.commit()
self.reschedule(self.machine)
if not self.rq.empty():
return self.rq.get()
return None
def reschedule(self, machine):
self.machine = machine
self.rq = Queue(-1)
session = Session()
programs = session.query(Program).order_by(desc(Program.priority)).filter(Program.finished == False).all()
if not programs:
return
log.debug("rescheduling, found %d runnable programs" % len(list(programs)))
for program in programs:
self.rq.put(program)
machine.wakeup()
def reschedule (self, machine):
self.machine = machine
self.rq = Queue(-1)
session = Session()
programs = session.query(Program).order_by(desc(Program.priority)).filter(Program.finished == False).all()
if not programs:
return
log.debug("rescheduling, found %d runnable programs" % len(list(programs)))
for program in programs:
self.rq.put(program)
machine.wakeup()
def setJD(self, jd=None):
'''
Configure time domain by specifing a julian day. It will use information on exposure time to build time bins that will be
filled when selecting targets.
'''
if not jd:
site = Site()
jd = np.floor(site.JD()) + 0.5
nightstart = _skysub.jd_sun_alt(self.sunMaxAlt, jd, self.sitelat,
self.sitelong)
nightend = _skysub.jd_sun_alt(self.sunMaxAlt, jd + 0.5, self.sitelat,
self.sitelong)
log.debug('Nigh Start @JD= %.3f # Night End @JD = %.3f' %
(nightstart, nightend))
tbin = np.max([np.max(self.sciExpTime),
np.max(self.stdExpTime)
]) * self.nfilters / 60. / 60. / 24.
self.obsTimeBins = np.arange(nightstart, nightend + tbin, tbin)
self.obsTimeMask = np.zeros(len(self.obsTimeBins))
self.obsTimeMask[-1] = 1.0
# Marking filled bins
session = Session()
scheduled = session.query(Program)
for target in scheduled:
tindex = np.abs(self.obsTimeBins - 2400000.5 -
target.slewAt).argmin()
self.obsTimeMask[tindex] = 1.0
self.isJD = True
def selectStandardTargets(self,flag,nstars=3,nairmass=3):
'''
Based on configuration parameters, select 'nstars' standard stars to run scheduler on a specified Julian Day. Ideally you
will select standard stars before your science targets so not to have a full queue. Usually standard stars are observed
more than once a night at different airmasses. The user can control this parameter with nairmass and the script will try
to take care of the rest.
'''
session = Session()
# query project information
projQuery = session.query(Projects).filter(Projects.flag == flag)
totobstime = 0.
# Calculate total observation time
for block in projQuery:
totobstime += block.exptime
totobstime /= 86400.0
# First of all, standard stars can be observed multiple times in sucessive nights. I will mark all
# stars as unscheduled.
targets = session.query(Targets).filter(Targets.scheduled == True).filter(Targets.type == flag)
for target in targets:
target.scheduled = False
session.commit()
# [To be done] Reject objects that are close to the moon
# [To be done] Apply all sorts of rejections
# Selecting standard stars is not only searching for the higher in that time but select stars than can be observed at 3
# or more (nairmass) different airmasses. It is also important to select stars with different colors (but this will be
# taken care in the future).
if nairmass*nstars > len(self.obsTimeBins):
self.log.warning('Requesting more stars/observations than it will be possible to schedule. Decreasing number of requests to fit in the night.')
nstars = len(self.obsTimeBins)/nairmass
# Build a grid of desired times for higher airmass observation of each standard star.
stdObsTimeBin = np.arange(10,len(self.obsTimeBins)-10,(len(self.obsTimeBins)-10)/nstars)
obsStandars = np.zeros(nstars)
print stdObsTimeBin
# selecting the closest bin without observation
stdObsTimeBin,status = self.findSuitableTimeBin(stdObsTimeBin)
if status != 0:
raise Exception('Could not find suitable time to start observations! Try cleaning queue.')
print stdObsTimeBin
site = Site()
calclst = lambda time: np.sum(np.array([float(tt) / 60.**i for i,tt in enumerate(str(site._getEphem(datetimeFromJD(time)).sidereal_time()).split(':'))]))
nightlst = np.array([calclst(obstime) for obstime in self.obsTimeBins])
for i,tbin in enumerate(stdObsTimeBin):
# selecting the closest bin without observation
closestcleanbin = tbin
while self.obsTimeMask[closestcleanbin] > 0.0:
closestcleanbin += 1
if i+1 < len(stdObsTimeBin):
if closestcleanbin > stdObsTimeBin[i+1]:
raise Exception('Could not find suitable place to start observations of standard star. Try cleaning queue.')
