#!/usr/bin/python -tt # # Werner Lustermann, Dominik Neise # ETH Zurich, TU Dortmund # from ctypes import * import numpy as np from scipy import signal # get the ROOT stuff + my shared libs from ROOT import gSystem # pyfits_h.so is made from pyfits.h and is used to access the data # make sure the location of pyfits_h.so is in LD_LIBRARY_PATH. # having it in PYTHONPATH is *not* sufficient gSystem.Load('pyfits_h.so') from ROOT import * class RawDataFeeder( object ): """ Wrapper class for RawData class capable of iterating over multiple RawData Files """ def __init__(self, filelist): """ *filelist* list of files to iterate over the list should contain tuples, or sublists of two filenames the first should be a data file (\*.fits.gz) the second should be an amplitude calibration file(\*.drs.fits.gz) """ # sanity check for input if type(filelist) != type(list()): raise TypeError('filelist should be a list') for entry in filelist: if len(entry) != 2: raise TypeError('the entries of filelist should have length == 2') for path in entry: if type(path) != type(str()): raise TypeError('the entries of filelist should be path, i.e. of type str()') #todo check if 'path' is a valid path # else: throw an Exception, or Warning? self.filelist = filelist self._current_RawData = RawData(filelist[0][0], filelist[0][1], return_dict=True) del filelist[0] def __iter__(self): return self def next(): """ Method being called by the iterator. Since the RawData Objects are simply looped over, the event_id from the RawData object will not be unique. Each RawData obejct will start with event_id = 1 as usual. """ try: return self._current_RawData.next() except StopIteration: # current_RawData was completely processed # delete it (I hope this calls the destructor of the fits file and/or closes it) del self._current_RawData # and remake it, if possible if len(self.filelist) > 0: self._current_RawData = RawData(filelist[0][0], filelist[0][1], return_dict=True) del filelist[0] else: raise class RawData( object ): """ raw data access and calibration - open raw data file and drs calibration file - performs amplitude calibration - performs baseline substraction if wanted - provides all data in an array: row = number of pixel col = length of region of interest """ def __init__(self, data_file_name, calib_file_name, user_action_calib=lambda acal_data, data, blm, tom, gm, scells, nroi: None, baseline_file_name='', return_dict = None, do_calibration = True): """ initialize object open data file and calibration data file get basic information about the data in data_file_name allocate buffers for data access data_file_name : fits or fits.gz file of the data including the path calib_file_name : fits or fits.gz file containing DRS calibration data baseline_file_name : npy file containing the baseline values """ self.__module__='pyfact' # manual implementation of default value, but I need to find out # if the user of this class is aware of the new option if return_dict == None: print 'Warning: Rawdata.__init__() has a new option "return_dict"' print 'the default value of this option is False, so nothing changes for you at the moment.' print print 'you probably want, to get a dictionary out of the next() method anyway' print ' so please change your scripts and set this option to True, for the moment' print 'e.g. like this: run = RawData(data_filename, calib_filename, return_dict = True)' print "after a while, the default value, will turn to True .. so you don't have to give the option anymore" print 'and some time later, the option will not be supported anymore' return_dict = False self.return_dict = return_dict self.do_calibration = do_calibration self.data_file_name = data_file_name self.calib_file_name = calib_file_name self.baseline_file_name = baseline_file_name self.user_action_calib = user_action_calib # baseline correction: True / False if len(baseline_file_name) == 0: self.correct_baseline = False else: self.correct_baseline = True # access data file try: data_file = fits(self.data_file_name) except IOError: print 'problem accessing data file: ', data_file_name raise # stop ! no data #: data file (fits object) self.data_file = data_file # get basic information about the data file #: region of interest (number of DRS slices read) self.nroi = data_file.GetUInt('NROI') #: number of pixels (should be 1440) self.npix = data_file.GetUInt('NPIX') #: number of events