more cleanup and add funciton for showing abs spectra with vernier sweep
This commit is contained in:
parent
327d2825c6
commit
0df7c7b2f1
|
@ -0,0 +1 @@
|
|||
dispersiveXanes_alignment.py
|
|
@ -68,77 +68,3 @@ def medianRatio(p1,p2,threshold=0.03):
|
|||
return np.ma.average(ratio,axis=0,weights=p1)
|
||||
|
||||
|
||||
# /--------------------\
|
||||
# | |
|
||||
# | PLOTS & CO. |
|
||||
# | |
|
||||
# \--------------------/
|
||||
|
||||
def plotShot(im1,im2,transf1=None,transf2=None,fig=None,ax=None,res=None,E=defaultE,save=None):
|
||||
if transf1 is not None: im1 = transf1.transformImage(im1)
|
||||
if transf2 is not None: im2 = transf2.transformImage(im2)
|
||||
if fig is None and ax is None:
|
||||
fig = plt.subplots(2,3,figsize=[7,5],sharex=True)[0]
|
||||
ax = fig.axes
|
||||
elif fig is not None:
|
||||
ax = fig.axes
|
||||
if E is None: E=np.arange(im1.shape[1])
|
||||
n = im1.shape[0]
|
||||
ax[0].imshow(im1,extent=(E[0],E[-1],0,n),**kw_2dplot)
|
||||
ax[1].imshow(im2,extent=(E[0],E[-1],0,n),**kw_2dplot)
|
||||
ax[2].imshow(im1-im2,extent=(E[0],E[-1],0,n),**kw_2dplot)
|
||||
if res is None:
|
||||
p1 = np.nansum(im1,axis=0)
|
||||
p2 = np.nansum(im2,axis=0)
|
||||
pr = p2/p1
|
||||
else:
|
||||
p1 = res.p1; p2 = res.p2; pr = res.ratio
|
||||
ax[3].plot(E,p1,lw=3)
|
||||
ax[4].plot(E,p1,lw=1)
|
||||
ax[4].plot(E,p2,lw=3)
|
||||
idx = (p1>p1.max()/10.)
|
||||
ax[5].plot(E[idx],pr[idx])
|
||||
if res is not None:
|
||||
ax[5].set_title("FOM: %.2f"%res.fom)
|
||||
else:
|
||||
ax[5].set_title("FOM: %.2f"% calcFOM(p1,p2,pr))
|
||||
if (save is not None) and (save is not False): plt.savefig(save,transparent=True,dpi=500)
|
||||
return fig
|
||||
|
||||
def plotRatios(r,shot='random',fig=None,E=defaultE,save=None):
|
||||
if fig is None: fig = plt.subplots(2,1,sharex=True)[0]
|
||||
ax = fig.axes
|
||||
n = r.shape[0]
|
||||
i = ax[0].imshow(r,extent=(E[0],E[-1],0,n),**kw_2dplot)
|
||||
i.set_clim(0,1.2)
|
||||
if shot == 'random' : shot = np.random.random_integers(0,n-1)
|
||||
ax[1].plot(E,r[shot],label="Shot n %d"%shot)
|
||||
ax[1].plot(E,np.nanmedian(r[:10],axis=0),label="median 10 shots")
|
||||
ax[1].plot(E,np.nanmedian(r,axis=0),label="all shots")
|
||||
ax[1].legend()
|
||||
ax[1].set_ylim(0,1.5)
|
||||
ax[1].set_xlabel("Energy")
|
||||
ax[1].set_ylabel("Transmission")
|
||||
ax[0].set_ylabel("Shot num")
|
||||
if (save is not None) and (save is not False): plt.savefig(save,transparent=True,dpi=500)
|
||||
|
||||
def plotSingleShots(r,nShots=10,fig=None,E=defaultE,save=None,ErangeForStd=(7090,7150)):
|
||||
if fig is None: fig = plt.subplots(2,1,sharex=True)[0]
|
||||
ax = fig.axes
|
||||
for i in range(nShots):
|
||||
ax[0].plot(E,r[i]+i)
|
||||
ax[0].set_ylim(0,nShots+0.5)
|
||||
av = (1,3,10,30,100)
|
||||
good = np.nanmedian(r,0)
|
||||
for i,a in enumerate(av):
|
||||
m = np.nanmedian(r[:a],0)
|
||||