time = self.obsTimeBins[closestcleanbin]
# 1 - Select objects from database that where not scheduled yet (standard stars may be repited)
# that fits our observing night
#targetSched = False
# Will try until a good match is obtained
#while( not targetSched ):
targets = session.query(Targets).filter(Targets.scheduled == 0).filter(Targets.type == flag)
if len(targets[:]) > 0:
#ephem = site._getEphem(datetimeFromJD(time))
lst = calclst(time) #np.sum(np.array([float(tt) / 60.**i for i,tt in enumerate(str(ephem.sidereal_time()).split(':'))]))
sitelat = np.sum(np.array([float(tt) / 60.**i for i,tt in enumerate(str(site['latitude']).split(':'))]))
alt = np.array([_skysub.altit(target.targetDec,lst - target.targetRa,sitelat)[0] for target in targets])
stg = alt.argmax()
print('Selecting %s'%(targets[stg]))
# Marking target as schedule
tst = session.query(Targets).filter(Targets.id == targets[stg].id)
# Build airmass table for object
objsecz = np.array([_skysub.true_airmass(_skysub.secant_z(_skysub.altit(targets[stg].targetDec,nlst - targets[stg].targetRa,sitelat)[0])) for nlst in nightlst])
# Build desired airmass table
#obsairmass = np.linspace(_skysub.true_airmass(_skysub.secant_z(alt[stg])),projQuery[0].maxairmass,nairmass)
obsairmass = np.logspace(np.log10(np.min(objsecz[objsecz > 0])),np.log10(projQuery[0].maxairmass),nairmass)
np.savetxt('airmass_%04i.dat'%(stg),X=zip(self.obsTimeBins,objsecz))
# Build mask with scheduled airmasses
#mask = np.zeros(len(objsecz),dtype=bool) == 1
pltobstime,pltobsairmass = np.array([]),np.array([])
# Try scheduling observations on all airmasses
for airmass in obsairmass:
# Get times where the object is close to the desired airmass and there are no observations scheduled
timeobsmask = np.bitwise_and(self.obsTimeMask < 1.0,np.abs(objsecz - airmass) < self.tolairmass)
# Check that there are times available
if not timeobsmask.any():
#raise Exception('No time available for scheduling observations of standard star %s at airmass %.3f'%(targets[stg],airmass))
self.log.warning('No time available for scheduling observations of standard star %s at airmass %.3f'%(targets[stg],airmass))
# Start trying to schedule observations
indexes = np.arange(len(self.obsTimeMask))[timeobsmask] #np.bitwise_and(self.obsTimeMask, timeobsmask)
obsSched = False
for index in indexes:
print('[%.3f] - Time bin available for observation of standard star at airmass %.3f'%(self.obsTimeBins[index], airmass))
print '- Require %i extra time bins'%(totobstime/self.tbin)
if (self.obsTimeMask[index:index+totobstime/self.tbin] < 1.0).all():
print 'Observation fit in this block.'
self.obsTimeMask[index:index+totobstime/self.tbin] = 1.0
self.log.info('Requesting observations of %s @airmass=%4.2f @mjd=%.3f...'%(target.name,airmass,self.obsTimeBins[index]-2400000.5))
pltobstime = np.append(pltobstime,self.obsTimeBins[index:index+totobstime/self.tbin])
pltobsairmass = np.append(pltobsairmass, objsecz[index:index+totobstime/self.tbin])
#for nblock,ii in enumerate(range(index,int(index+totobstime/self.tbin),1)):
self.addObservation(targets[stg],self.obsTimeBins[index],projQuery)
break
np.savetxt('obsairmass_%04i.dat'%stg,X = zip(pltobstime,pltobsairmass))
#self.obsTimeMask[index] = 1.0
#for iobsbins in range(index+1,index+int(totobstime/self.tbin)):
#print '[%i] - require extra time bin'%(iobsbins)
#if self.obsTimeMask[iobsbins] < 1.0:
# self.obsTimeMask[iobsbins] = 1.0
#else:
# raise Exception('Time bin [%i/%i] not available for observation of standard star at airmass %.3f'%(iobsbins,len(self.obsTimeMask),airmass))
#else:
#raise Exception('Time bin not available for observation of standard star at airmass %.3f'%(airmass))
for t in tst:
t.scheduled = True
session.commit()
obsStandars[i] = t.id
else:
self.log.warning('No suitable standard star for jd:%.3f in database...'%(time))
return 0
return 0
if len(obsStandars[obsStandars >= 0]) < nstars:
self.log.warning('Could not find %i suitable standard stars in catalog. Only %i where found.'%(nstars,len(obsStandars[obsStandars >= 0])))
obsStandars = np.zeros(len(self.obsTimeBins))-1 # first selection of observable standards
for tbin,time in enumerate(self.obsTimeBins):
if self.obsTimeMask[tbin] < 1.0:
# 1 - Select objects from database that where not scheduled yet (standard stars may be repited)
# that fits our observing night
targets = session.query(Targets).filter(Targets.scheduled == 0).filter(Targets.type == flag)
if len(targets[:]) > 0:
ephem = site._getEphem(datetimeFromJD(time))
lst = np.sum(np.array([float(tt) / 60.**i for i,tt in enumerate(str(ephem.sidereal_time()).split(':'))]))
sitelat = np.sum(np.array([float(tt) / 60.**i for i,tt in enumerate(str(site['latitude']).split(':'))]))
secz = np.array([_skysub.secant_z(_skysub.altit(target.targetDec,lst - target.targetRa,sitelat)[0]) for target in targets])
stg = secz.argmax()
self.log.info('Selecting %s'%(targets[stg]))
# Marking target as schedule
tst = session.query(Targets).filter(Targets.id == targets[stg].id)
for t in tst:
t.scheduled = True
session.commit()
obsStandars[tbin] = t.id
else:
print('No suitable target for jd:%.3f in database...'%(time))
break
else:
self.log.info('Bin already filled up with observations. Skipping...')
if len(obsStandars[obsStandars >= 0]) < nstars:
self.log.warning('Could not find %i suitable standard stars in catalog. Only %i where found.'%(nstars,len(obsStandars[obsStandars >= 0])))
#
# Unmarking potential targets as scheduled
#
for id in obsStandars[obsStandars >= 0]:
target = session.query(Targets).filter(Targets.id == id)
for t in target:
t.scheduled = False
session.commit()
tbin+=1
#
# Preparing a grid of altitudes for each target for each observing window
#
amGrid = np.zeros(len(obsStandars)*len(obsStandars)).reshape(len(obsStandars),len(obsStandars))
for i in np.arange(len(obsStandars))[obsStandars >= 0]:
target = session.query(Targets).filter(Targets.id == obsStandars[i])[0]
for j in range(len(obsStandars)):
lst = _skysub.lst(self.obsTimeBins[j],self.sitelong)
amGrid[i][j] = _skysub.true_airmass(_skysub.secant_z(_skysub.altit(target.targetDec,lst - target.targetRa,self.sitelat)[0]))
if amGrid[i][j] < 0:
amGrid [i][j] = 99.
#
# Build a grid mask that specifies the position in time each target should be observed. This means that, when
# selecting a single target we ocuppy more than one, non consecutive, position in the night. This grid shows where are these
# positions.
#
obsMask = np.zeros(len(obsStandars)*len(obsStandars),dtype=np.bool).reshape(len(obsStandars),len(obsStandars))
for i in np.arange(len(obsStandars))[obsStandars >= 0]:
amObs = np.linspace(amGrid[i].min(),self.stdMaxAirmass,nairmass) # requested aimasses
dam = np.mean(np.abs(amGrid[i][amGrid[i]<self.stdMaxAirmass][1:] - amGrid[i][amGrid[i]<self.stdMaxAirmass][:-1])) # how much airmass changes in average
for j,am in enumerate(amObs):
# Mark positions where target is at specified airmass
if j == 0:
obsMask[i] = np.bitwise_or(obsMask[i],amGrid[i] == am)
else:
obsMask[i] = np.bitwise_or(obsMask[i],np.bitwise_and(amGrid[i]>am-dam,amGrid[i]<am+dam))
#print amGrid[i][np.where(obsMask[i])]
#
# Now it is time to actually select the targets. It will start with the first target and then try the others
# until it find enough standard stars, as specified by the user.
#
# Para cada bin em tempo, varro o bin em massa de ar por coisas observaveis. Se acho um, vejo se posso agendar
# os outros bins. Se sim, marco o alvo para observacao, se nao, passo para o proximo. Repito ate completar a
# lista de alvos
#
obsMaskTimeGrid = np.zeros(len(obsStandars),dtype=np.bool)
nrequests = 0
reqId = np.zeros(nstars,dtype=np.int)-1
for tbin,time in enumerate(self.obsTimeBins[:-1]):
# Evaluates if time slots are all available. If yes, mark orbservation and ocuppy slots.
if ( (not obsMaskTimeGrid[obsMask[tbin]].any()) and (len(amGrid[tbin][obsMask[tbin]])>=nairmass) ):
obsMaskTimeGrid = np.bitwise_or(obsMaskTimeGrid,obsMask[tbin])
reqId[nrequests] = tbin
nrequests += 1
if nrequests >= nstars:
break
# Finally, requesting observations
for id in reqId[reqId >= 0]:
target = session.query(Targets).filter(Targets.id == obsStandars[id])[0]
secz = amGrid[id][obsMask[id]]
seczreq = np.zeros(nairmass,dtype=np.bool)
amObs = np.linspace(amGrid[id].min(),self.stdMaxAirmass,nairmass) # requested aimasses
for i,obstime in enumerate(self.obsTimeBins[obsMask[id]]):
sindex = np.abs(amObs-secz[i]).argmin()
if not seczreq[sindex]:
self.log.info('Requesting observations of %s @airmass=%4.2f @mjd=%.3f...'%(target.name,secz[i],obstime-2400000.5))
seczreq[sindex] = True
target.scheduled = True
session.commit()
self.addObservation(target,obstime)
self.obsTimeMask[obsMask[id]] = 1.0
#print self.obsTimeBins[obsMask[id]]
#print
#print i
return 0 #targets
def targets(self):
'''
After selecting targets, you can generate a list of potential targets to run the scheduler.
'''
session = Session()
fp1 = open(os.path.join(self.PATH,'Fixed.txt'),'w')
fp2 = open(os.path.join(self.PATH,'request.stg'),'w')
# Write header
fp1.write('''P|Designation | RA | dec |mag.
-|------------|hh mm ss.ss|sdd mm ss.s|nn.n
''')
config = { 'name' : '',
'user' : '',
'nimages' : 1,
'expt' : 0,
'filter' : '',
'time' : ''}
for obstype in [self.stdFlag,self.sciFlag]:
targets = session.query(Targets,Program).join((Program,Targets.id==Program.tid)).filter(Targets.type == obstype).order_by(Targets.name)
tobs = []
tname = ''
FlagFilterClear = True
for target,program in targets:
p = Position.fromRaDec(target.targetRa,target.targetDec)
ra = p.ra.HMS
dec = p.dec.DMS
#
# Write Fixed.txt
#
objname = '%12s'%(target.name).replace(' ','_')
objname = objname.replace(' ','_')
fp1.write('%1s %s %02.0f %02.0f %05.2f %+03.0f %02.0f %04.1f %04.1f\n'%( target.type,
objname,
ra[1],
ra[2],
ra[3],
dec[0]*dec[1],
dec[2],
dec[3],target.targetMag))
#
# Write stg file with observation requests
#
config['name'] = objname
config['user'] = self.stdUser
filterExpt = self.stdExpTime
if target.type == self.sciFlag:
config['user'] = self.sciUser
filterExpt = self.sciExpTime
config['time'] = ''
dt = np.max(filterExpt)*self.nfilters/60./60./24.
if target.type ==self.stdFlag:
FlagFilterClear = True
tname = target.name
tstart = datetimeFromJD(program.slewAt + 2400000.5)
tend = datetimeFromJD(program.slewAt+dt + 2400000.5)
config['time'] = 't>%s t<%s'%(tstart.strftime('%y%m%d-%H:%M'),tend.strftime('%y%m%d-%H:%M'))
for i in range(self.nfilters):
config['expt'] = filterExpt[i]
config['filter'] = self.filters[i]
tstart = datetimeFromJD(program.slewAt + 2400000.5)
tend = datetimeFromJD(program.slewAt+dt*1.1 + 2400000.5)
fp2.write('%(name)12s; %(user)s %(nimages)ii exp=%(expt).2f opt filter=%(filter)s %(time)s\n'%config)
fp1.close()
fp2.close()
return 0
def selectStandardTargets(self,nstars=3,nairmass=3):
'''
Based on configuration parameters, select 'nstars' standard stars to run scheduler on a specified Julian Day. Ideally you
will select standard stars before your science targets so not to have a full queue. Usually standard stars are observed
more than once a night at different airmasses. The user can control this parameter with nairmass and the script will try
to take care of the rest.
'''
session = Session()
# First of all, standard stars can be obsered multiple times in sucessive nights. I will mark all
# stars an unscheduled.
targets = session.query(Targets).filter(Targets.scheduled == True).filter(Targets.type == self.stdFlag)
for target in targets:
target.scheduled = False
session.commit()
# [To be done] Reject objects that are close to the moon
# Selecting standard stars is not only searching for the higher in that time but select stars than can be observed at 3
# or more (nairmass) different airmasses. It is also important to select stars with different colors (but this will be
# taken care in the future).
if nairmass*nstars > len(self.obsTimeBins):
self.log.warning('Requesting more stars/observations than it will be possible to schedule. Decreasing number of requests to fit in the night.')
nstars = len(self.obsTimeBins)/nairmass
obsStandars = np.zeros(len(self.obsTimeBins))-1 # first selection of observable standards
for tbin,time in enumerate(self.obsTimeBins):
if self.obsTimeMask[tbin] < 1.0:
# 1 - Select objects from database that where not scheduled yet (standard stars may be repited)
# that fits our observing night
targets = session.query(Targets).filter(Targets.scheduled == 0).filter(Targets.type == self.stdFlag)
lst = _skysub.lst(time,self.sitelong) #*360./24.
alt = np.array([_skysub.altit(target.targetDec,lst - target.targetRa,self.sitelat)[0] for target in targets])
stg = alt.argmax()
self.log.info('Selecting %s'%(targets[stg]))
# Marking target as schedule
tst = session.query(Targets).filter(Targets.id == targets[stg].id)
for t in tst:
t.scheduled = True
session.commit()
obsStandars[tbin] = t.id
else:
self.log.info('Bin already filled up with observations. Skipping...')
