def ex_01(fns=[
'powder_collection_25JUL12/25JUL12_0002Data1Phi-90.txt',
'powder_collection_25JUL12/25JUL12_0002Data2Phi0.txt',
'powder_collection_25JUL12/25JUL12_0002Data3Phi45.txt',
'powder_collection_25JUL12/25JUL12_0002Data4Phi135.txt'
]):
from MP.lib.mpl_lib import wide_fig as wf
from MP.lib.mpl_lib import fancy_legend
from MP.lib.axes_label import __deco__ as deco
fig = wf(nh=1, nw=1, uw=4, left=0.2)
ax = fig.axes[0]
ehkl_dat = []
for i in xrange(len(fns)):
sign_sin2psi, ehkl, dspacing, psi = read(fns[i], iopt=0, isort=True)
ax.plot(sign_sin2psi,
ehkl,
'-x',
label=fns[i].split('Phi')[1].split('.')[0])
ehkl_dat.append(ehkl)
print 'mean:', np.array(ehkl_dat).mean()
print 'std:', np.array(ehkl_dat).std()
deco(iopt=0, ft=15, ax=ax, ipsi_opt=1)
fancy_legend(ax=ax, size=10, ncol=2)
def ex_01(fns=[
'powder_collection_25JUL12/25JUL12_0002Data1Phi-90.txt',
'powder_collection_25JUL12/25JUL12_0002Data2Phi0.txt',
'powder_collection_25JUL12/25JUL12_0002Data3Phi45.txt',
'powder_collection_25JUL12/25JUL12_0002Data4Phi135.txt']):
from MP.lib.mpl_lib import wide_fig as wf
from MP.lib.mpl_lib import fancy_legend
from MP.lib.axes_label import __deco__ as deco
fig = wf(nh=1,nw=1,uw=4,left=0.2); ax=fig.axes[0]
ehkl_dat = []
for i in range(len(fns)):
sign_sin2psi,ehkl,dspacing,psi = read(fns[i],iopt=0,isort=True)
ax.plot(sign_sin2psi,ehkl,'-x',label=fns[i].split('Phi')[1].split('.')[0])
ehkl_dat.append(ehkl)
print 'mean:', np.array(ehkl_dat).mean()
print 'std:', np.array(ehkl_dat).std()
deco(iopt=0,ft=15,ax=ax,ipsi_opt=1)
fancy_legend(ax=ax,size=10,ncol=2)
def ex_igb(fn='igstrain_fbulk_ph1.out',ifig=2,iphi=0,isf=0,
mxnst=None,flow=None):
"""
Read ig strain file and analyze... (igstrain_fbulk_ph1.out)
igstrain_fbulk_ph1.out contains diffraction for each and every
reloading from 'unloaded' polycrystal.
Arguments
=========
fn = 'igstrain_fbulk_ph1.out'
ifig = 2
iphi=0
isf=0
mxnst=None
"""
import matplotlib.pyplot as plt
from sff_converter import condition
from MP.ssort import sh as sort
from MP.mat import voigt
from MP.lib.mpl_lib import wide_fig as wf
difl, nphi, phis, nbeta, dum1, dum2 = condition(fn=None)
if flow==None: raise IOError, 'Flow stress is missing'
eps = flow.epsilon_vm[::]
neps = flow.nstp
nw = 4
remainder = np.mod(neps,nw)
if remainder==0:
nh = neps/nw
elif remainder>0:
nh = neps/nw + 1
fig =wf(nw=2, nh=2,ifig=ifig,iarange=True); axes =fig.axes
ax01, ax02, ax03, ax04 = axes
fig01=wf(nw=nw,nh=nh,ifig=ifig+1,iarange=True,nfig=neps);axes01=fig01.axes
fig02=wf(nw=nw,nh=nh,ifig=ifig+2,iarange=True,nfig=neps);axes02=fig02.axes
fig03=wf(nw=nw,nh=nh,ifig=ifig+3,iarange=True,nfig=neps);axes03=fig03.axes
fig04=wf(nw=nw,nh=nh,ifig=ifig+4,iarange=True,nfig=neps);axes04=fig04.axes
tdat, usf = reader2(fn,iopt=1)
print tdat.shape
nsf = len(usf); nst = len(tdat[0]); npsi = len(tdat[0,0,0,0,:])
if nst!=neps: raise IOError, 'Inconsistency between neps and nst'
markers = ['o','x','+','^','d','*','o','x','+','^','d','*''o','x','+','^','d','*']
colors = ['r','g','b','k','m','y','r','g','b','k','m','y','r','g','b','k','m','y']
from string import ascii_lowercase as alphabet
axe_lab=[]
for ab in alphabet: axe_lab.append('(%s)'%ab)
yl = 0; yh = 0
avg = np.zeros((2,nst,npsi)) #
sft = np.zeros((6,nst,npsi)) # f^{hkl}
rsqt = np.zeros((6,nst,npsi)) # f^{hkl}
if mxnst!=None: nst=mxnst
for i in range(nst):
ehkl = tdat[0,i,isf,iphi,:] # e(hkl,phi,psi)
e = tdat[1,i,isf,iphi,:] # macro
ehkle = tdat[2,i,isf,iphi,:] # e - macro
fhkl = tdat[3,i,isf,iphi,:] # Fhkl
fbulk = tdat[4,i,isf,iphi,:] # Fbulk
ige = tdat[5,i,isf,iphi,:] # e - F_ij *Sij
sij = tdat[6,i,isf,iphi,:] # Sij
psi = tdat[8,i,isf,iphi,:]
rsq = tdat[9,i,isf,iphi,:]
sin2psi = np.sin(psi*np.pi/180.)**2
axes01[i].set_title(r'%s $\bar{E}^{\mathrm{eff}} = %4.2f$'%(
axe_lab[i],eps[i]))
axes02[i].set_title(r'%s $\bar{E}^{\mathrm{eff}} = %4.2f$'%(
axe_lab[i],eps[i]))
axes03[i].set_title(r'%s $\bar{E}^{\mathrm{eff}} = %4.2f$'%(
axe_lab[i],eps[i]))
axes04[i].set_title(r'%s $\bar{E}^{\mathrm{eff}} = %4.2f$'%(
axe_lab[i],eps[i]))
igys = []; igos = []; igts = []
for j in range(nsf): ## Number of elastic loading for stress factor calc.
ehkl_ = tdat[0,i,j,iphi,:] # e(hkl,phi,psi)
e_ = tdat[1,i,j,iphi,:] # macro
fhkl_ = tdat[3,i,j,iphi,:] # Fhkl
fbulk_= tdat[4,i,j,iphi,:] # Fbulk
sij_ = tdat[6,i,j,iphi,:] # Sij
rsq_ = tdat[6,i,j,iphi,:] # R^2
sft[j,i,:] = fhkl_[::] #
rsqt[j,i,:] = rsq_[::]
i1,i2 = usf[j]
igo = ehkl_ - e_ ## old way
igt = []
m = 0
for k in range(len(fbulk_)):
if fbulk_[k]==0:
m=m+1
igt.append(0.)
