save = kwargs.get('save', False)

fmt = kwargs.get('fmt', '%9.4g')

print "Fetching data..."

for k in gdat.keys():

try:

gdat[k]['data'] = np.loadtxt(k+'.dat')

print "Loaded %s succesfully." % k

except:

data, errmsg, rawdat = qrange.qrange_eval( gdat[k]['dir'], gdat[k]['shots'], datakeys)

if save:

np.savetxt(k+'.dat', data, fmt=fmt)

gdat[k]['data'] = data

print "Done."

K = lambda key: datakeys.index(key)

dat1 = np.array(np.copy(dat))

#return dat1

for c in conds:

Kc = K(c[0])

cval = c[1]

if len(dat1) > 0 :

try:

if len(c) == 2:

dat1 = dat1[ dat1[:,Kc] == cval ]

elif len(c) == 3:

if c[2] == '>':

dat1 = dat1[ dat1[:,Kc] > cval ]

elif c[2] == '<':

dat1 = dat1[ dat1[:,Kc] < cval ]

elif c[2] == '=':

dat1 = dat1[ dat1[:,Kc] == cval ]

elif c[2] == '>=':

dat1 = dat1[ dat1[:,Kc] >= cval ]

elif c[2] == '<=':

dat1 = dat1[ dat1[:,Kc] <= cval ]

elif c[2] == '!=':

dat1 = dat1[ dat1[:,Kc] != cval ]

except:

print "Condition failed:"

print "key index = ", Kc

print "looking for cval = ",cval

print "in ", dat1

raise

# First get the data for each condition

dat1 = data_pick( dat, cond1, K)

dat2 = data_pick( dat, cond2, K)

# Then determine the x values they have in common

x1= np.unique(dat1[:,K(Xkey)])

x2= np.unique(dat2[:,K(Xkey)])

set1 = set( x1.tolist())

set2 = set( x2.tolist())

common = sorted(list( set1 & set2 ))

# Decide which operation will be performed on the two sets

operation = kwargs.get( 'operation', 'ratio')

# Then loop through the common x values to get the

# quantities of interest

num = dat1 ; den = dat2 ;

rval = []; rerr = [] ; nsamples=[]

for c in common:

numi = num[ num[:,K(Xkey)] == c ][ :, K(Ykey) ]

deni = den[ den[:,K(Xkey)] == c ][ :, K(Ykey) ]

ns = max( len(numi), len(deni))

if operation == 'ratio':

val = ufloat(( np.mean(numi), stats.sem(numi) )) / ufloat(( np.mean(deni), stats.sem(deni) ))

elif operation == 'sum':

val = ufloat(( np.mean(numi), stats.sem(numi) )) + ufloat(( np.mean(deni), stats.sem(deni) ))

else:

raise Exception('undefined operation in DataHandling.data_ratio' )

rval.append( val.nominal_value )

rerr.append( val.std_dev() )

# nsamples.append( ns ) # Changed the behaviour of nsamples 140428

nsamples.append( [len(numi), len(deni)] )

rval = np.array( rval )

rerr = np.array( rerr )

xc = np.array(common)

xsets = []

for s in sets:

xsets.append( set( s[0].tolist()) )

# Determine the x values they are common for all sets

common = set.intersection( *xsets )

def inset(x, xset):

if x in xset:

return True

X=[]; Y=[]; YERR=[];

for s in sets:

xs = []; ys = []; yerrs=[];

for i in range( len(s[0] )):

if s[0][i] in common:

xs.append( s[0][i] )

ys.append( s[1][i] )

yerrs.append( s[2][i] )

xs = np.array( xs )

ys = np.array( ys )

yerrs = np.array( yerrs)

index = np.argsort (xs )

X.append( xs[ index ] )

Y.append( ys[ index ] )

YERR.append( yerrs[ index] )

X = np.array(X)

Y = np.array(Y)

YERR = np.array(YERR)

#print "\nX="

#print X

#print "\nY="

#print Y

#print "\nYERR="

#print YERR

# Perform weighted average

weights = 1/ np.power( YERR, 2. )

y_wav = np.sum( Y * weights , axis=0) / np.sum(weights, axis=0 )

y_err_wav = 1./ np.sqrt( np.sum(weights, axis=0) )

#print "\nWeighted average"

#print y_wav

#print "\nError"