if len(obsStandars[obsStandars >= 0]) < nstars:
self.log.warning('Could not find %i suitable standard stars in catalog. Only %i where found.'%(nstars,len(obsStandars[obsStandars >= 0])))
#
# Unmarking potential targets as scheduled
#
for id in obsStandars[obsStandars >= 0]:
target = session.query(Targets).filter(Targets.id == id)
for t in target:
t.scheduled = False
session.commit()
tbin+=1
#
# Preparing a grid of altitudes for each target for each observing window
#
amGrid = np.zeros(len(obsStandars)*len(obsStandars)).reshape(len(obsStandars),len(obsStandars))
for i in np.arange(len(obsStandars))[obsStandars >= 0]:
target = session.query(Targets).filter(Targets.id == obsStandars[i])[0]
for j in range(len(obsStandars)):
lst = _skysub.lst(self.obsTimeBins[j],self.sitelong)
amGrid[i][j] = _skysub.true_airmass(_skysub.secant_z(_skysub.altit(target.targetDec,lst - target.targetRa,self.sitelat)[0]))
if amGrid[i][j] < 0:
amGrid [i][j] = 99.
#
# Build a grid mask that specifies the position in time each target should be observed. This means that, when
# selecting a single target we ocuppy more than one, non consecutive, position in the night. This grid shows where are these
# positions.
#
obsMask = np.zeros(len(obsStandars)*len(obsStandars),dtype=np.bool).reshape(len(obsStandars),len(obsStandars))
for i in np.arange(len(obsStandars))[obsStandars >= 0]:
amObs = np.linspace(amGrid[i].min(),self.stdMaxAirmass,nairmass) # requested aimasses
dam = np.mean(np.abs(amGrid[i][amGrid[i]<self.stdMaxAirmass][1:] - amGrid[i][amGrid[i]<self.stdMaxAirmass][:-1])) # how much airmass changes in average
for j,am in enumerate(amObs):
# Mark positions where target is at specified airmass
if j == 0:
obsMask[i] = np.bitwise_or(obsMask[i],amGrid[i] == am)
else:
obsMask[i] = np.bitwise_or(obsMask[i],np.bitwise_and(amGrid[i]>am-dam,amGrid[i]<am+dam))
#print amGrid[i][np.where(obsMask[i])]
#
# Now it is time to actually select the targets. It will start with the first target and then try the others
# until it find enough standard stars, as specified by the user.
#
# Para cada bin em tempo, varro o bin em massa de ar por coisas observaveis. Se acho um, vejo se posso agendar
# os outros bins. Se sim, marco o alvo para observacao, se nao, passo para o proximo. Repito ate completar a
# lista de alvos
#
obsMaskTimeGrid = np.zeros(len(obsStandars),dtype=np.bool)
nrequests = 0
reqId = np.zeros(nstars,dtype=np.int)-1
for tbin,time in enumerate(self.obsTimeBins[:-1]):
# Evaluates if time slots are all available. If yes, mark orbservation and ocuppy slots.
if ( (not obsMaskTimeGrid[obsMask[tbin]].any()) and (len(amGrid[tbin][obsMask[tbin]])>=nairmass) ):
obsMaskTimeGrid = np.bitwise_or(obsMaskTimeGrid,obsMask[tbin])
reqId[nrequests] = tbin
nrequests += 1
if nrequests >= nstars:
break
# Finally, requesting observations
for id in reqId[reqId >= 0]:
target = session.query(Targets).filter(Targets.id == obsStandars[id])[0]
secz = amGrid[id][obsMask[id]]
seczreq = np.zeros(nairmass,dtype=np.bool)
amObs = np.linspace(amGrid[id].min(),self.stdMaxAirmass,nairmass) # requested aimasses
for i,obstime in enumerate(self.obsTimeBins[obsMask[id]]):
sindex = np.abs(amObs-secz[i]).argmin()
if not seczreq[sindex]:
self.log.info('Requesting observations of %s @airmass=%4.2f @mjd=%.3f...'%(target.name,secz[i],obstime-2400000.5))
seczreq[sindex] = True
target.scheduled = True
session.commit()
self.addObservation(target,obstime)
self.obsTimeMask[obsMask[id]] = 1.0
#print self.obsTimeBins[obsMask[id]]
#print
#print i
return 0 #targets
def process(check):
import yaml
from chimera.util.position import Position
from chimera.util.coord import Coord
from chimera.controllers.scheduler.model import (Session, Program,
AutoFocus, AutoFlat,
PointVerify, Point,
Expose)
actionDict = {
'autofocus': AutoFocus,
'autoflat': AutoFlat,
'pointverify': PointVerify,
'point': Point,
'expose': Expose,
}
manager = BaseResponse.manager
# sched = ConfigureScheduler.scheduler
# delete all programs
session = Session()
programs = session.query(Program).all()
for program in programs:
session.delete(program)
session.commit()
def generateDatabase(options):
with open(os.path.join(os.path.expanduser('~/'), options.filename),
'r') as stream:
try:
prgconfig = yaml.load(stream)
except yaml.YAMLError as exc:
manager.broadCast(exc)
raise
except Exception, e:
manager.broadCast(
'Exception trying to start scheduler: %s' % repr(e))
raise
def _validateOffset(value):
try:
offset = Coord.fromAS(int(value))
except ValueError:
offset = Coord.fromDMS(value)
return offset
session = Session()
programs = []
for prg in prgconfig['programs']:
# process program
program = Program()
for key in prg.keys():
if hasattr(program, key) and key != 'actions':
try:
setattr(program, key, prg[key])
except:
manager.broadCast(
'Could not set attribute %s = %s on Program' %
(key, prg[key]))
# self.out("# program: %s" % program.name)
# process actions
for actconfig in prg['actions']:
act = actionDict[actconfig['action']]()
# self.out('Action: %s' % actconfig['action'])
if actconfig['action'] == 'point':
if 'ra' in actconfig.keys(
) and 'dec' in actconfig.keys():
epoch = 'J2000' if 'epoch' not in actconfig.keys(
) else actconfig['epoch']
position = Position.fromRaDec(
actconfig['ra'], actconfig['dec'], epoch)
# self.out('Coords: %s' % position)
act.targetRaDec = position
# act = Point(targetRaDec=position)
elif 'alt' in actconfig.keys(
) and 'az' in actconfig.keys():
position = Position.fromAltAz(
actconfig['alt'], actconfig['az'])
# self.out('Coords: %s' % position)
act.targetAltAz = position
elif 'name' in actconfig:
# self.out('Target name: %s' % actconfig['name'])
act.targetName = actconfig['name']
elif 'offset' not in actconfig:
manager.broadCast(
'Empty Point action. No target to point to or offset to perform!'
