def fcs_import_method(fit_obj,file_path):
text =[0]
r_obj = csv.reader(open(file_path, 'rb'),delimiter='\t')
title = r_obj.next()
line = r_obj.next()
line = r_obj.next()
read = True
ind = 0
while read == True:
corrObj = corrObject(file_path,fit_obj);
line = r_obj.next()
text =[]
for part in line:
if part != '':
text.append(part)
while text[0].split(' = ')[0] != 'CountRateArray':
if text[0].split(' = ')[0] == 'Channel':
channel_str = text[0].split(' = ')[1]
if channel_str == 'Auto-correlation detector Meta1':
channel = 0
name = 'CH0'
ind +=1
if channel_str == 'Auto-correlation detector Meta2':
channel = 1
name = 'CH1'
if channel_str == 'Cross-correlation detector Meta2 versus detector Meta1':
channel = 3
name = 'CH10'
if channel_str == 'Cross-correlation detector Meta1 versus detector Meta2':
channel = 2
name = 'CH01'
try:
line = r_obj.next()
except:
read = False
break
text =[]
for part in line:
if part != '':
text.append(part)
if read == False:
break
line = r_obj.next()
dimensions = text[0].split(' = ')[1]
len_of_seq = int(dimensions.split(' ')[0])
if len_of_seq >0:
cdata = []
cscale = []
for v in range(0,len_of_seq):
text =[]
for part in line:
if part != '':
text.append(part)
if text.__len__() >1:
cscale.append(float(text[0]))
cdata.append(float(text[1]))
line = r_obj.next()
#Reads to first correlation array text.
while text[0].split(' = ')[0] != 'CorrelationArray':
try:
line = r_obj.next()
except:
break
read = False
text =[]
for part in line:
if part != '':
text.append(part)
dimensions = text[0].split(' = ')[1]
len_of_seq = int(dimensions.split(' ')[0])
if len_of_seq >0:
tdata = []
tscale = []
line = r_obj.next()
for v in range(0,len_of_seq):
text =[]
for part in line:
if part != '':
text.append(part)
if text.__len__() >1:
tscale.append(float(text[0]))
tdata.append(float(text[1]))
line = r_obj.next()
fit_obj.objIdArr.append(corrObj)
corrObj.name = corrObj.name+'_'+str(ind)+'_'+str(name)
corrObj.parent_name = '.fcs files'
corrObj.parent_uqid = '0'
corrObj.siblings = None
corrObj.autoNorm = np.array(tdata).astype(np.float64).reshape(-1)
corrObj.autotime = np.array(tscale).astype(np.float64).reshape(-1)*1000
#Average counts per bin. For it to be seconds (Hz), we have cscale-1 because the first value corresponds to zero recording time.
unit = cscale[-1]/(cscale.__len__()-1)
#must divide by duration (in seconds).And to be in kHz we divide by 1000.
corrObj.kcount = np.average(np.array(cdata)/unit)/1000
corrObj.ch_type = channel;
corrObj.param = copy.deepcopy(fit_obj.def_param)
corrObj.calculate_suitability()
#if fit_obj.diffModEqSel.currentIndex()+1 == 3:
# GS.calc_param_fcs(fit_obj,corrObj)
#else:
# SE.calc_param_fcs(fit_obj,corrObj)
if fit_obj.def_options['Diff_eq'] == 4:
VD.calc_param_fcs(fit_obj,corrObj)
elif fit_obj.def_options['Diff_eq'] == 3:
GS.calc_param_fcs(fit_obj,corrObj)
else:
SE.calc_param_fcs(fit_obj,corrObj)
