def peakFitBBA(filepath, config):
'''
Wrapper for Bayesian Block Analysis of 1D Plots.
Takes file path
Assumes 1D files live in (filepath.dirname)/Processed/
'''
print('\n')
print('******************************************** Begin peak fitting...')
##############################################################
############ Parse filepath input ############################
processedPath = os.path.join(os.path.dirname(filepath), 'Processed/')
folder_path = os.path.dirname(filepath)
filename = os.path.basename(filepath)
fileRoot, ext = os.path.splitext(filename)
savePath = processedPath + 'peak_details/'
csvFilepath = os.path.join(processedPath, fileRoot + '_1D.csv')
# Generate Master CSV path
### name w/o ind
base_filename = re.match('(.*?)[0-9]+.[a-zA-Z]+$', filename).group(1)
### index
index = re.match('.*?([0-9]+).[a-zA-Z]+$', filename).group(1)
masterPath = os.path.join(folder_path, base_filename + 'master.csv')
attDict = {'scanNo': index}
if not config:
peakShape = 'Voigt'
numCurves = 2
fit_order = 2
else:
peakShape = config['peakShape']
numCurves = config['peakNo']
fit_order = config['fit_order']
useBkgdImg = config['bkgdImg']
print('config read')
##############End Input#######################################
##############################################################
if not os.path.exists(savePath):
os.makedirs(savePath)
peakCnt = 0
# File data into array
print csvFilepath
data = np.genfromtxt(csvFilepath, delimiter=',')
Qlist = data[:, 0]
IntAve = data[:, 1]
dataArray = np.array([Qlist, IntAve])
##############################################################
#### Data Structure object instantiation (data, fit_order, ncp_prior)
dataIn = BlockData(dataArray, fit_order, 0.5, peakShape)
#### has various functions
##############################################################
dataIn.trimData(trimLen=50)
if useBkgdImg: # if a background image has been supplied
print(config['bkgdPath'])
bkgdData = np.genfromtxt(config['bkgdPath'], delimiter=',')
bkgdX = bkgdData[:, 0]
bkgdY = bkgdData[:, 1]
dataIn.bkgdSubImg(np.array([bkgdX, bkgdY]))
# background subtracted with polynomial of order = fit_order, trims ends
elif type(fit_order) is str:
dataIn.bkgdSub() # Trim using chebyshev
else:
dataIn.bkgdSubPoly(fit_order=fit_order)
# Plot bkgdSub Data
plt.figure(figsize=(8, 8))
plt.plot(Qlist, IntAve, label='Raw data', marker='s', color='k')
plt.plot(dataIn.subData[0],
dataIn.bkgd,
'--',
label='Background',
color='g')
plt.plot(dataIn.subData[0],
dataIn.subData[1],
label='Background subtracted',
color='r')
try:
plt.plot(dataIn.downData[0, :],
dataIn.downData[1, :],
label='Downsampled',
color='b',
marker='o',
linestyle='None')
except Exception as e:
print(e)
plt.legend()
plt.savefig(savePath + basename(csvFilepath)[:-7] + '_plot.png')
plt.close()
save_1Dcsv(dataIn.subData[0], dataIn.subData[1], fileRoot + '_bkgdSub',
savePath)
# Guess at noise level
hld = dataIn.subData
sigmaGuess = np.std(hld[1][hld[1] <= np.median(hld[1])])
dataIn.cellData = sigmaGuess * np.ones(len(dataIn.subData[0]))
# incorporate block information into data struct
dataIn.blockFinder()
# Get optimized parameters from fitting each block and plot
paramDict, litFWHM = bumpFindFit(dataIn, peakShape, numCurves, config,
savePath,
basename(csvFilepath)[:-7])
# Print information to terminal, print data to csv
print('---------Fitting Finished')
print('Fit ({0}) curve(s) per peak'.format(paramDict['numCurves']))
print('Using ({0}) peak shape'.format(paramDict['peakShape']))
# Generate residual plot using stored optParams
pctErr = dataIn.genResidPlot(savePath, csvFilepath)
genOptParamCSV(savePath, csvFilepath, paramDict)
genPeakReportCSV(savePath, filepath, litFWHM, pctErr)
###############################################################################
# Add features to master metadata
# hard pull items for now
attDict['scanNo'] = int(index)
(attDict['FSDP_loc'], attDict['FSDP_FWHM'], attDict['FSDP_Intens'],
attDict['FSDP_yMax']) = findFitFSDP(paramDict, litFWHM, config)
(attDict['maxPeak_loc'], attDict['maxPeak_FWHM'],
attDict['maxPeak_Intens']) = findFitMaxPeak(paramDict, litFWHM, config)
addFeatsToMaster(attDict, masterPath)
# incorporate block information into data struct
dataIn.blockFinder()
# Get optimized parameters from fitting each block and plot
paramDict, litFWHM = bumpFindFit(dataIn, peakShape, numCurves, savePath,
basename(file)[:-7])
# Print information to terminal, print data to csv
print('---------Fitting Finished')
print('Fit ({0}) curve(s) per peak'.format(paramDict['numCurves']))
print('Using ({0}) peak shape'.format(paramDict['peakShape']))
# Generate residual plot using stored optParams
pctErr = dataIn.genResidPlot(savePath, file)
genOptParamCSV(savePath, file, paramDict)
genPeakReportCSV(savePath, file, litFWHM, pctErr)
end = time.time()
loopTime += [(end - start)]
break
# Evaluate performance
avgTime = np.mean(loopTime)
maxTime = np.max(loopTime)
print('============================================================')
print('============================================================')
print('Files finished processing')
print('-----Average {:.4f}s / file, max {:.4f}s / file'.format(