idx = (E>ErangeForStd[0]) & (E<ErangeForStd[1])
|
||||
fom = np.nanstd( m[idx]/good[idx] )
|
||||
print("n shots %d, std %.2f"%(a,fom) )
|
||||
ax[1].plot(E,m+i,label="%d shots, std :%.2f"%(a,fom))
|
||||
ax[1].legend()
|
||||
ax[1].set_ylim(0,len(av)+0.5)
|
||||
ax[1].set_xlabel("Energy")
|
||||
ax[1].set_ylabel("Transmission")
|
||||
if (save is not None) and (save is not False): plt.savefig(save,transparent=True,dpi=500)
|
||||
|
||||
|
|
|
@ -4,6 +4,7 @@ import copy
|
|||
import argparse
|
||||
import collections
|
||||
import dispersiveXanes_alignment as alignment
|
||||
import dispersiveXanes_utils as utils
|
||||
import xanes_analyzeRun
|
||||
|
||||
parser = argparse.ArgumentParser(description='Process argv')
|
||||
|
@ -102,16 +103,16 @@ def calcRef(r1,r2,calibs=None,threshold=0.05):
|
|||
out["ratioOfAverage"] = dict()
|
||||
out["medianOfRatios"] = dict()
|
||||
for p1,p2,n in zip(r1,r2,calibs):
|
||||
out["ratioOfAverage"][n] = alignment.ratioOfAverage(p1,p2,threshold=threshold)
|
||||
out["medianOfRatios"][n] = alignment.medianRatio(p1,p2,threshold=threshold)
|
||||
out["ratioOfAverage"][n] = utils.ratioOfAverage(p1,p2,threshold=threshold)
|
||||
out["medianOfRatios"][n] = utils.medianRatio(p1,p2,threshold=threshold)
|
||||
# add curves with all calib together
|
||||
p1 = np.vstack(r1)
|
||||
p2 = np.vstack(r2)
|
||||
n = ",".join(map(str,calibs))
|
||||
ref1 = xanes_analyzeRun.ratioOfAverage(p1,p2,threshold=threshold)
|
||||
ref2 = xanes_analyzeRun.medianRatio(p1,p2,threshold=threshold)
|
||||
out["ratioOfAverage"][n] = alignment.ratioOfAverage(p1,p2,threshold=threshold)
|
||||
out["medianOfRatios"][n] = alignment.medianRatio(p1,p2,threshold=threshold)
|
||||
ref1 = utils.ratioOfAverage(p1,p2,threshold=threshold)
|
||||
ref2 = utils.medianRatio(p1,p2,threshold=threshold)
|
||||
out["ratioOfAverage"][n] = utils.ratioOfAverage(p1,p2,threshold=threshold)
|
||||
out["medianOfRatios"][n] = utils.medianRatio(p1,p2,threshold=threshold)
|
||||
out["ratioOfAverage"]['all'] = out["ratioOfAverage"][n]
|
||||
out["medianOfRatios"]['all'] = out["medianOfRatios"][n]
|
||||
return out
|
||||
|
@ -150,7 +151,7 @@ def calcSampleAbs(run=82,refCalibs=slice(None,None,2),threshold=0.05,refKind="me
|
|||
p1 = np.vstack( sample.p1 )
|
||||
p2 = np.vstack( sample.p2 )
|
||||
print(p1.shape)
|
||||
p1,p2 = alignment.maskLowIntensity(p1,p2,threshold=threshold)
|
||||
p1,p2 = utils.maskLowIntensity(p1,p2,threshold=threshold)
|
||||
ratio = p2/p1
|
||||
ratio = ratio/ref
|
||||
return p1,p2,-np.log10(ratio)
|
||||
|
@ -185,13 +186,26 @@ def showSpectra(run=82,shots=slice(5),calibs=0,averageEachCalib=False,
|
|||
ax[0][0].set_title("Run %d"%run)
|
||||
if not averageEachCalib: ax[0][0].set_ylim(0,shifty*(len(spectra_norm)))
|
||||
|
||||
def showAbs(run=82,shots=slice(5),normalization="auto",shifty=1,xlim=(7080,7180),showAv=True,smoothWidth=0,threshold=0.01):
|
||||
|
||||
color_ss = '#08519c'
|
||||
color_av = '#238b45'
|
||||
color_av_all = '#d95f0e'
|
||||
|
||||
def showAbs(run=82,shots=slice(5),normalization="auto",shifty=1,xlim=(7080,7180),showAv=True,showAvOverAll=True,smoothWidth=0,threshold=0.01,filterShot=0.1):
|
||||
""" normalization: if "auto", the max of the spectra that will be plotted will be used
|
||||
filterShot = means that it filters out the filterShot*100 percentile
|
||||
"""
|
||||
E = alignment.defaultE
|
||||
p1,p2,abs = calcSampleAbs(run=run,threshold=threshold)
|
||||
p1_sum = p1.sum(-1)
|
||||
if filterShot>0:
|
||||
idx = p1_sum>np.percentile(p1_sum,filterShot*100)
|
||||
p1 = p1[idx]
|
||||
p2 = p2[idx]
|
||||
abs = abs[idx]
|
||||
p1_av = np.nanmean(p1,axis=0)
|
||||
p2_av = np.nanmean(p2,axis=0)
|
||||
# somehow nanmedian screws up when array is too big ... so using nanmean
|
||||
abs_av = np.nanmean(abs,axis=0)
|
||||
p1 = p1[shots]; p2=p2[shots]; abs = abs[shots]
|
||||
if smoothWidth > 0: abs = smoothSpectra(E,abs,res=smoothWidth)
|
||||
|
@ -204,17 +218,26 @@ def showAbs(run=82,shots=slice(5),normalization="auto",shifty=1,xlim=(7080,7180)
|
|||
color = gradual_colors[ishot%len(gradual_colors)]
|
||||
ax[0].axhline(ishot*shifty,ls='--',color=color)
|
||||
ax[1].axhline(ishot*shifty,ls='--',color=color)
|
||||
if showAvOverAll:
|
||||
if ishot == 0: ax[0].fill_between(E,ishot*shifty,p1_av/normalization+ishot*shifty,color=color_av_all,alpha=0.6)
|
||||
ax[1].plot(E,abs_av+ishot*shifty,color=color_av_all,lw=2,zorder=20)
|
||||
if showAv:
|
||||
if ishot == 0: ax[0].fill_between(E,ishot*shifty,p1_av/normalization+ishot*shifty,color='#d95f0e',alpha=0.6)
|
||||
ax[1].plot(E,abs_av+ishot*shifty,color='#d95f0e',lw=2,zorder=20)
|
||||
ax[1].plot(E,np.nanmedian(abs,0)+ishot*shifty,color=color_av,lw=2,zorder=10)
|
||||
ax[0].plot(E,s1_norm+ishot*shifty,ls = '-' ,color='0.8',lw=2)
|
||||
ax[0].plot(E,s2_norm+ishot*shifty,ls = '-' ,color='0.3',lw=2)
|
||||
ax[1].plot(E,a+ishot*shifty,color=color,lw=2)
|
||||
ax[1].plot(E,a+ishot*shifty,color=color_ss,lw=2)
|
||||
ax[0].set_xlim(*xlim)
|
||||
ax[0].set_title("Run %d"%run)
|
||||
ax[0].set_title("Run %s"%str(run))
|
||||
ax[1].set_ylabel("Sample Absorption")
|
||||
ax[0].set_ylabel("Normalized Spectra")
|
||||
ax[0].set_ylim(0,shifty*(p1.shape[0]))
|
||||
print("STD of (average over shown shots) - (average over all): %.3f"%np.nanstd(np.nanmedian(abs,0)-abs_av))
|
||||
ax[1].set_title("STD of (average_{shown}) - (average_{all}): %.3f"%np.nanstd(np.nanmedian(abs,0)-abs_av))
|
||||
|
||||
|
||||
def showAbsWithSweep(run=(155,156),first=0,period=150,nSpectra=10,**kwards):
|
||||
shots = slice(first,first+period,int(period/nSpectra))
|
||||
showAbs(run=run,shots=shots,**kwards)
|
||||
|
||||
def smoothSpectra(E,abs_spectra,res=0.5):
|
||||
from scipy import integrate
|
||||
|
|
Loading…
Reference in New Issue