def sin_import_method(fit_obj,file_path): tscale = []; tdata = []; tdata2 = []; tdata3 = []; tdata4 = []; int_tscale =[]; int_tdata = []; int_tdata2 = []; proceed = False for line in csv.reader(open(file_path, 'rb'),delimiter='\t'): if proceed =='correlated': if line ==[]: proceed =False; else: tscale.append(float(line[0])) tdata.append(float(line[1])) if line.__len__()>2: tdata2.append(float(line[2])) if line.__len__()>3: tdata3.append(float(line[3])) if line.__len__()>4: tdata4.append(float(line[4])) if proceed =='intensity': if line ==[]: proceed=False; elif line.__len__()> 1: int_tscale.append(float(line[0])) int_tdata.append(float(line[1])) if line.__len__()>2: int_tdata2.append(float(line[2])) if (str(line) == "[\'[CorrelationFunction]\']"): proceed = 'correlated'; elif (str(line) == "[\'[IntensityHistory]\']"): proceed = 'intensity'; corrObj1 = corrObject(file_path,fit_obj) corrObj1.siblings = None corrObj1.autoNorm= np.array(tdata).astype(np.float64).reshape(-1) corrObj1.autotime= np.array(tscale).astype(np.float64).reshape(-1)*1000 corrObj1.name = corrObj1.name+'-CH0' corrObj1.parent_name = '.sin files' corrObj1.parent_uqid = '0' corrObj1.ch_type = 0; #Average counts per bin. Apparently they are normalised to time. unit = int_tscale[-1]/(int_tscale.__len__()-1) #And to be in kHz we divide by 1000. corrObj1.kcount = np.average(np.array(int_tdata))/1000 corrObj1.param = copy.deepcopy(fit_obj.def_param) corrObj1.calculate_suitability() if fit_obj.def_options['Diff_eq'] == 4: VD.calc_param_fcs(fit_obj,corrObj1) elif fit_obj.def_options['Diff_eq'] == 3: GS.calc_param_fcs(fit_obj,corrObj1) else: SE.calc_param_fcs(fit_obj,corrObj1) fit_obj.objIdArr.append(corrObj1) #Basic if tdata2 !=[]: corrObj2 = corrObject(file_path,fit_obj) corrObj2.autoNorm= np.array(tdata2).astype(np.float64).reshape(-1) corrObj2.autotime= np.array(tscale).astype(np.float64).reshape(-1)*1000 corrObj2.name = corrObj2.name+'-CH1' corrObj2.parent_name = '.sin files' corrObj2.parent_uqid = '0' corrObj2.ch_type = 1; #And to be in kHz we divide by 1000. corrObj2.kcount = np.average(np.array(int_tdata2))/1000 corrObj2.param = copy.deepcopy(fit_obj.def_param) corrObj1.siblings = [corrObj2] corrObj2.siblings = [corrObj1] corrObj2.calculate_suitability() if fit_obj.def_options['Diff_eq'] == 4: VD.calc_param_fcs(fit_obj,corrObj2) elif fit_obj.def_options['Diff_eq'] == 3: GS.calc_param_fcs(fit_obj,corrObj2) else: SE.calc_param_fcs(fit_obj,corrObj2) fit_obj.objIdArr.append(corrObj2) if tdata3 !=[]: corrObj3 = corrObject(file_path,fit_obj) corrObj3.autoNorm= np.array(tdata3).astype(np.float64).reshape(-1) corrObj3.autotime= np.array(tscale).astype(np.float64).reshape(-1)*1000 corrObj3.ch_type = 2; corrObj3.name = corrObj3.name+'-CH10' corrObj3.parent_name = '.sin files' corrObj3.parent_uqid = '0' #And to be in kHz we divide by 1000. #corrObj3.kcount = np.average(np.array(int_tdata3)/unit)/1000 corrObj3.param = copy.deepcopy(fit_obj.def_param) corrObj1.siblings = [corrObj2,corrObj3] corrObj2.siblings = [corrObj1,corrObj3] corrObj3.siblings = [corrObj1,corrObj2] corrObj3.calculate_suitability() if fit_obj.def_options['Diff_eq'] == 4: VD.calc_param_fcs(fit_obj,corrObj3) elif fit_obj.def_options['Diff_eq'] == 3: GS.calc_param_fcs(fit_obj,corrObj3) else: SE.calc_param_fcs(fit_obj,corrObj3) fit_obj.objIdArr.append(corrObj3) if tdata4 !