)
continue
if 'offset' in actconfig:
if 'north' in actconfig['offset']:
offset = _validateOffset(
actconfig['offset']['north'])
act.offsetNS = offset
elif 'south' in actconfig['offset']:
offset = _validateOffset(
actconfig['offset']['south'])
act.offsetNS = Coord.fromAS(-offset.AS)
if 'west' in actconfig['offset']:
offset = _validateOffset(
actconfig['offset']['west'])
act.offsetEW = offset
elif 'east' in actconfig['offset']:
offset = _validateOffset(
actconfig['offset']['east'])
act.offsetEW = Coord.fromAS(-offset.AS)
else:
for key in actconfig.keys():
if hasattr(act, key) and key != 'action':
# self.out('\t%s: %s' % (key,actconfig[key]))
try:
setattr(act, key, actconfig[key])
except:
manager.broadCast(
'Could not set attribute %s = %s on action %s'
% (key, actconfig[key],
actconfig['action']))
program.actions.append(act)
# self.out("")
programs.append(program)
# self.out("List contain %i programs" % len(programs))
session.add_all(programs)
session.commit()
return 0
def process(check):
import yaml
from chimera.util.position import Position
from chimera.controllers.scheduler.model import (Session, Program, AutoFocus, AutoFlat,
PointVerify, Point,
Expose)
actionDict = {'autofocus' : AutoFocus,
'autoflat' : AutoFlat,
'pointverify' : PointVerify,
'point' : Point,
'expose' : Expose,
}
manager = BaseResponse.manager
sched = ConfigureScheduler.scheduler
# delete all programs
session = Session()
programs = session.query(Program).all()
for program in programs:
session.delete(program)
session.commit()
def generateDatabase(options):
with open(os.path.join(os.path.expanduser('~/'),
options.filename), 'r') as stream:
try:
prgconfig = yaml.load(stream)
except yaml.YAMLError as exc:
manager.broadCast(exc)
return -1
session = Session()
programs = []
for prg in prgconfig['programs']:
# process program
program = Program()
for key in prg.keys():
if hasattr(program,key) and key != 'actions':
try:
setattr(program,key,prg[key])
except:
manager.broadCast('Could not set attribute %s = %s on Program' % (key,prg[key]))
# self.out("# program: %s" % program.name)
# process actions
for actconfig in prg['actions']:
act = actionDict[actconfig['action']]()
# self.out('Action: %s' % actconfig['action'])
if actconfig['action'] == 'point':
if 'ra' in actconfig.keys() and 'dec' in actconfig.keys():
epoch = 'J2000' if 'epoch' not in actconfig.keys() else actconfig['epoch']
position = Position.fromRaDec(actconfig['ra'], actconfig['dec'], epoch)
# self.out('Coords: %s' % position)
act.targetRaDec = position
# act = Point(targetRaDec=position)
elif 'alt' in actconfig.keys() and 'az' in actconfig.keys():
position = Position.fromAltAz(actconfig['alt'], actconfig['az'])
# self.out('Coords: %s' % position)
act.targetAltAz = position
else:
# self.out('Target name: %s' % actconfig['name'])
act.targetName = actconfig['name']
else:
for key in actconfig.keys():
if hasattr(act,key) and key != 'action':
# self.out('\t%s: %s' % (key,actconfig[key]))
try:
setattr(act,key,actconfig[key])
except:
manager.broadCast('Could not set attribute %s = %s on action %s' % (key,
actconfig[key],
actconfig['action']))
program.actions.append(act)
# self.out("")
programs.append(program)
# self.out("List contain %i programs" % len(programs))
session.add_all(programs)
session.commit()
return 0
# self.out("Restart the scheduler to run it with the new database.")
if generateDatabase(check) < 0:
manager.broadCast("Could not configure scheduler with provided arguments.")