else: igt.append(
ehkl_[k] - \
fhkl_[k] / fbulk_[k] * e_[k])
if m!=0: print m, 'case of fbulk'+\
'(phi,psi)=0 happened'
igt = np.array(igt)
igy = ehkl_ - fhkl_ * sij_ ## new way (YJ)
if j==0:
igy_avg = igy/nsf
igo_avg = igo/nsf
igt_avg = igt/nsf
else:
igy_avg = igy_avg + igy/nsf
igo_avg = igo_avg + igo/nsf
igt_avg = igt_avg + igt/nsf
igys.append(igy);igos.append(igo);igts.append(igt)
# The old method
if j==4: m = markers[j]
else: m= '--'
y = igo
x = np.sin(psi[:]*np.pi/180.)**2
newx,newy = sort(x,y) # shell sort
axes01[i].plot(
newx,newy*1e6,m,color=colors[j],
label=r'$\Sigma_{%1i%1i}$'%(i1,i2))
if i==0:axes01[i].legend(loc='best',
fancybox=True).get_frame().set_alpha(0.5)
if j==0:axes01[i].grid('on')
yl0, yh0 = axes01[i].get_ylim()
if yl0<yl: yl = yl0
if yh0>yh: yh = yh0
# GHT
y = igt
x = np.sin(psi[:]*np.pi/180.)**2
newx,newy = sort(x,y) # shell sort
axes02[i].plot(
newx,newy*1e6,m,color=colors[j],
label=r'$\Sigma_{%1i%1i}$'%(i1,i2))
if i==0:axes02[i].legend(loc='best',
fancybox=True).get_frame().set_alpha(0.5)
if j==0:axes02[i].grid('on')
yl0, yh0 = axes02[i].get_ylim()
if yl0<yl: yl = yl0
if yh0>yh: yh = yh0
# YJ
y = igy
x = np.sin(psi[:]*np.pi/180.)**2
newx,newy = sort(x,y) # shell sort
axes03[i].plot(
newx,newy*1e6,m,color=colors[j],
label=r'$\Sigma_{%1i%1i}$'%(i1,i2))
if i==0:axes03[i].legend(loc='best',
fancybox=True).get_frame().set_alpha(0.5)
if j==0:axes03[i].grid('on')
yl0, yh0 = axes03[i].get_ylim()
if yl0<yl: yl = yl0
if yh0>yh: yh = yh0
pass # over nsf
# errors for 6 cases of SF loading
igos = np.array(igos).swapaxes(0,1)
igys = np.array(igys).swapaxes(0,1)
igts = np.array(igts).swapaxes(0,1)
igoe=[]; igye=[]; igte=[]
for m in range(len(igo_avg)):
igoe.append(igos[m].std())
igte.append(igts[m].std())
igye.append(igys[m].std())
igoe,igte,igye = np.array(igoe),np.array(igte),np.array(igye)
##
# average old
y = igo_avg
ye= igoe
x = np.sin(psi[:]*np.pi/180.)**2
newx, newy, newye = sort(x,y,ye)
axes04[i].errorbar(
x=newx,y=newy*1e6,yerr=newye*1e6,
ls='-',color='r',label='old')
# average YU
y = igy_avg
avg[0,i,:] = psi[:]; avg[1,i,:] = y[:]
ye = igye
x = np.sin(psi[:]*np.pi/180.)**2
newx, newy, newye = sort(x,y,ye)
axes04[i].errorbar(
x=newx,y=newy*1e6,yerr=newye*1e6,
ls='-',ms=3,color='b',label='YJ')
axes04[i].grid('on')
axes04[0].legend(loc='best',fancybox=True)\
.get_frame().set_alpha(0.5)
i1,i2 = voigt.ijv[:,isf]
sin2psi = np.sin(psi*np.pi/180.)**2
ax01.plot(sin2psi,ige*1e6,markers[i]#,color='k',
,label=r'$\bar{E}^{\mathrm{eff}} = %4.2f$'%eps[i])
ax02.plot(sin2psi,ehkle*1e6,markers[i]#,color='k'
,label=r'$\bar{E}^{\mathrm{eff}} = %4.2f$'%eps[i])
l, = ax04.plot(sin2psi,fhkl*1e6,markers[i]#,color='k'
,label=r'$\bar{E}^{\mathrm{eff}} = %4.2f$'%eps[i])
if i==0: ax04.plot(sin2psi,fbulk*1e6,'-',alpha=0.2,color='k')
# ax03.plot(sin2psi,e*1e6,markers[i],
# label=r'$\bar{E}^{\mathrm{eff}} = %4.2f$'%eps[i])
ax03.plot(sin2psi,(ehkl-fhkl/fbulk*e)*1e6,markers[i]
,label=r'$\bar{E}^{\mathrm{eff}} = %4.2f$'%eps[i])
sin2psi = np.sin(psi*np.pi/180.)**2
## Deco the axes
for i in range(nst):
axes01[i].set_ylim(yl,yh)
axes02[i].set_ylim(yl,yh)
axes03[i].set_ylim(yl,yh)
#axes04[i].set_ylim(yl,yh)
axes01[i].set_xlim(0.,); axes02[i].set_xlim(0.,)
axes03[i].set_xlim(0.,); axes04[i].set_xlim(0.,)
for i in range(len(axes)):
axes[i].set_xlabel(r'$\sin^2{\psi}$', dict(fontsize=20))
axes01[i].set_xlabel(r'$\sin^2{\psi}$', dict(fontsize=20))
axes02[i].set_xlabel(r'$\sin^2{\psi}$', dict(fontsize=20))
axes03[i].set_xlabel(r'$\sin^2{\psi}$', dict(fontsize=20))
axes04[i].set_xlabel(r'$\sin^2{\psi}$', dict(fontsize=20))
for i in range(len(axes)): axes[i].grid('on')
for i in range(len(axes)): axes[i].set_xlim(0.,0.5)
ax01.set_ylabel(
r'$\varepsilon^{hkl} - F^{hkl}_{ij}\bar{\Sigma}_{ij}$'+\
r' $[\mu\varepsilon]$',dict(fontsize=15))
ax02.set_ylabel(
r'$\varepsilon^{hkl} - E$ $[\mu\varepsilon]$',dict(fontsize=15))
ax03.set_ylabel(r'$\varepsilon^{hkl} - $'+
r'$F_{ij}^{hkl}/F^{\mathrm{bulk}}_{ij}$'+
r'$ \bar{E} $ $[\mu \varepsilon]$',dict(fontsize=15))
ax04.set_ylabel(r'$F_{ij}^{hkl}, F^{\mathrm{bulk}}_{ij}$ $[TPa^{-1}]$',
dict(fontsize=15))
ax01.set_title(r'(a) $\varepsilon(hkl,%3.0f^\circ,\psi)$'\
r'$ - F_{%1i%1i}(hkl,%3.0f^\circ,\psi)$'\
r'$\bar{\Sigma}_{%1i%1i}$'%(phis[iphi],i1,i2,phis[iphi],i1,i2),
dict(fontsize=13))
ax02.set_title(
r'(b) $\varepsilon(hkl,%3.0f^\circ,\psi)$'%(phis[iphi])+\
r'$ - \bar{E}(%3.0f^\circ,\psi)$'%\
(phis[iphi]),dict(fontsize=13))
ax03.set_title(