=[]: corrObj4 = corrObject(file_path,fit_obj) corrObj4.autoNorm= np.array(tdata4).astype(np.float64).reshape(-1) corrObj4.autotime= np.array(tscale).astype(np.float64).reshape(-1)*1000 corrObj4.ch_type = 3; corrObj4.name = corrObj4.name+'-CH01' corrObj4.parent_name = '.sin files' corrObj4.parent_uqid = '0' #And to be in kHz we divide by 1000. #corrObj4.kcount = np.average(np.array(int_tdata4)/unit)/1000 corrObj4.param = copy.deepcopy(fit_obj.def_param) corrObj1.siblings = [corrObj2,corrObj3,corrObj4] corrObj2.siblings = [corrObj1,corrObj3,corrObj4] corrObj3.siblings = [corrObj1,corrObj2,corrObj4] corrObj4.siblings = [corrObj1,corrObj2,corrObj3] corrObj4.calculate_suitability() if fit_obj.def_options['Diff_eq'] == 4: VD.calc_param_fcs(fit_obj,corrObj4) elif fit_obj.def_options['Diff_eq'] == 3: GS.calc_param_fcs(fit_obj,corrObj4) else: SE.calc_param_fcs(fit_obj,corrObj4) fit_obj.objIdArr.append(corrObj4)
def csv_import_method(fit_obj,file_path):
r_obj = csv.reader(open(file_path, 'rb'))
line_one = r_obj.next()
if line_one.__len__()>1:
if float(line_one[1]) == 2:
version = 2
else:
print('version not known:',line_one[1])
else:
version = 1
corrObj1 = corrObject(file_path,fit_obj);
if version == 1:
fit_obj.objIdArr.append(corrObj1)
c = 0
for line in csv.reader(open(file_path, 'rb')):
if (c >0):
tscale.append(line[0])
tdata.append(line[1])
c +=1;
corrObj1.autoNorm= np.array(tdata).astype(np.float64).reshape(-1)
corrObj1.autotime= np.array(tscale).astype(np.float64).reshape(-1)
corrObj1.calculate_suitability()
corrObj1.name = corrObj1.name+'-CH'+str(0)
corrObj1.ch_type = 0
corrObj1.datalen= len(tdata)
corrObj1.param = copy.deepcopy(fit_obj.def_param)
fit_obj.fill_series_list()
if version >= 2:
numOfCH = float(r_obj.next()[1])
if numOfCH == 1:
fit_obj.objIdArr.append(corrObj1)
corrObj1.type =str(r_obj.next()[1])
corrObj1.ch_type = int(r_obj.next()[1])
corrObj1.name = corrObj1.name+'-CH'+str(corrObj1.ch_type)
corrObj1.parent_name = 'no name'
corrObj1.parent_uqid = '0'
line = r_obj.next()
while line[0] != 'Time (ns)':
if line[0] == 'kcount':
corrObj1.kcount = float(line[1])
if line[0] == 'numberNandB':
corrObj1.numberNandB = float(line[1])
if line[0] == 'brightnessNandB':
corrObj1.brightnessNandB = float(line[1])
if line[0] == 'CV':
corrObj1.CV = float(line[1])
if line[0] == 'carpet pos':
carpet = int(line[1])
if line[0] == 'pc':
pc_text = int(line[1])
if line[0] == 'pbc_f0':
corrObj1.pbc_f0 = float(line[1])
if line[0] == 'pbc_tb':
corrObj1.pbc_tb = float(line[1])
if line[0] == 'parent_name':
corrObj1.parent_name = str(line[1])
if line[0] == 'parent_uqid':
corrObj1.parent_uqid = str(line[1])
line = r_obj.next()
if pc_text != False:
corrObj1.name = corrObj1.name +'_pc_m'+str(pc_text)
tscale = []
tdata = []
#null = r_obj.next()
line = r_obj.next()
while line[0] != 'end':
tscale.append(line[0])
tdata.append(line[1])
line = r_obj.next()
corrObj1.autoNorm= np.array(tdata).astype(np.float64).reshape(-1)
corrObj1.autotime= np.array(tscale).astype(np.float64).reshape(-1)
corrObj1.siblings = None
corrObj1.param = copy.deepcopy(fit_obj.def_param)
if numOfCH == 2:
corrObj2 = corrObject(file_path,fit_obj);
corrObj3 = corrObject(file_path,fit_obj);
fit_obj.objIdArr.append(corrObj1)
fit_obj.objIdArr.append(corrObj2)
fit_obj.objIdArr.append(corrObj3)
corrObj1.parent_name = 'no name'
corrObj1.parent_uqid = '0'
corrObj2.parent_name = 'no name'
corrObj2.parent_uqid = '0'