in the data run self.nevents = data_file.GetNumRows() # allocate the data memories self.event_id = c_ulong() self.trigger_type = c_ushort() #: 1D array with raw data self.data = np.zeros( self.npix * self.nroi, np.int16 ).reshape(self.npix ,self.nroi) #: slice where drs readout started self.start_cells = np.zeros( self.npix, np.int16 ) #: time when the FAD was triggered, in some strange units... self.board_times = np.zeros( 40, np.int32 ) # set the pointers to the data++ data_file.SetPtrAddress('EventNum', self.event_id) data_file.SetPtrAddress('TriggerType', self.trigger_type) data_file.SetPtrAddress('StartCellData', self.start_cells) data_file.SetPtrAddress('Data', self.data) data_file.SetPtrAddress('BoardTime', self.board_times) # open the calibration file try: calib_file = fits(self.calib_file_name) except IOError: print 'problem accessing calibration file: ', calib_file_name raise #: drs calibration file self.calib_file = calib_file baseline_mean = calib_file.GetN('BaselineMean') gain_mean = calib_file.GetN('GainMean') trigger_offset_mean = calib_file.GetN('TriggerOffsetMean') self.Nblm = baseline_mean / self.npix self.Ngm = gain_mean / self.npix self.Ntom = trigger_offset_mean / self.npix self.blm = np.zeros(baseline_mean, np.float32).reshape(self.npix , self.Nblm) self.gm = np.zeros(gain_mean, np.float32).reshape(self.npix , self.Ngm) self.tom = np.zeros(trigger_offset_mean, np.float32).reshape(self.npix , self.Ntom) calib_file.SetPtrAddress('BaselineMean', self.blm) calib_file.SetPtrAddress('GainMean', self.gm) calib_file.SetPtrAddress('TriggerOffsetMean', self.tom) calib_file.GetRow(0) # make calibration constants double, so we never need to roll self.blm = np.hstack((self.blm, self.blm)) self.gm = np.hstack((self.gm, self.gm)) self.tom = np.hstack((self.tom, self.tom)) self.v_bsl = np.zeros(self.npix) # array of baseline values (all ZERO) def __iter__(self): """ iterator """ return self def __add__(self, jump_over): self.data_file.GetRow(jump_over) return self def next(self): """ used by __iter__ """ if self.data_file.GetNextRow() == False: raise StopIteration else: if self.do_calibration == True: self.calibrate_drs_amplitude() #print 'nevents = ', self.nevents, 'event_id = ', self.event_id.value if self.return_dict: return self.__dict__ else: return self.acal_data, self.start_cells, self.trigger_type.value def next_event(self): """ load the next event from disk and calibrate it """ self.data_file.GetNextRow() self.calibrate_drs_amplitude() def calibrate_drs_amplitude(self): """ perform the drs amplitude calibration of the event data """ # shortcuts blm = self.blm gm = self.gm tom = self.tom to_mV = 2000./4096. #: 2D array with amplitude calibrated dat in mV acal_data = self.data * to_mV # convert ADC counts to mV for pixel in range( self.npix ): #shortcuts sc = self.start_cells[pixel] roi = self.nroi # rotate the pixel baseline mean to the Data startCell acal_data[pixel,:] -= blm[pixel,sc:sc+roi] # the 'trigger offset mean' does not need to be rolled # on the contrary, it seems there is an offset in the DRS data, # which is related to its distance to the startCell, not to its # distance to the beginning of the physical pipeline in the DRS chip acal_data[pixel,:] -= tom[pixel,0:roi] # rotate the pixel gain mean to the Data startCell acal_data[pixel,:] /= gm[pixel,sc:sc+roi] self.acal_data = acal_data * 1907.35 self.user_action_calib( self.acal_data, np.reshape(self.data, (self.npix, self.nroi) ), blm, tom, gm, self.start_cells, self.nroi) def baseline_read_values(self, file, bsl_hist='bsl_sum/hplt_mean'): """ open ROOT file with baseline histogram and read baseline values file name of the root file bsl_hist path to the histogram containing the basline values """ try: f = TFile(file) except: print 'Baseline data file could not be read: ', file return h = f.Get(bsl_hist) for i in range(self.npix): self.v_bsl[i] = h.GetBinContent(i+1) f.Close() def baseline_correct(self): """ subtract baseline from the data """ for pixel in range(self.npix): self.acal_data[pixel,:] -= self.v_bsl[pixel] def info(self): """ print run information """ print 'data file: ', data_file_name print 'calib file: ', calib_file_name print 'calibration file' print 'N baseline_mean: ', self.Nblm print 'N gain mean: ', self.Ngm print 'N TriggeroffsetMean: ', self.Ntom # ----------------------------------------------------------------------------- class RawDataFake( object ): """ raw data FAKE access similar to real RawData access """ def __init__(self, data_file_name, calib_file_name, user_action_calib=lambda acal_data, data, blm, tom, gm, scells, nroi: None, baseline_file_name=''): self.