r'(c) $\varepsilon(hkl,%3.0f^\circ,\psi)$'%(phis[iphi])+
r'$-F_{%1i%1i}(hkl,%3.0f^\circ,\psi)/$'%(i1,i2,phis[iphi])+
r'$F_{%1i%1i}^{\mathrm{bulk}}(%3.0f^\circ,\psi)$'%(
i1,i2,phis[iphi])+
r'$\bar{E}(%3.0f^\circ,\psi)$'%(phis[iphi]),dict(fontsize=13))
ax04.set_title(r'(d) $F^{hkl}_{%1i%1i}$ and '%(i1,i2)+\
r'$F^{\mathrm{bulk}}_{%1i%1i}$'%
(i1,i2), dict(fontsize=13))
ax04.legend(loc='lower right',fancybox=True).get_frame().set_alpha(0.5)
# fig.tight_layout(); fig01.tight_layout(); fig02.tight_layout()
# fig03.tight_layout(); fig04.tight_layout()
fig.savefig('ig_bulk.pdf');fig01.savefig('ig_bulk_Old.pdf')
fig02.savefig('ig_bulk_GHT.pdf'); fig03.savefig('ig_bulk_YJ.pdf')
fig04.savefig('ig_bulk_avg.pdf')
## save the averaged IG strain
return avg, sft # avg[2,nst,npsi], sft[6,nst,npsi]
def main_disc(fn='BB/2011NOV10/node_disc_36.csv',
dist_bb=[10],
dist_rb=[5],
fang=90.,
ixy_flip=False,
ioff=True,
ntot_col=14,
fig=None,
fout=None):
"""
fn = file name: " /dum/[strainpath]/[date].csv "
dist = distances from the critical point
fang = fracture angle
ixy_flip = flag if x-y cooridnate should be switched
ioff = interactive mode switch for MPL
ntot_col = total number of column belonging to a dat block
## below is the column information for a block
icol: 0 1 2 3 4 5 6 7 8 9 10 11 12
variable: x,y,z,exx,eyy,exy,e1,e2,vm,exxdot,eyydot,exydot,e33
"""
import os
from MP.lib.axes_label import __deco_fld__ as deco_fld
import matplotlib.pyplot as plt
from MP.lib.mpl_lib import wide_fig as wf
from MP.lib.mpl_lib import rm_all_lab as ral
from matplotlib.backends.backend_pdf import PdfPages
from readers import read_disc
## output file name
strain_path = fn.split(os.sep)[1]
date = fn.split(os.sep)[2].split('.')[0]
fn_out = '%s_%s' % (strain_path, date)
pdf_all = PdfPages('%s.pdf' % fn_out)
## MPL interactive mode switch
if ioff: plt.ioff()
else: plt.ion()
## Read data from VIC-3d software
dat = read_disc(fn=fn, ndat_col=ntot_col, ixy_flip=ixy_flip)
## through thickness strain rate minimum
fig1 = wf(nw=1,
nh=1,
w0=0.0,
ws=1.0,
w1=0.0,
h0=0.0,
hs=1.0,
h1=0.0,
uw=2.3,
uh=2.3,
left=0.2,
up=0.1,
down=0.2,
iarange=True)
n_image = 15 ## last n-th image
## from which maximum reference is taken
ax = fig1.axes[0]
fig2, fig3, mdat, c = plot_xy_disc_e33dot(dat,
istp=-n_image - 1,
icol=12,
ax=ax,
fang=fang,
dist_bb=dist_bb,
dist_rb=dist_rb)
ax.set_xlim(-40, 40)
ax.set_ylim(-40, 40)
ax.set_aspect('equal')
ral(fig1.axes)
col = ['path', 'date', 'dist [mm]', 'Exx', 'Eyy', 'Std1', 'Std2', 'color']
row = ['%9s', '%9s', '%9i', '%9.3f', '%9.3f', '%9.3f', '%9.3f', '%9s']
header = ''
fmt = ''
for i in range(len(col)):
header = '%s %s' % (header, '%9s' % col[i])
fmt = '%s %s' % (fmt, row[i])
print header
if fout != None: fout.write(header + '\n')
db = np.array(dist_bb)
dum = (db[::-1]).tolist() + [0]
ndat = mdat.shape[-1]
for i in range(ndat):
dat = mdat[:, :, i]
ex = dat[0, 0]
ey = dat[0, 1]
exe = dat[1, 0]
eye = dat[1, 1]
fig.axes[0].errorbar(ey, ex, xerr=eye, yerr=exe, fmt='.', color=c[i])
line = fmt % (strain_path, date, dum[i], ex, ey, exe, eye, c[i])
print line
if fout != None:
fout.write(line + '\n')
pdf_all.savefig(fig1)
# pdf_all.savefig(fig2)
pdf_all.savefig(fig3)
pdf_all.close()
deco_fld(ax=fig.axes[0], iopt=2, ft=15, iasp=True)
## if ioff: plt.close('all')
return fig1, fig2, fig3
def ex02(
## non-smooth
fns=[ #'Bsteel_non_smooth/BB/2011NOV10.csv',
'Bsteel_non_smooth/BB/2012JUL11.csv',
#'Bsteel_non_smooth/PSRD/2011NOV17.csv',
'Bsteel_non_smooth/PSRD/2012JUL12.csv',
'Bsteel_non_smooth/PSRD/2012JUL13.csv',
#'Bsteel_non_smooth/PSTD/2011NOV17.csv',
'Bsteel_non_smooth/PSTD/2012JUL17.csv'
],
fangs=[ #90,
90,
#0,
90,
90,
#90,
0
],
dist_bb=[2, 4, 6],
dist_rb=[2, 3, 4, 5, 6, 7, 8],
## dist=[5,10,15,20],
ixy_flip=[ #False,
False,
#False,
True,
True,
#False,
False
],
ntot_col=[ #14,
14,
#14,
14,
14,
#14,
14
],
ioff=True):
from MP.lib.mpl_lib import wide_fig as wf
import matplotlib.pyplot as plt
from MP.lib.axes_label import __deco_fld__ as deco_fld
fn = open('EXP_FLD.txt', 'w')
fig4 = wf(nw=1, nh=1)
for i in range(len(fns)):
#try:
## print fns[i]
fig1, fig2, fig3 = main_disc(fn=fns[i],
dist_bb=dist_bb,
dist_rb=dist_rb,
fang=fangs[i],
ixy_flip=ixy_flip[i],
ntot_col=ntot_col[i],
ioff=ioff,
fig=fig4,
fout=fn)
# except ValueError:
# print 'error occured in %s'%fns[i]
print 'Analysis results have been saved to %s' % fn.name
fn.close()
fig4.savefig('EXP_FLD.pdf')
plt.close('all')
def plot_xy_disc_e33dot(
dat,
istp=-1,
ax=None,
icol=11,
fang=90,
dist_bb=[5, 10], ## backbone
dist_rb=[1, 5]):
"""
Extract DIC data from 'disc' area
"""
from MP.lib.mpl_lib import wide_fig as wf
from MP.lib.mpl_lib import rm_all_lab as ral
import matplotlib as mpl
from MP.lib import mpl_lib
from MP import ssort
sort = ssort.shellSort
d = dat[istp, :, :]
dz = np.log10(abs(d[:, icol]))
for i in range(len(dz)):
if dz[i] == np.inf or dz[i] == -np.inf:
dz[i] = np.nan
mn = min(dz)
mx = max(dz)
norm = mpl.colors.Normalize(vmin=mn, vmax=mx)
cmap, m = mpl_lib.norm_cmap(mn=mn, mx=mx, cm_name='brg')
n = d.shape[0]
## find values in the back-bone branch shape
ids = []
xs = []
ys = []
zs = []
for i in range(n):
x, y, z = d[i, 0], d[i, 1], -d[i, icol]
if np.isnan(z): pass
else:
zs.append(d[i, icol])
ids.append(i)
xs.append(d[i, 0])
ys.append(d[i, 1])
c = m.to_rgba(np.log10(z))
ax.plot(x, y, 'o', ms=1.5, color=c, mfc=c, mec='None')
val, ind = sort(zs)
imn = ids[ind[0]]
ax.plot(d[imn, 0], d[imn, 1], 'g.') ## minimum
## draw a backbone grid
x0, y0 = d[imn, 0], d[imn, 1] # center of backbone
db = np.array(dist_bb)
dr = np.array(dist_rb)
DB = []
DR = []
for i in range(len(db)):
DB.append(-db[-i - 1])
DB.append(0)
for i in range(len(db)):
DB.append(db[i])
for i in range(len(dr)):
DR.append(-dr[-i - 1])
DR.append(0)
for i in range(len(dr)):
DR.append(dr[i])
db = DB[::]
dr = DR[::]
xy_db_dr = np.zeros((len(db), len(dr), 2))
for i in range(len(db)):
for j in range(len(dr)):
xy_db_dr[i, j, :] = db[i], dr[j]
## rotate xy by fang
for i in range(len(db)):
for j in range(len(dr)):
y, x = xy_db_dr[i, j, :]
xy_db_dr[i, j, :] = rot_xy([x, y], fang)
## translate
xy_db_dr[:, :, 0] = xy_db_dr[:, :, 0] + x0
xy_db_dr[:, :, 1] = xy_db_dr[:, :, 1] + y0
## print xy_db_dr.shape
## return xy_db_dr
if len(db) > 3:
nw = len(db) / 3
nh = 4
fig = wf(nw=nw,
nh=nh,
w0=0.0,
ws=1.0,
w1=0.0,
h0=0.0,
hs=1.0,
h1=0.0,
uw=2.3,
uh=2.3,
left=0.2,
up=0.1,
down=0.2,
iarange=True)
fig1 = wf(nw=2,
nh=2,
w0=0.1,
ws=0.7,
w1=0.2,
h0=0.1,
hs=0.7,
h1=0.2,
uw=2.3,
uh=2.3,
left=0.2,
up=0.1,
down=0.2,
iarange=True)
iax = 0
sym = [
'k^', 'k+', 'kd', 'kx', 'k<', 'k>', 'k*', 'k1', 'k2', 'k3', 'k4', 'k8',
'ks', 'kp', 'kh', 'kH', 'kD'
]
## thickness strain rate
epst_dot_av = []
epst_dot_std = []
## epsxx
exx_av = []
exx_std = []
## epsyy
eyy_av = []
eyy_std = []
for i in range(len(db)):
indices_along_ribs = []
for j in range(len(dr)):
x, y = xy_db_dr[i, j, :]
ind = find_nearest([x, y], xs, ys, err=2)
if not (np.isnan(ind)):
indices_along_ribs.append(ids[ind])
epsx = []
epsy = []
epst_dot = []
coord = []
for k in range(len(indices_along_ribs)):
indx = indices_along_ribs[k]
x = d[indx, 0]
y = d[indx, 1]
coord.append([x, y])
## ax.plot(x,y,sym[i],ms=3)
ex = dat[:, indx, 3]
ey = dat[:, indx, 4]
et = dat[:, indx, 12]
ex_d = dat[:, indx, 9]
ey_d = dat[:, indx, 10]
ez_d = ex_d + ey_d ## positive value
epst_dot.append(ez_d)
epsx.append(ex)
epsy.append(ey)
fig.axes[iax].plot(ez_d)
iax = iax + 1
x, y = np.array(coord).T[:]
ax.plot(x, y, '-', color='k', alpha=0.5)
epst_dot = np.array(epst_dot).T
epsx = np.array(epsx).T
epsy = np.array(epsy).T
D = []
D_e = []
E_ex = []
E_exe = []
E_ey = []
E_eye = []
for j in range(len(epst_dot)): ## number of ribs
D.append(np.mean(epst_dot[j]))
D_e.append(np.std(epst_dot[j]))
E_ex.append(np.mean(epsx[j]))
E_exe.append(np.std(epsx[j]))
E_ey.append(np.mean(epsy[j]))
E_eye.append(np.std(epsy[j]))
## fig1.axes[0].plot(D,label=db[i])
epst_dot_av.append(D)
epst_dot_std.append(D_e)
exx_av.append(E_ex)
exx_std.append(E_exe)
eyy_av.append(E_ey)
eyy_std.append(E_eye)
## find maximum D value's index to trim the data
ndat = (len(epst_dot_av) - 1) / 2 + 1
ref = epst_dot_av[ndat - 1] ## data at necking center spot
_ref_ = []
_ind_ = []
for i in range(len(ref)):
if not (np.isnan(ref[i])):
_ref_.append(ref[i])
_ind_.append(i)
_val_, _i_ = sort(_ref_)
ix = _ind_[_i_[-1]]
mdat = np.zeros((2, 2, ndat))
c = []
for i in range(ndat):
i0 = i
i1 = -i - 1
## thickness strain rate
d1 = epst_dot_av[i0]
d2 = epst_dot_av[i1]
e1 = epst_dot_std[i0]
e2 = epst_dot_std[i1]
d = []
e = []
## epsilon xx
ex1 = exx_av[i0]
ex2 = exx_av[i1]
exe1 = exx_std[i0]
exe2 = exx_std[i1]
ex = []
exe = []
## epsilon yy
ey1 = eyy_av[i0]
ey2 = eyy_av[i1]
eye1 = eyy_std[i0]
eye2 = eyy_std[i1]
ey = []
eye = []
for j in range(len(d1)): # along the time axis
d.append((d1[j] + d2[j]) / 2.)
e.append((e1[j] + e2[j]) / 2.)
ex.append((ex1[j] + ex2[j]) / 2.)
exe.append((exe1[j] + exe2[j]) / 2.)
ey.append((ey1[j] + ey2[j]) / 2.)
eye.append((eye1[j] + eye2[j]) / 2.)
xind = np.arange(len(d)) + 1
p = 0.7
i0 = int(float(len(xind)) * p)
## find maximum value and index?
l, = fig1.axes[0].plot(xind, d, '-', label=abs(db[-i - 1]))
fig1.axes[1].plot(xind[i0:ix + 1], d[i0:ix + 1], '+')
fig1.axes[2].errorbar(xind[i0:ix + 1],
d[i0:ix + 1],
yerr=e1[i0:ix + 1],
fmt='+')
l, = fig1.axes[3].plot(ex[i0:ix + 1], ey[i0:ix + 1])
fig1.axes[3].errorbar(ex[ix],
ey[ix],
color=l.get_color(),
xerr=exe[ix],
yerr=eye[ix])
mdat[0, 0, i] = ex[ix]
mdat[0, 1, i] = ey[ix]
mdat[1, 0, i] = exe[ix]
mdat[1, 1, i] = eye[ix]
c.append(l.get_color())
## deco
from MP.lib.mpl_lib import tune_xy_lim
tune_xy_lim(fig.axes)
fig1.axes[2].set_yscale('log')
fig1.axes[2].set_ylim(1e-3)
fig1.axes[0].legend(loc='best',ncol=2,fontsize=5).\
get_frame().set_alpha(0.5)
return fig, fig1, mdat, c
def main_disc(fn='BB/2011NOV10/node_disc_36.csv',
dist_bb=[10],
dist_rb=[5],
fang=90.,ixy_flip=False,ioff=True,
ntot_col=14,
fig=None,
fout=None):
"""
fn = file name: " /dum/[strainpath]/[date].csv "
dist = distances from the critical point
fang = fracture angle
ixy_flip = flag if x-y cooridnate should be switched
ioff = interactive mode switch for MPL
ntot_col = total number of column belonging to a dat block
## below is the column information for a block
icol: 0 1 2 3 4 5 6 7 8 9 10 11 12
variable: x,y,z,exx,eyy,exy,e1,e2,vm,exxdot,eyydot,exydot,e33
"""
import os
from MP.lib.axes_label import __deco_fld__ as deco_fld
import matplotlib.pyplot as plt
from MP.lib.mpl_lib import wide_fig as wf
from MP.lib.mpl_lib import rm_all_lab as ral
from matplotlib.backends.backend_pdf import PdfPages
from readers import read_disc
## output file name
strain_path = fn.split(os.sep)[1]
date = fn.split(os.sep)[2].split('.')[0]
fn_out = '%s_%s'%(strain_path,date)
pdf_all = PdfPages('%s.pdf'%fn_out)
## MPL interactive mode switch
if ioff: plt.ioff()
else: plt.ion()
## Read data from VIC-3d software
dat=read_disc(fn=fn,ndat_col=ntot_col,
ixy_flip=ixy_flip)
## through thickness strain rate minimum
fig1=wf(nw=1,nh=1,w0=0.0,ws=1.0,w1=0.0,
h0=0.0,hs=1.0,h1=0.0,uw=2.3,uh=2.3,
left=0.2,up=0.1,down=0.2,iarange=True)
n_image=15 ## last n-th image
## from which maximum reference is taken
ax=fig1.axes[0]
fig2,fig3,mdat,c=plot_xy_disc_e33dot(
dat,istp=-n_image-1,icol=12,
ax=ax,fang=fang,dist_bb=dist_bb,
dist_rb=dist_rb)
ax.set_xlim(-40,40);ax.set_ylim(-40,40)
ax.set_aspect('equal')
ral(fig1.axes)
col=['path','date','dist [mm]','Exx','Eyy','Std1','Std2','color']
row=['%9s','%9s','%9i','%9.3f','%9.3f','%9.3f','%9.3f','%9s']
header =''
fmt =''
for i in range(len(col)):
header='%s %s'%(header,'%9s'%col[i])
fmt ='%s %s'%(fmt,row[i])
print header
if fout!=None: fout.write(header+'\n')
db=np.array(dist_bb)
dum = (db[::-1]).tolist()+[0]
ndat = mdat.shape[-1]
for i in range(ndat):
dat = mdat[:,:,i]
ex = dat[0,0]
ey = dat[0,1]
exe= dat[1,0]
eye= dat[1,1]
fig.axes[0].errorbar(ey,ex,xerr=eye,yerr=exe,fmt='.',
color=c[i])
line= fmt%(strain_path,date,dum[i],ex,ey,exe,eye,c[i])
print line
if fout!=None:
fout.write(line + '\n')
pdf_all.savefig(fig1)
# pdf_all.savefig(fig2)
pdf_all.savefig(fig3)
pdf_all.close()
deco_fld(ax=fig.axes[0],iopt=2,ft=15,iasp=True)
## if ioff: plt.close('all')
return fig1, fig2, fig3
def ex02(
## non-smooth
fns=[#'Bsteel_non_smooth/BB/2011NOV10.csv',
'Bsteel_non_smooth/BB/2012JUL11.csv',
#'Bsteel_non_smooth/PSRD/2011NOV17.csv',
'Bsteel_non_smooth/PSRD/2012JUL12.csv',
'Bsteel_non_smooth/PSRD/2012JUL13.csv',
#'Bsteel_non_smooth/PSTD/2011NOV17.csv',
'Bsteel_non_smooth/PSTD/2012JUL17.csv'],
fangs=[#90,
90,
#0,
90,
90,
#90,
0],
dist_bb=[2,4,6],
dist_rb=[2,3,4,5,6,7,8],
## dist=[5,10,15,20],
ixy_flip=[#False,
False,
#False,
True,
True,
#False,
False],
ntot_col=[#14,
14,
#14,
14,
14,
#14,
14],
ioff=True):
from MP.lib.mpl_lib import wide_fig as wf
import matplotlib.pyplot as plt
from MP.lib.axes_label import __deco_fld__ as deco_fld
fn = open('EXP_FLD.txt','w')
fig4=wf(nw=1,nh=1)
for i in range(len(fns)):
#try:
## print fns[i]
fig1, fig2, fig3 = main_disc(
fn=fns[i],
dist_bb=dist_bb,
dist_rb=dist_rb,
fang=fangs[i],
ixy_flip=ixy_flip[i],
ntot_col=ntot_col[i],
ioff=ioff,fig=fig4,fout=fn)
# except ValueError:
# print 'error occured in %s'%fns[i]
print 'Analysis results have been saved to %s'%fn.name
fn.close()
fig4.savefig('EXP_FLD.pdf')
plt.close('all')
def plot_xy_disc_e33dot(
dat,istp=-1,ax=None,icol=11,fang=90,
dist_bb=[5,10], ## backbone
dist_rb=[1,5]):
"""
Extract DIC data from 'disc' area
"""
from MP.lib.mpl_lib import wide_fig as wf
from MP.lib.mpl_lib import rm_all_lab as ral
import matplotlib as mpl
from MP.lib import mpl_lib
from MP import ssort
sort=ssort.shellSort
d=dat[istp,:,:]
dz = np.log10(abs(d[:,icol]))
for i in range(len(dz)):
if dz[i]==np.inf or dz[i]==-np.inf:
dz[i] = np.nan
mn=min(dz);mx=max(dz)
norm=mpl.colors.Normalize(vmin=mn, vmax=mx)
cmap, m = mpl_lib.norm_cmap(
mn=mn,mx=mx,cm_name='brg')
n=d.shape[0]
## find values in the back-bone branch shape
ids = []; xs = []; ys = []; zs = []
for i in range(n):
x,y,z=d[i,0],d[i,1],-d[i,icol]
if np.isnan(z): pass
else:
zs.append(d[i,icol])
ids.append(i)
xs.append(d[i,0])
ys.append(d[i,1])
c = m.to_rgba(np.log10(z))
ax.plot(x,y,'o',ms=1.5,
color=c,mfc=c,mec='None')
val, ind = sort(zs)
imn = ids[ind[0]]
ax.plot(d[imn,0],d[imn,1],'g.') ## minimum
## draw a backbone grid
x0,y0=d[imn,0],d[imn,1] # center of backbone
db=np.array(dist_bb); dr=np.array(dist_rb)
DB=[]; DR=[]
for i in range(len(db)):
DB.append(-db[-i-1])
DB.append(0)
for i in range(len(db)):
DB.append(db[i])
for i in range(len(dr)):
DR.append(-dr[-i-1])
DR.append(0)
for i in range(len(dr)):
DR.append(dr[i])
db=DB[::];dr=DR[::]
xy_db_dr = np.zeros((len(db),len(dr),2))
for i in range(len(db)):
for j in range(len(dr)):
xy_db_dr[i,j,:] = db[i], dr[j]
## rotate xy by fang
for i in range(len(db)):
for j in range(len(dr)):
y,x = xy_db_dr[i,j,:]
xy_db_dr[i,j,:] = rot_xy([x,y],fang)
## translate
xy_db_dr[:,:,0] = xy_db_dr[:,:,0] + x0
xy_db_dr[:,:,1] = xy_db_dr[:,:,1] + y0
## print xy_db_dr.shape
## return xy_db_dr
if len(db)>3:
nw = len(db)/3
nh = 4
fig=wf(nw=nw,nh=nh,w0=0.0,ws=1.0,w1=0.0,
h0=0.0,hs=1.0,h1=0.0,uw=2.3,uh=2.3,
left=0.2,up=0.1,down=0.2,iarange=True)
fig1=wf(nw=2,nh=2,w0=0.1,ws=0.7,w1=0.2,
h0=0.1,hs=0.7,h1=0.2,uw=2.3,uh=2.3,
left=0.2,up=0.1,down=0.2,iarange=True)
iax=0
sym=['k^','k+','kd','kx','k<','k>','k*',
'k1','k2','k3','k4','k8','ks','kp',
'kh','kH','kD']
## thickness strain rate
epst_dot_av = []; epst_dot_std = []
## epsxx
exx_av = []; exx_std = []
## epsyy
eyy_av = []; eyy_std = []
for i in range(len(db)):
indices_along_ribs = []
for j in range(len(dr)):
x,y = xy_db_dr[i,j,:]
ind = find_nearest([x,y],xs,ys,err=2)
if not(np.isnan(ind)):
indices_along_ribs.append(ids[ind])
epsx = []; epsy = []; epst_dot = []
coord = []
for k in range(len(indices_along_ribs)):
indx = indices_along_ribs[k]
x = d[indx,0]; y = d[indx,1]
coord.append([x,y])
## ax.plot(x,y,sym[i],ms=3)
ex=dat[:,indx,3]
ey=dat[:,indx,4]
et=dat[:,indx,12]
ex_d=dat[:,indx,9]
ey_d=dat[:,indx,10]
ez_d=ex_d+ey_d ## positive value
epst_dot.append(ez_d)
epsx.append(ex)
epsy.append(ey)
fig.axes[iax].plot(ez_d)
iax = iax + 1
x,y = np.array(coord).T[:]
ax.plot(x,y,'-',color='k',alpha=0.5)
epst_dot=np.array(epst_dot).T
epsx =np.array(epsx).T
epsy =np.array(epsy).T
D = []; D_e = [];
E_ex = []; E_exe = [];
E_ey = []; E_eye = [];
for j in range(len(epst_dot)): ## number of ribs
D.append(np.mean(epst_dot[j]))
D_e.append(np.std(epst_dot[j]))
E_ex.append(np.mean(epsx[j]))
E_exe.append(np.std(epsx[j]))
E_ey.append(np.mean(epsy[j]))
E_eye.append(np.std(epsy[j]))
## fig1.axes[0].plot(D,label=db[i])
epst_dot_av.append(D)
epst_dot_std.append(D_e)
exx_av.append(E_ex)
exx_std.append(E_exe)
eyy_av.append(E_ey)
eyy_std.append(E_eye)
## find maximum D value's index to trim the data
ndat = (len(epst_dot_av)-1)/2+1
ref=epst_dot_av[ndat-1] ## data at necking center spot
_ref_=[]; _ind_ =[]
for i in range(len(ref)):
if not(np.isnan(ref[i])):
_ref_.append(ref[i]); _ind_.append(i)
_val_, _i_ = sort(_ref_)
ix = _ind_[_i_[-1]]
mdat = np.zeros((2,2,ndat))
c = []
for i in range(ndat):
i0 = i
i1 = -i -1
## thickness strain rate
d1=epst_dot_av[i0]; d2=epst_dot_av[i1]
e1=epst_dot_std[i0]; e2=epst_dot_std[i1]
d =[]; e=[]
## epsilon xx
ex1 = exx_av[i0]; ex2 = exx_av[i1]
exe1 = exx_std[i0]; exe2 = exx_std[i1]
ex=[]; exe=[]
## epsilon yy
ey1 = eyy_av[i0]; ey2 = eyy_av[i1]
eye1 = eyy_std[i0]; eye2 = eyy_std[i1]
ey=[]; eye=[]
for j in range(len(d1)): # along the time axis
d.append((d1[j]+d2[j])/2.)
e.append((e1[j]+e2[j])/2.)
ex.append((ex1[j]+ex2[j])/2.)
exe.append((exe1[j]+exe2[j])/2.)
ey.append((ey1[j]+ey2[j])/2.)
eye.append((eye1[j]+eye2[j])/2.)
xind=np.arange(len(d))+1
p = 0.7
i0 = int(float(len(xind))*p)
## find maximum value and index?
l, =fig1.axes[0].plot(xind,d,'-',label=abs(db[-i-1]))
fig1.axes[1].plot(xind[i0:ix+1],d[i0:ix+1],'+')
fig1.axes[2].errorbar(xind[i0:ix+1],d[i0:ix+1],
yerr=e1[i0:ix+1],fmt='+')
l, =fig1.axes[3].plot(ex[i0:ix+1],ey[i0:ix+1])
fig1.axes[3].errorbar(
ex[ix],ey[ix],color=l.get_color(),
xerr=exe[ix],yerr=eye[ix])
mdat[0,0,i] = ex[ix]
mdat[0,1,i] = ey[ix]
mdat[1,0,i] = exe[ix]
mdat[1,1,i] = eye[ix]
c.append(l.get_color())
## deco
from MP.lib.mpl_lib import tune_xy_lim
tune_xy_lim(fig.axes)
fig1.axes[2].set_yscale('log')
fig1.axes[2].set_ylim(1e-3)
fig1.axes[0].legend(loc='best',ncol=2,fontsize=5).\
get_frame().set_alpha(0.5)
return fig,fig1,mdat,c
def ex_igb(fn='igstrain_fbulk_ph1.out',ifig=2,iphi=0,isf=0,
mxnst=None,flow=None):
"""
Read ig strain file and analyze... (igstrain_fbulk_ph1.out)
igstrain_fbulk_ph1.out contains diffraction for each and every
reloading from 'unloaded' polycrystal.
Arguments
=========
fn = 'igstrain_fbulk_ph1.out'
ifig = 2
iphi=0
isf=0
mxnst=None
"""
import matplotlib.pyplot as plt
from sff_converter import condition
from MP.ssort import sh as sort
from MP.mat import voigt
from MP.lib.mpl_lib import wide_fig as wf
difl, nphi, phis, nbeta, dum1, dum2 = condition(fn=None)
if flow==None: raise IOError, 'Flow stress is missing'
eps = flow.epsilon_vm[::]
neps = flow.nstp
nw = 4
remainder = np.mod(neps,nw)
if remainder==0:
nh = neps/nw
elif remainder>0:
nh = neps/nw + 1
fig =wf(nw=2, nh=2,ifig=ifig,iarange=True); axes =fig.axes
ax01, ax02, ax03, ax04 = axes
fig01=wf(nw=nw,nh=nh,ifig=ifig+1,iarange=True,nfig=neps);axes01=fig01.axes
fig02=wf(nw=nw,nh=nh,ifig=ifig+2,iarange=True,nfig=neps);axes02=fig02.axes
fig03=wf(nw=nw,nh=nh,ifig=ifig+3,iarange=True,nfig=neps);axes03=fig03.axes
fig04=wf(nw=nw,nh=nh,ifig=ifig+4,iarange=True,nfig=neps);axes04=fig04.axes
tdat, usf = reader2(fn,iopt=1)
print tdat.shape
nsf = len(usf); nst = len(tdat[0]); npsi = len(tdat[0,0,0,0,:])
if nst!=neps: raise IOError, 'Inconsistency between neps and nst'
markers = ['o','x','+','^','d','*','o','x','+','^','d','*''o','x','+','^','d','*']
colors = ['r','g','b','k','m','y','r','g','b','k','m','y','r','g','b','k','m','y']
from string import ascii_lowercase as alphabet
axe_lab=[]
for ab in alphabet: axe_lab.append('(%s)'%ab)
yl = 0; yh = 0
avg = np.zeros((2,nst,npsi)) #
sft = np.zeros((6,nst,npsi)) # f^{hkl}
rsqt = np.zeros((6,nst,npsi)) # f^{hkl}
if mxnst!=None: nst=mxnst
for i in xrange(nst):
ehkl = tdat[0,i,isf,iphi,:] # e(hkl,phi,psi)
e = tdat[1,i,isf,iphi,:] # macro
ehkle = tdat[2,i,isf,iphi,:] # e - macro
fhkl = tdat[3,i,isf,iphi,:] # Fhkl
fbulk = tdat[4,i,isf,iphi,:] # Fbulk
ige = tdat[5,i,isf,iphi,:] # e - F_ij *Sij
sij = tdat[6,i,isf,iphi,:] # Sij
psi = tdat[8,i,isf,iphi,:]
rsq = tdat[9,i,isf,iphi,:]
sin2psi = np.sin(psi*np.pi/180.)**2
axes01[i].set_title(r'%s $\bar{E}^{\mathrm{eff}} = %4.2f$'%(
axe_lab[i],eps[i]))
axes02[i].set_title(r'%s $\bar{E}^{\mathrm{eff}} = %4.2f$'%(
axe_lab[i],eps[i]))
axes03[i].set_title(r'%s $\bar{E}^{\mathrm{eff}} = %4.2f$'%(
axe_lab[i],eps[i]))
axes04[i].set_title(r'%s $\bar{E}^{\mathrm{eff}} = %4.2f$'%(
axe_lab[i],eps[i]))
igys = []; igos = []; igts = []
for j in xrange(nsf): ## Number of elastic loading for stress factor calc.
ehkl_ = tdat[0,i,j,iphi,:] # e(hkl,phi,psi)
e_ = tdat[1,i,j,iphi,:] # macro
fhkl_ = tdat[3,i,j,iphi,:] # Fhkl
fbulk_= tdat[4,i,j,iphi,:] # Fbulk
sij_ = tdat[6,i,j,iphi,:] # Sij
rsq_ = tdat[6,i,j,iphi,:] # R^2
sft[j,i,:] = fhkl_[::] #
rsqt[j,i,:] = rsq_[::]
i1,i2 = usf[j]
igo = ehkl_ - e_ ## old way
igt = []
m = 0
for k in xrange(len(fbulk_)):
if fbulk_[k]==0:
m=m+1
igt.append(0.)
else: igt.append(
ehkl_[k] - \
fhkl_[k] / fbulk_[k] * e_[k])
if m!=0: print m, 'case of fbulk'+\
'(phi,psi)=0 happened'
igt = np.array(igt)
igy = ehkl_ - fhkl_ * sij_ ## new way (YJ)
if j==0:
igy_avg = igy/nsf
igo_avg = igo/nsf
igt_avg = igt/nsf
else:
igy_avg = igy_avg + igy/nsf
igo_avg = igo_avg + igo/nsf
igt_avg = igt_avg + igt/nsf
igys.append(igy);igos.append(igo);igts.append(igt)
# The old method
if j==4: m = markers[j]
else: m= '--'
y = igo
x = np.sin(psi[:]*np.pi/180.)**2
newx,newy = sort(x,y) # shell sort
axes01[i].plot(
newx,newy*1e6,m,color=colors[j],
label=r'$\Sigma_{%1i%1i}$'%(i1,i2))
if i==0:axes01[i].legend(loc='best',
fancybox=True).get_frame().set_alpha(0.5)
if j==0:axes01[i].grid('on')
yl0, yh0 = axes01[i].get_ylim()
if yl0<yl: yl = yl0
if yh0>yh: yh = yh0
# GHT
y = igt
x = np.sin(psi[:]*np.pi/180.)**2
newx,newy = sort(x,y) # shell sort
axes02[i].plot(
newx,newy*1e6,m,color=colors[j],
label=r'$\Sigma_{%1i%1i}$'%(i1,i2))
if i==0:axes02[i].legend(loc='best',
fancybox=True).get_frame().set_alpha(0.5)
if j==0:axes02[i].grid('on')
yl0, yh0 = axes02[i].get_ylim()
if yl0<yl: yl = yl0
if yh0>yh: yh = yh0
# YJ
y = igy
x = np.sin(psi[:]*np.pi/180.)**2
newx,newy = sort(x,y) # shell sort
axes03[i].plot(
newx,newy*1e6,m,color=colors[j],
label=r'$\Sigma_{%1i%1i}$'%(i1,i2))
if i==0:axes03[i].legend(loc='best',
fancybox=True).get_frame().set_alpha(0.5)
if j==0:axes03[i].grid('on')
yl0, yh0 = axes03[i].get_ylim()
if yl0<yl: yl = yl0
if yh0>yh: yh = yh0
pass # over nsf
# errors for 6 cases of SF loading
igos = np.array(igos).swapaxes(0,1)
igys = np.array(igys).swapaxes(0,1)
igts = np.array(igts).swapaxes(0,1)
igoe=[]; igye=[]; igte=[]
for m in xrange(len(igo_avg)):
igoe.append(igos[m].std())
igte.append(igts[m].std())
igye.append(igys[m].std())
igoe,igte,igye = np.array(igoe),np.array(igte),np.array(igye)
##
# average old
y = igo_avg
ye= igoe
x = np.sin(psi[:]*np.pi/180.)**2
newx, newy, newye = sort(x,y,ye)
axes04[i].errorbar(
x=newx,y=newy*1e6,yerr=newye*1e6,
ls='-',color='r',label='old')
# average YU
y = igy_avg
avg[0,i,:] = psi[:]; avg[1,i,:] = y[:]
ye = igye
x = np.sin(psi[:]*np.pi/180.)**2
newx, newy, newye = sort(x,y,ye)
axes04[i].errorbar(
x=newx,y=newy*1e6,yerr=newye*1e6,
ls='-',ms=3,color='b',label='YJ')
axes04[i].grid('on')
axes04[0].legend(loc='best',fancybox=True)\
.get_frame().set_alpha(0.5)
i1,i2 = voigt.ijv[:,isf]
sin2psi = np.sin(psi*np.pi/180.)**2
ax01.plot(sin2psi,ige*1e6,markers[i]#,color='k',
,label=r'$\bar{E}^{\mathrm{eff}} = %4.2f$'%eps[i])
ax02.plot(sin2psi,ehkle*1e6,markers[i]#,color='k'
,label=r'$\bar{E}^{\mathrm{eff}} = %4.2f$'%eps[i])
l, = ax04.plot(sin2psi,fhkl*1e6,markers[i]#,color='k'
,label=r'$\bar{E}^{\mathrm{eff}} = %4.2f$'%eps[i])
if i==0: ax04.plot(sin2psi,fbulk*1e6,'-',alpha=0.2,color='k')
# ax03.plot(sin2psi,e*1e6,markers[i],
# label=r'$\bar{E}^{\mathrm{eff}} = %4.2f$'%eps[i])
ax03.plot(sin2psi,(ehkl-fhkl/fbulk*e)*1e6,markers[i]
,label=r'$\bar{E}^{\mathrm{eff}} = %4.2f$'%eps[i])
sin2psi = np.sin(psi*np.pi/180.)**2
## Deco the axes
for i in xrange(nst):
axes01[i].set_ylim(yl,yh)
axes02[i].set_ylim(yl,yh)
axes03[i].set_ylim(yl,yh)
#axes04[i].set_ylim(yl,yh)
axes01[i].set_xlim(0.,); axes02[i].set_xlim(0.,)
axes03[i].set_xlim(0.,); axes04[i].set_xlim(0.,)
for i in xrange(len(axes)):
axes[i].set_xlabel(r'$\sin^2{\psi}$', dict(fontsize=20))
axes01[i].set_xlabel(r'$\sin^2{\psi}$', dict(fontsize=20))
axes02[i].set_xlabel(r'$\sin^2{\psi}$', dict(fontsize=20))
axes03[i].set_xlabel(r'$\sin^2{\psi}$', dict(fontsize=20))
axes04[i].set_xlabel(r'$\sin^2{\psi}$', dict(fontsize=20))
for i in xrange(len(axes)): axes[i].grid('on')
for i in xrange(len(axes)): axes[i].set_xlim(0.,0.5)
ax01.set_ylabel(
r'$\varepsilon^{hkl} - F^{hkl}_{ij}\bar{\Sigma}_{ij}$'+\
r' $[\mu\varepsilon]$',dict(fontsize=15))
ax02.set_ylabel(
r'$\varepsilon^{hkl} - E$ $[\mu\varepsilon]$',dict(fontsize=15))
ax03.set_ylabel(r'$\varepsilon^{hkl} - $'+
r'$F_{ij}^{hkl}/F^{\mathrm{bulk}}_{ij}$'+
r'$ \bar{E} $ $[\mu \varepsilon]$',dict(fontsize=15))
ax04.set_ylabel(r'$F_{ij}^{hkl}, F^{\mathrm{bulk}}_{ij}$ $[TPa^{-1}]$',
dict(fontsize=15))
ax01.set_title(r'(a) $\varepsilon(hkl,%3.0f^\circ,\psi)$'\
r'$ - F_{%1i%1i}(hkl,%3.0f^\circ,\psi)$'\
r'$\bar{\Sigma}_{%1i%1i}$'%(phis[iphi],i1,i2,phis[iphi],i1,i2),
dict(fontsize=13))
ax02.set_title(
r'(b) $\varepsilon(hkl,%3.0f^\circ,\psi)$'%(phis[iphi])+\
r'$ - \bar{E}(%3.0f^\circ,\psi)$'%\
(phis[iphi]),dict(fontsize=13))
ax03.set_title(
r'(c) $\varepsilon(hkl,%3.0f^\circ,\psi)$'%(phis[iphi])+
r'$-F_{%1i%1i}(hkl,%3.0f^\circ,\psi)/$'%(i1,i2,phis[iphi])+
r'$F_{%1i%1i}^{\mathrm{bulk}}(%3.0f^\circ,\psi)$'%(
i1,i2,phis[iphi])+
r'$\bar{E}(%3.0f^\circ,\psi)$'%(phis[iphi]),dict(fontsize=13))
ax04.set_title(r'(d) $F^{hkl}_{%1i%1i}$ and '%(i1,i2)+\
r'$F^{\mathrm{bulk}}_{%1i%1i}$'%
(i1,i2), dict(fontsize=13))
ax04.legend(loc='lower right',fancybox=True).get_frame().set_alpha(0.5)
# fig.tight_layout(); fig01.tight_layout(); fig02.tight_layout()
# fig03.tight_layout(); fig04.tight_layout()
fig.savefig('ig_bulk.pdf');fig01.savefig('ig_bulk_Old.pdf')
fig02.savefig('ig_bulk_GHT.pdf'); fig03.savefig('ig_bulk_YJ.pdf')
fig04.savefig('ig_bulk_avg.pdf')
## save the averaged IG strain
return avg, sft # avg[2,nst,npsi], sft[6,nst,npsi]