corrObj3.parent_name = 'no name'
corrObj3.parent_uqid = '0'
line_type = r_obj.next()
corrObj1.type = str(line_type[1])
corrObj2.type = str(line_type[1])
corrObj3.type = str(line_type[1])
line_ch = r_obj.next()
corrObj1.ch_type = int(line_ch[1])
corrObj2.ch_type = int(line_ch[2])
corrObj3.ch_type = int(line_ch[3])
corrObj1.name = corrObj1.name+'-CH'+str(corrObj1.ch_type)
corrObj2.name = corrObj2.name+'-CH'+str(corrObj2.ch_type)
corrObj3.name = corrObj3.name+'-CH'+str(corrObj3.ch_type)
line = r_obj.next()
while line[0] != 'Time (ns)':
if line[0] == 'kcount':
corrObj1.kcount = float(line[1])
corrObj2.kcount = float(line[2])
if line[0] == 'numberNandB':
corrObj1.numberNandB = float(line[1])
corrObj2.numberNandB = float(line[2])
if line[0] == 'brightnessNandB':
corrObj1.brightnessNandB = float(line[1])
corrObj2.brightnessNandB = float(line[2])
if line[0] == 'CV':
corrObj1.CV = float(line[1])
corrObj2.CV = float(line[2])
corrObj3.CV = float(line[3])
if line[0] == 'carpet pos':
corrObj1.carpet_position = int(line[1])
if line[0] == 'pc':
pc_text = int(line[1])
if line[0] == 'pbc_f0':
corrObj1.pbc_f0 = float(line[1])
corrObj2.pbc_f0 = float(line[2])
if line[0] == 'pbc_tb':
corrObj1.pbc_tb = float(line[1])
corrObj2.pbc_tb = float(line[2])
if line[0] == 'parent_name':
corrObj1.parent_name = str(line[1])
corrObj2.parent_name = str(line[1])
corrObj3.parent_name = str(line[1])
if line[0] == 'parent_uqid':
corrObj1.parent_uqid = str(line[1])
corrObj2.parent_uqid = str(line[1])
corrObj3.parent_uqid = str(line[1])
print(line)
line = r_obj.next()
if pc_text != False:
corrObj1.name = corrObj1.name +'_pc_m'+str(pc_text)
corrObj2.name = corrObj2.name +'_pc_m'+str(pc_text)
corrObj3.name = corrObj3.name +'_pc_m'+str(pc_text)
#null = r_obj.next()
line = r_obj.next()
tscale = []
tdata0 = []
tdata1 = []
tdata2 = []
while line[0] != 'end':
tscale.append(line[0])
tdata0.append(line[1])
tdata1.append(line[2])
tdata2.append(line[3])
line = r_obj.next()
corrObj1.autotime= np.array(tscale).astype(np.float64).reshape(-1)
corrObj2.autotime= np.array(tscale).astype(np.float64).reshape(-1)
corrObj3.autotime= np.array(tscale).astype(np.float64).reshape(-1)
corrObj1.autoNorm= np.array(tdata0).astype(np.float64).reshape(-1)
corrObj2.autoNorm= np.array(tdata1).astype(np.float64).reshape(-1)
corrObj3.autoNorm= np.array(tdata2).astype(np.float64).reshape(-1)
corrObj1.siblings = [corrObj2,corrObj3]
corrObj2.siblings = [corrObj1,corrObj3]
corrObj3.siblings = [corrObj1,corrObj2]
corrObj1.param = copy.deepcopy(fit_obj.def_param)
corrObj2.param = copy.deepcopy(fit_obj.def_param)
corrObj3.param = copy.deepcopy(fit_obj.def_param)
corrObj1.calculate_suitability()
corrObj2.calculate_suitability()
corrObj3.calculate_suitability()
SE.calc_param_fcs(fit_obj,corrObj1)
def fitToParameters(self):
#Populate param for lmfit.
param = Parameters()
#self.def_param.add('A1', value=1.0, min=0,max=1.0, vary=False)
for art in self.param:
if self.param[art]['to_show'] == True and self.param[art]['calc'] == False:
param.add(art, value=float(self.param[art]['value']), min=float(self.param[art]['minv']), max=float(self.param[art]['maxv']), vary=self.param[art]['vary']);
#Find the index of the nearest point in the scale.