manager.setFlag("scheduler",
IOFlag.ERROR)
else:
manager.setFlag("scheduler",
IOFlag.READY)
manager.broadCast("Scheduler configured. Restart it to run with the new database.")
def selectStandardTargets(self, nstars=3, nairmass=3):
'''
Based on configuration parameters, select 'nstars' standard stars to run scheduler on a specified Julian Day. Ideally you
will select standard stars before your science targets so not to have a full queue. Usually standard stars are observed
more than once a night at different airmasses. The user can control this parameter with nairmass and the script will try
to take care of the rest.
'''
session = Session()
# First of all, standard stars can be obsered multiple times in sucessive nights. I will mark all
# stars an unscheduled.
targets = session.query(Targets).filter(
Targets.scheduled == True).filter(Targets.type == self.stdFlag)
for target in targets:
target.scheduled = False
session.commit()
# [To be done] Reject objects that are close to the moon
# Selecting standard stars is not only searching for the higher in that time but select stars than can be observed at 3
# or more (nairmass) different airmasses. It is also important to select stars with different colors (but this will be
# taken care in the future).
if nairmass * nstars > len(self.obsTimeBins):
log.warning(
'Requesting more stars/observations than it will be possible to schedule. Decreasing number of requests to fit in the night.'
)
nstars = len(self.obsTimeBins) / nairmass
obsStandars = np.zeros(len(
self.obsTimeBins)) - 1 # first selection of observable standards
for tbin, time in enumerate(self.obsTimeBins):
if self.obsTimeMask[tbin] < 1.0:
# 1 - Select objects from database that where not scheduled yet (standard stars may be repited)
# that fits our observing night
targets = session.query(Targets).filter(
Targets.scheduled == 0).filter(
Targets.type == self.stdFlag)
lst = _skysub.lst(time, self.sitelong) #*360./24.
alt = np.array([
_skysub.altit(target.targetDec, lst - target.targetRa,
self.sitelat)[0] for target in targets
])
stg = alt.argmax()
log.info('Selecting %s' % (targets[stg]))
# Marking target as schedule
tst = session.query(Targets).filter(
Targets.id == targets[stg].id)
for t in tst:
t.scheduled = True
session.commit()
obsStandars[tbin] = t.id
else:
log.info(
'Bin already filled up with observations. Skipping...')
if len(obsStandars[obsStandars >= 0]) < nstars:
log.warning(
'Could not find %i suitable standard stars in catalog. Only %i where found.'
% (nstars, len(obsStandars[obsStandars >= 0])))
#
# Unmarking potential targets as scheduled
#
for id in obsStandars[obsStandars >= 0]:
target = session.query(Targets).filter(Targets.id == id)
for t in target:
t.scheduled = False
session.commit()
tbin += 1
#
# Preparing a grid of altitudes for each target for each observing window
#
amGrid = np.zeros(len(obsStandars) * len(obsStandars)).reshape(
len(obsStandars), len(obsStandars))
for i in np.arange(len(obsStandars))[obsStandars >= 0]:
target = session.query(Targets).filter(
Targets.id == obsStandars[i])[0]
for j in range(len(obsStandars)):
lst = _skysub.lst(self.obsTimeBins[j], self.sitelong)
amGrid[i][j] = _skysub.true_airmass(
_skysub.secant_z(
_skysub.altit(target.targetDec, lst - target.targetRa,
self.sitelat)[0]))
if amGrid[i][j] < 0:
amGrid[i][j] = 99.
#
# Build a grid mask that specifies the position in time each target should be observed. This means that, when
# selecting a single target we ocuppy more than one, non consecutive, position in the night. This grid shows where are these
# positions.
#
obsMask = np.zeros(len(obsStandars) * len(obsStandars),
dtype=np.bool).reshape(len(obsStandars),
len(obsStandars))
for i in np.arange(len(obsStandars))[obsStandars >= 0]:
amObs = np.linspace(amGrid[i].min(), self.stdMaxAirmass,
nairmass) # requested aimasses
dam = np.mean(
np.abs(amGrid[i][amGrid[i] < self.stdMaxAirmass][1:] -
amGrid[i][amGrid[i] < self.stdMaxAirmass][:-1])
) # how much airmass changes in average
for j, am in enumerate(amObs):
# Mark positions where target is at specified airmass
if j == 0:
obsMask[i] = np.bitwise_or(obsMask[i], amGrid[i] == am)
else:
obsMask[i] = np.bitwise_or(
obsMask[i],
np.bitwise_and(amGrid[i] > am - dam,
amGrid[i] < am + dam))
#print amGrid[i][np.where(obsMask[i])]