__module__='pyfact' self.nroi = 300 self.npix = 9 self.nevents = 1000 self.simulator = None self.time = np.ones(1024) * 0.5 self.event_id = c_ulong(0) self.trigger_type = c_ushort(4) self.data = np.zeros( self.npix * self.nroi, np.int16 ).reshape(self.npix ,self.nroi) self.start_cells = np.zeros( self.npix, np.int16 ) self.board_times = np.zeros( 40, np.int32 ) def __iter__(self): """ iterator """ return self def next(self): """ used by __iter__ """ self.event_id = c_ulong(self.event_id.value + 1) self.board_times = self.board_times + 42 if self.event_id.value >= self.nevents: raise StopIteration else: self._make_event_data() return self.__dict__ def _make_event_data(self): sample_times = self.time.cumsum() - time[0] # random start cell self.start_cells = np.ones( self.npix, np.int16 ) * np.random.randint(0,1024) starttime = self.start_cells[0] signal = self._std_sinus_simu(sample_times, starttime) data = np.vstack( (signal,signal) ) for i in range(8): data = np.vstack( (data,signal) ) self.data = data def _std_sinus_simu(self, times, starttime): period = 10 # in ns # give a jitter on starttime starttime = np.random.normal(startime, 0.05) phase = 0.0 signal = 10 * np.sin(times * 2*np.pi/period + starttime + phase) # add some noise noise = np.random.normal(0.0, 0.5, signal.shape) signal += noise return signal def info(self): """ print run information """ print 'data file: ', data_file_name print 'calib file: ', calib_file_name print 'calibration file' print 'N baseline_mean: ', self.Nblm print 'N gain mean: ', self.Ngm print 'N TriggeroffsetMean: ', self.Ntom # ----------------------------------------------------------------------------- class fnames( object ): """ organize file names of a FACT data run """ def __init__(self, specifier = ['012', '023', '2011', '11', '24'], rpath = '/scratch_nfs/res/bsl/', zipped = True): """ specifier : list of strings defined as: [ 'DRS calibration file', 'Data file', 'YYYY', 'MM', 'DD'] rpath : directory path for the results; YYYYMMDD will be appended to rpath zipped : use zipped (True) or unzipped (Data) """ self.specifier = specifier self.rpath = rpath self.zipped = zipped self.make( self.specifier, self.rpath, self.zipped ) def make( self, specifier, rpath, zipped ): """ create (make) the filenames names : dictionary of filenames, tags { 'data', 'drscal', 'results' } data : name of the data file drscal : name of the drs calibration file results : radikal of file name(s) for results (to be completed by suffixes) """ self.specifier = specifier if zipped: dpath = '/data00/fact-construction/raw/' ext = '.fits.gz' else: dpath = '/data03/fact-construction/raw/' ext = '.fits' year = specifier[2] month = specifier[3] day = specifier[4] yyyymmdd = year + month + day dfile = specifier[1] cfile = specifier[0] rpath = rpath + yyyymmdd + '/' self.rpath = rpath self.names = {} tmp = dpath + year + '/' + month + '/' + day + '/' + yyyymmdd + '_' self.names['data'] = tmp + dfile + ext self.names['drscal'] = tmp + cfile + '.drs' + ext self.names['results'] = rpath + yyyymmdd + '_' + dfile + '_' + cfile self.data = self.names['data'] self.drscal = self.names['drscal'] self.results = self.names['results'] def info( self ): """ print complete filenames """ print 'file names:' print 'data: ', self.names['data'] print 'drs-cal: ', self.names['drscal'] print 'results: ', self.names['results'] # end of class definition: fnames( object ) def _test_iter( nevents ): """ test for function __iter__ """ data_file_name = '/data00/fact-construction/raw/2011/11/24/20111124_117.fits.gz' calib_file_name = '/data00/fact-construction/raw/2011/11/24/20111124_114.drs.fits.gz' # data_file_name = '/home/luster/win7/FACT/data/raw/20120114/20120114_028.fits.gz' # calib_file_name = '/home/luster/win7/FACT/data/raw/20120114/20120114_022.drs.fits.gz' run = RawData( data_file_name, calib_file_name , return_dict=True) for event in run: print 'ev ', event['event_id'].value, 'data[0,0] = ', event['acal_data'][0,0], 'start_cell[0] = ', event['start_cells'][0], 'trigger type = ', event['trigger_type'] if run.event_id.value == nevents: break if __name__ == '__main__': """ tests """ _test_iter(10)