data = np.array(self.autoNorm).astype(np.float64).reshape(-1)
scale = np.array(self.autotime).astype(np.float64).reshape(-1)
self.indx_L = int(np.argmin(np.abs(scale - self.parentFn.dr.xpos)))
self.indx_R = int(np.argmin(np.abs(scale - self.parentFn.dr1.xpos)))
if self.parentFn.bootstrap_enable_toggle == True:
num_of_straps = self.parentFn.bootstrap_samples.value()
aver_data = {}
lim_scale = scale[self.indx_L:self.indx_R+1].astype(np.float64)
lim_data = data[self.indx_L:self.indx_R+1].astype(np.float64)
#Populate a container which will store our output variables from the bootstrap.
for art in self.param:
if self.param[art]['to_show'] == True and self.param[art]['calc'] == False:
aver_data[art] = []
for i in range(0,num_of_straps):
#Bootstrap our sample, but remove duplicates.
boot_ind = np.random.choice(np.arange(0,lim_data.shape[0]),size=lim_data.shape[0],replace=True);
#boot_ind = np.arange(0,lim_data.shape[0])#np.sort(boot_ind)
boot_scale = lim_scale[boot_ind]
boot_data = lim_data[boot_ind]
res = minimize(self.residual, param, args=(boot_scale,boot_data, self.parentFn.def_options))
for art in self.param:
if self.param[art]['to_show'] == True and self.param[art]['calc'] == False:
aver_data[art].append(res.params[art].value)
for art in self.param:
if self.param[art]['to_show'] == True and self.param[art]['calc'] == False:
self.param[art]['value'] = np.average(aver_data[art])
self.param[art]['stderr'] = np.std(aver_data[art])
#Run the fitting.
if self.parentFn.bootstrap_enable_toggle == False:
res = minimize(self.residual, param, args=(scale[self.indx_L:self.indx_R+1],data[self.indx_L:self.indx_R+1], self.parentFn.def_options))
#Repopulate the parameter object.
for art in self.param:
if self.param[art]['to_show'] == True and self.param[art]['calc'] == False:
self.param[art]['value'] = res.params[art].value
self.param[art]['stderr'] = float(res.params[art].stderr)
#Extra parameters, which are not fit or inherited.
#self.param['N_FCS']['value'] = np.round(1/self.param['GN0']['value'],4)
#Populate param for the plotting. Would like to use same method as
plot_param = Parameters()
#self.def_param.add('A1', value=1.0, min=0,max=1.0, vary=False)
for art in self.param:
if self.param[art]['to_show'] == True and self.param[art]['calc'] == False:
plot_param.add(art, value=float(self.param[art]['value']), min=float(self.param[art]['minv']), max=float(self.param[art]['maxv']), vary=self.param[art]['vary']);
#Display fitted equation.
if self.parentFn.def_options['Diff_eq'] == 5:
self.model_autoNorm = PB.equation_(plot_param, scale[self.indx_L:self.indx_R+1],self.parentFn.def_options)
elif self.parentFn.def_options['Diff_eq'] == 4:
self.model_autoNorm = VD.equation_(plot_param, scale[self.indx_L:self.indx_R+1],self.parentFn.def_options)
elif self.parentFn.def_options['Diff_eq'] == 3:
self.model_autoNorm = GS.equation_(plot_param, scale[self.indx_L:self.indx_R+1],self.parentFn.def_options)
else:
self.model_autoNorm = SE.equation_(plot_param, scale[self.indx_L:self.indx_R+1],self.parentFn.def_options)
self.model_autotime = scale[self.indx_L:self.indx_R+1]
if self.parentFn.bootstrap_enable_toggle == False:
self.residualVar = res.residual
else:
self.residualVar = self.model_autoNorm - data[self.indx_L:self.indx_R+1].astype(np.float64)
output = fit_report(res.params)
print('residual',res.chisqr)
if(res.chisqr>self.parentFn.chisqr):
print('CAUTION DATA DID NOT FIT WELL CHI^2 >',self.parentFn.chisqr,' ',res.chisqr)
self.goodFit = False
else:
self.goodFit = True
self.fitted = True
self.chisqr = res.chisqr
self.localTime = time.asctime( time.localtime(time.time()) )
#Update the displayed options.
if self.parentFn.def_options['Diff_eq'] == 5:
PB.calc_param_fcs(self.parentFn,self)
elif self.parentFn.def_options['Diff_eq'] == 4:
VD.calc_param_fcs(self.parentFn,self)
elif self.parentFn.def_options['Diff_eq'] == 3:
GS.calc_param_fcs(self.parentFn,self)
else:
SE.calc_param_fcs(self.parentFn,self)