#
# Now it is time to actually select the targets. It will start with the first target and then try the others
# until it find enough standard stars, as specified by the user.
#
# Para cada bin em tempo, varro o bin em massa de ar por coisas observaveis. Se acho um, vejo se posso agendar
# os outros bins. Se sim, marco o alvo para observacao, se nao, passo para o proximo. Repito ate completar a
# lista de alvos
#
obsMaskTimeGrid = np.zeros(len(obsStandars), dtype=np.bool)
nrequests = 0
reqId = np.zeros(nstars, dtype=np.int) - 1
for tbin, time in enumerate(self.obsTimeBins[:-1]):
# Evaluates if time slots are all available. If yes, mark orbservation and ocuppy slots.
if ((not obsMaskTimeGrid[obsMask[tbin]].any())
and (len(amGrid[tbin][obsMask[tbin]]) >= nairmass)):
obsMaskTimeGrid = np.bitwise_or(obsMaskTimeGrid, obsMask[tbin])
reqId[nrequests] = tbin
nrequests += 1
if nrequests >= nstars:
break
# Finally, requesting observations
for id in reqId[reqId >= 0]:
target = session.query(Targets).filter(
Targets.id == obsStandars[id])[0]
secz = amGrid[id][obsMask[id]]
seczreq = np.zeros(nairmass, dtype=np.bool)
amObs = np.linspace(amGrid[id].min(), self.stdMaxAirmass,
nairmass) # requested aimasses
for i, obstime in enumerate(self.obsTimeBins[obsMask[id]]):
sindex = np.abs(amObs - secz[i]).argmin()
if not seczreq[sindex]:
log.info(
'Requesting observations of %s @airmass=%4.2f @mjd=%.3f...'
% (target.name, secz[i], obstime - 2400000.5))
seczreq[sindex] = True
target.scheduled = True
session.commit()
self.addObservation(target, obstime)
self.obsTimeMask[obsMask[id]] = 1.0
#print self.obsTimeBins[obsMask[id]]
#print
#print i
return 0 #targets
def targets(self):
'''
After selecting targets, you can generate a list of potential targets to run the scheduler.
'''
import subprocess
session = Session()
request = os.path.join(self.PATH, 'targets/request.stg')
fp1 = open(os.path.join(self.PATH, 'targets/Fixed.txt'), 'w')
fp2 = open(request, 'w')
# Write header
fp1.write('''P|Designation | RA | dec |mag.
-|------------|hh mm ss.ss|sdd mm ss.s|nn.n
''')
config = {
'name': '',
'user': '',
'nimages': 1,
'expt': 0,
'filter': '',
'time': ''
}
for obstype in [self.stdFlag, self.sciFlag]:
targets = session.query(Targets, Program).join(
(Program, Targets.id == Program.tid)).filter(
Targets.type == obstype).order_by(Targets.name)
tobs = []
tname = ''
FlagFilterClear = True
for target, program in targets:
p = Position.fromRaDec(target.targetRa, target.targetDec)
ra = p.ra.HMS
dec = p.dec.DMS
#
# Write Fixed.txt
#
objname = '%12s' % (target.name).replace(' ', '_')
objname = objname.replace(' ', '_')
fp1.write(
'%1s %s %02.0f %02.0f %05.2f %+03.0f %02.0f %04.1f %04.1f\n'
% (target.type, objname, ra[1], ra[2], ra[3],
dec[0] * dec[1], dec[2], dec[3], target.targetMag))
#
# Write stg file with observation requests
#
config['name'] = objname
config['user'] = self.stdUser
filterExpt = self.stdExpTime
if target.type == self.sciFlag:
config['user'] = self.sciUser
filterExpt = self.sciExpTime
config['time'] = ''
dt = np.max(filterExpt) * self.nfilters / 60. / 60. / 24.
if target.type == self.stdFlag:
FlagFilterClear = True
tname = target.name
tstart = datetimeFromJD(program.slewAt + 2400000.5)
tend = datetimeFromJD(program.slewAt + dt + 2400000.5)
config['time'] = 't>%s t<%s' % (tstart.strftime(
'%y%m%d-%H:%M'), tend.strftime('%y%m%d-%H:%M'))
for i in range(self.nfilters):
config['expt'] = filterExpt[i]
config['filter'] = self.filters[i]
tstart = datetimeFromJD(program.slewAt + 2400000.5)
tend = datetimeFromJD(program.slewAt + dt * 1.1 +
2400000.5)
fp2.write(
'%(name)12s; %(user)s %(nimages)ii exp=%(expt).2f opt filter=%(filter)s %(time)s\n'
% config)
fp1.close()
fp2.close()
#
# Calling targets from TAO to generate targets list.
#
fp1 = open(os.path.join(self.PATH, 'targets/targets.log'), 'w')
bin = os.path.expanduser(os.path.join(self.PATH, 'targets/targets'))
runTargets = subprocess.Popen([bin, '-s', request],
stdout=fp1,
stderr=fp1)
runTargets.wait()
fp1.close()
return 0