#print y_err_wav

try:

defmarker = gdict['marker']

except:

defmarker = 'o'

marker = kwargs.get( 'marker', defmarker)

ms = kwargs.get( 'ms', 5. )

mew = kwargs.get( 'mew', 1.0 )

mec = kwargs.get( 'mec', gdict['ec'])

mfc = kwargs.get( 'mfc', gdict['fc'])

ew = kwargs.get( 'ew', 1.0 )

ecap = kwargs.get( 'ecap', 0. )

save = kwargs.get( 'save', False )

exceptions = kwargs.get( 'exceptions', False)

raw = kwargs.get( 'raw', True )

raw_offset = kwargs.get( 'raw_offset', 0.)

labelstr = kwargs.get( 'labelstr', gdict['label'] )

xkey0 = xkey

ykey0 = ykey

xkey = K(xkey)

ykey = K(ykey)

if kwargs.get('use_stddev', False):

error_index = 2 # standard deviation

else:

error_index = 3 # standard error

discard = kwargs.pop('discard',None)

if discard is not None:

if 'y>' in discard.keys():

dat = dat[ dat[:,ykey] < discard['y>'] ]

if 'y<' in discard.keys():

dat = dat[ dat[:,ykey] > discard['y<'] ]

try:

both_offset = kwargs.get('both_offset', 0.)

yf = kwargs.get( 'yf', None)

if yf is not None:

ydict = kwargs.get('yf_kwargs', {} )

yf_usex = kwargs.get('yf_usex', False)

if yf_usex is True:

ydict['x'] = fx(xc)

if raw:

rawcolor = kwargs.pop('rawcolor', 'gray')

rawalpha = kwargs.pop('rawalpha', 0.5)

rawms = kwargs.pop('rawalpha', 4.5)

xraw = fx(dat[:, xkey] ) + raw_offset + both_offset

yraw = dat[:,ykey]/base

if yf is not None:

yraw = yf( yraw, **ydict )

ax.plot( xraw, yraw , '.',

marker='.', ms=rawms,\

color=rawcolor, alpha=rawalpha)

datS = statdat.statdat( dat, xkey, ykey, **kwargs)

xplot = fx(datS[:,0])

yplot = datS[:,1]/base

yploterr = datS[:,error_index]/base

if yf is not None:

yunc = unumpy.uarray(( yplot, yploterr ))

yunc = yf( yunc, **ydict )

yplot = unumpy.nominal_values( yunc )

yploterr = unumpy.std_devs( yunc )

ax.errorbar( xplot+both_offset, yplot, yerr=yploterr,\

capsize=ecap, elinewidth=ew,\

fmt='.', ecolor=mec, mec=mec,\

mew=mew, marker=marker, mfc=mfc, ms=ms,\

label=labelstr)

guide = kwargs.get( 'guide', False)

if guide:

guide_color = kwargs.get('guide_color', mec)

f = interp1d( xplot, yplot, kind='linear')

xnew = np.linspace( xplot.min(), xplot.max(), 120)

ax.plot( xnew, f(xnew), color=guide_color, zorder = 1.1 )

if save:

fname = xkey0.replace(':','_') + '_' + labelstr[-3:] + '.rawdat'

X = np.transpose(np.vstack( ( xplot, yplot, yploterr )))

np.savetxt( fname, X, fmt='%10.2f', delimiter='\t', newline='\n')

if kwargs.get('return_raw', False):

return xraw, yraw

return xplot, unumpy.uarray(( yplot, yploterr ))

except:

print "Exception occured in plotkey %s vs %s"% (ykey0,xkey0)

if exceptions:

try:

defmarker = gdict['marker']

except:

defmarker = 'o'

marker = kwargs.get( 'marker', defmarker)

ms = kwargs.get( 'ms', 5. )

mew = kwargs.get( 'mew', 1.0 )

mec = kwargs.get( 'mec', gdict['ec'])

mfc = kwargs.get( 'mfc', gdict['fc'])

save = kwargs.get( 'save', False )

exceptions = kwargs.get( 'exceptions', False)

raw = kwargs.get( 'raw', True )

raw_offset = kwargs.get( 'raw_offset', 0.)

labelstr = kwargs.get( 'labelstr', gdict['label'] )

xkey0 = xkey

ykey0 = ykey

xkey = K(xkey)

ykey = K(ykey)

try:

datS = statdat.statdat( dat, xkey, ykey )

Ystderr = datS[:,3]/base

Ystddev = datS[:,2]/base

X = datS[:,0]

Y = datS[:,1]/base

def meanY( x ):

try:

index = np.where( X == x )[0][0]

assert isinstance( index, int )

return Y[index]

except:

print "Error finding mean value of Y at X = ", x

raise

if raw:

datY = dat[:,ykey]/base

datX = dat[:,xkey]

datYnormed = np.array([ datY[i] / meanY( datX[i] ) \

for i in range(len(datX)) ] )

rawcolor = kwargs.pop('rawcolor', 'gray')

ax.plot( fx(datX)+raw_offset, datYnormed, '.',\

marker='.', ms=4.5,\

color=rawcolor, alpha=0.5)

ax.plot( fx(X), 100.*2.*Ystderr/Y, '.',\

mec=mec, mew=mew, marker=marker, mfc=mfc, ms=ms,\

label=labelstr)

#ax.plot( fx(X), 100.*Ystddev/Y, '.',\

# mec=mec, mew=mew, marker=marker, mfc=mfc, ms=ms,\

# label=labelstr)

if save:

fname = xkey0.replace(':','_') + '_' + labelstr[-3:] + 'RELERR.rawdat'

X = np.transpose(np.vstack( ( fx(datS[:,0]), datS[:,1]/base, datS[:,3]/base )))

np.savetxt( fname, X, fmt='%10.2f', delimiter='\t', newline='\n')

return fx(datS[:,0]), unumpy.uarray(( datS[:,1]/base, datS[:,3]/base ))

except:

print "Exception occured in plotkey %s vs %s"% (ykey0,xkey0)

if exceptions:

raise

try:

defmarker = gdict['marker']

except:

defmarker = 'o'

marker = kwargs.get( 'marker', defmarker)

ms = kwargs.get( 'ms', 5. )

mec = kwargs.get( 'mec', gdict['ec'])

mfc = kwargs.get( 'mfc', gdict['fc'])

mew = kwargs.get( 'mew', 1.0 )

ew = kwargs.get( 'ew', 1.0 )

ecap = kwargs.get( 'ecap', 0. )

save = kwargs.get( 'save', False )

labelstr = kwargs.get( 'labelstr', gdict['label'] )

xoffset = kwargs.get( 'xoffset', 0.)

exceptions = kwargs.get( 'exceptions', False)

operation = kwargs.get( 'operation', 'ratio')

try:

xc, rval, rerr, nsamples = data_ratio( dat, cond1, cond2, K, xkey, ykey, \

operation=operation )

rval = rval / base

rerr = rerr / base

#print '\t%s: x points in common = '%ykey,len(xc)

yf = kwargs.get( 'yf', None)

if yf is not None:

ydict = kwargs.get('yf_kwargs', {} )

yf_usex = kwargs.get('yf_usex', False)

if yf_usex is True:

ydict['x'] = fx(xc)

runc = unumpy.uarray(( rval, rerr ))

runc = yf( runc, **ydict )

rval = unumpy.nominal_values( runc )

rerr = unumpy.std_devs( runc )

if len(xc) > 0:

## Debugging

#print "plotkey_ratio debug"

#print "len( x ) = ", len(fx(xc))

#print "len( y ) = ", len(rval/base)

#print "len( yerr ) = ", len(rerr/base)

ax.errorbar( fx(xc) + xoffset, rval/base, yerr=rerr/base,\

capsize=ecap, elinewidth=ew,\

fmt='.', ecolor=mec, mec=mec,\

mew=mew, marker=marker, mfc=mfc, ms=ms,\

label=labelstr)

ax_relerr = kwargs.get('ax_relerr', None)

if ax_relerr is not None:

mec_relerr = kwargs.get('mec_relerr', mec )

mew_relerr = kwargs.get('mec_relerr', mew )

marker_relerr = kwargs.get('mec_relerr', marker )

mfc_relerr = kwargs.get('mfc_relerr', mfc )

ms_relerr = kwargs.get('ms_relerr', ms )

ax_relerr.plot( fx(xc), 100.*2.*rerr/rval, '.',\

mec=mec_relerr, mew=mew_relerr, \

marker=marker_relerr, mfc=mfc_relerr, \

ms=ms_relerr,\

label=labelstr)

if save:

fname = xkey.replace(':','_') + '_' + labelstr[-3:] + '.rawdatNOTCorrected'

X = np.transpose(np.vstack( ( fx(xc), rval/base, rerr/base )))

np.savetxt( fname, X, fmt='%10.2f', delimiter='\t', newline='\n')

if kwargs.get('return_unumpy', False):

return {'x':fx(xc), 'y':unumpy.uarray(( rval/base, rerr/base)),\

'nsamples':nsamples }

else:

return [fx(xc), rval/base, rerr/base, nsamples]

except:

print "Exception occured in plotkey_ratio %s vs %s"%(ykey,xkey)

if exceptions: