def process_file(self, normfiledata):
""" Process a given SPECS file. Iterate through the regions and if the scan
mode is ConstantFinalState we normalize it.
"""
# Try to make a folder called rootfile-NEXAFS."
dfname = self.rootfile + "-NEXAFS"
if os.path.exists(dfname):
print "Error: Unpack folder already exists. Delete it and then click Go again."
return
else:
os.mkdir(dfname)
olddir = os.getcwd()
os.chdir(dfname)
for g in self.specsobj.groups:
if DEBUG:
print "=====================", g.name, "====================="
# Make a folder with the group name - need to make sure it's unique.
have_name = False
name_counter = 0
while not have_name:
if name_counter == 0:
tmpname = g.name
else:
tmpname = g.name + "-" + str(name_counter)
if not os.path.exists(tmpname):
os.mkdir(tmpname)
os.chdir(tmpname)
have_name = True
name_counter = 0
else:
# Darn, the directory prob already exists.
name_counter += 1
for r in g.regions:
newdata = None
if r.scan_mode == "ConstantFinalState":
newdata = self.process_region(r, normfiledata)
print "Processing region: ", r.name
else:
print "Region %s is not NEXAFS - skipping." % (r.name)
if newdata is not None:
laey = None
ltey = None
lpey = None
if self.preedge_treatment == "linear":
laey = specs.preedge_calculate(newdata[:,0], newdata[:,9])
ltey = specs.preedge_calculate(newdata[:,0], newdata[:,10])
lpey = specs.preedge_calculate(newdata[:,0], newdata[:,11])
if laey is None:
laey = newdata[0,9] * ones(newdata[:,9].shape)
if ltey is None:
ltey = newdata[0,10] * ones(newdata[:,10].shape)
if lpey is None:
lpey = newdata[0,11] * ones(newdata[:,11].shape)
subdata = zeros((newdata.shape[0], newdata.shape[1] + 6))
subdata[:,:12] = newdata
subdata[:,12] = laey
subdata[:,13] = ltey
subdata[:,14] = lpey
try:
subdata[:,15] = newdata[:,9] - laey
except TypeError:
subdata[:,15] = newdata[:,9]
print "Couldn't subtract a linear pre-edge from the aey spectrum for this region."
try:
subdata[:,16] = newdata[:,10] - ltey
except TypeError:
subdata[:,16] = newdata[:,10]
print "Couldn't subtract a linear pre-edge from the tey spectrum for this region."
try:
subdata[:,17] = newdata[:,11] - lpey
except TypeError:
subdata[:,17] = newdata[:,11]
print "Couldn't subtract a linear pre-edge from the pey spectrum for this region."
# Individually normalize the components
#try:
# subdata[:,15] = subdata[:,15] / subdata[-1,15]
#except FloatingPointError:
# print "The AEY channel cannot be 1-normalized for this file."
#try:
# subdata[:,16] = subdata[:,16] / subdata[-1,16]
#except FloatingPointError:
# print "The TEY channel cannot be 1-normalized for this file."
#try:
# subdata[:,17] = subdata[:,17] / subdata[-1,17]
#except FloatingPointError:
# print "The PEY channel cannot be 1-normalized for this file."
have_name = False
name_counter = 0
while not have_name:
if name_counter == 0:
r_write_name = r.name + ".xy"
else:
r_write_name = r.name + "-" + str(name_counter) + ".xy"
if not os.path.exists(r_write_name):
have_name = True
name_counter = 0
else:
# Darn, the directory prob already exists.
name_counter += 1
savetxt(r_write_name, subdata)
# If we didn't actually put any files in this folder, delete it.
tmpdir = os.getcwd()
os.chdir(dfname)
if os.listdir(tmpdir) == []:
os.rmdir(tmpdir)
os.chdir(olddir)
def loadxml(self):
xmlfile = tkFileDialog.askopenfilename(title="Choose a SPECS .xml file.", initialdir=os.getcwd())
if xmlfile != None:
xmlcontents = specs.SPECS(xmlfile)
# Make a folder: call it "xmlfile"-unpacked
base_path = os.path.split(xmlfile)[0]
cwd = os.getcwd()
# Remove whitespace either side of the name. It's important. :o)
unpack_path = os.path.join(base_path, xmlfile+"-unpacked").strip()
os.mkdir(unpack_path)
os.chdir(unpack_path)
for g in xmlcontents.groups:
if DEBUG:
print "=================== OUTPUT: %s ====================" % (g.name)
# We want to make a folder matching the group name, but we might not
# be able to because it might already exist (group names are not
# required to be unique in SPECS). So we have to fiddle here a bit.
have_name = False
name_counter = 0
while not have_name:
if name_counter == 0:
tmpname = g.name.strip()
else:
tmpname = (g.name + "-" + str(name_counter)).strip()
if not os.path.exists(tmpname):
os.mkdir(tmpname)
os.chdir(tmpname)
have_name = True
name_counter = 0
else:
# Darn, the directory prob already exists.
name_counter += 1
for r in g.regions:
if DEBUG:
print "Constructing output for region %s..." % (r.name)
# Decide how many outputs based on whether we have extended channels
# and so on.
have_nine_channels = False
num_outputs = 2 # Bare output is energy and counts
if r.channel_counts.shape[1] > 1:
have_nine_channels = True
num_outputs += r.channel_counts.shape[1]
chan_out_idx = 2 + r.channel_counts.shape[1]
if r.extended_channels is not None:
# Output all extended channels plus counts / I0.
ex_out_sidx = num_outputs
num_outputs += r.extended_channels.shape[1] + 1
ex_out_fidx = ex_out_sidx + 9
# For now, automatically add the shirley and linear subtractions.
# Eventually this will be an option.
num_outputs += 5 # Linear preedge, background
# scaling factor, shirley, counts - L - S and
# (counts - L - S) / I0.
data = zeros((r.values_per_curve,num_outputs))
hdr_string = ""
if r.scan_mode == "FixedAnalyzerTransmission":
data[:,0] = r.binding_axis
x = r.binding_axis
hdr_string += "Binding_energy_(eV) "
elif r.scan_mode == "ConstantFinalState":
data[:,0] = r.excitation_axis
x = r.excitation_axis
hdr_string += "Photon_energy_(eV) "
else:
data[:,0] = r.kinetic_axis
x = r.kinetic_axis
hdr_string += "Kinetic_energy_(eV) "
data[:,1] = r.counts
y = r.counts
hdr_string += "Counts "
if have_nine_channels:
data[:,2:chan_out_idx] = r.channel_counts
for i in range(2,chan_out_idx):
hdr_string += "Channeltron_#%d " % (i-1)
if r.extended_channels is not None:
data[:,ex_out_sidx:ex_out_fidx] = r.extended_channels
for i in range(1,ex_out_fidx-ex_out_sidx+1):
hdr_string += "Extended_Channel_#%d " % (i)
# Need to check for divide-by-zero here
try:
data[:,ex_out_fidx] = r.counts / r.extended_channels[:,2]
# Removed the normalization here because we want un-normalized y for the
# pre-edge fitting.
#y = r.counts / r.extended_channels[:,2]
hdr_string += "Counts/I0 "
except FloatingPointError:
print "Divide by zero in I0 channel: skipping I0 normalization."
data[:,ex_out_fidx] = r.counts
#y = r.counts
hdr_string += "Counts "
else:
print "Extended channels are not present: No I0 normalization."
# The order of operations here is:
#
# 1. Calculate the linear pre-edge (gives L).
# 2. Subtract the linear pre-edge (With next step gives y1).
# 3. Divide by an intensity offset to normalize against changes
# in the overall intensity of the background, eg. from analyzer.
# We take this offset to be the value of the linear pre-edge at the
# lowest energy in the region. This could be more accurate but it
# is pretty good. This gives y1.
# (We skip if the background is zero or contains zeros)
# 4. Calculate Shirley background using this new spectrum (gives S).
# 5. Subtract the Shirley background (gives y2).
# 6. Divide by I0 if present. We don't want to take y to be the normalized
# spectrum to start with because the linear pre-edge fit is not robust
# in those circumstances. Note this means we get TWO normalized spectra
# in the file - one column before all the automatic stuff and one at the very
# last column.
# We really want to only do this stuff for XPS scans: if the analyzer mode is
# not in FixedAnalyzerTransmission these extra columns will be zeroed. Strictly
# speaking, we should also allow this for FRR mode (FixedRetardingRatio) but
# since the input routine doesn't account for FRR properly yet there isn't
# much point.
# Have to check for a zero-division here. If we encounter one,
# don't normalize against the pre-edge or divide by I0.
print "Scan mode is: ", r.scan_mode
if r.scan_mode == "FixedAnalyzerTransmission":
print "FixedAnalyzerTransmission mode detected: attempting pre-edge and Shirley subtraction."
try:
L = specs.preedge_calculate(x,y)
y1 = (y - L) / L[x.argmin()]
S = specs.shirley_calculate(x,y1)
y2 = y1 - S
data[:,-5] = L
data[:,-4] = L[x.argmin()]
data[:,-3] = S
data[:,-2] = y2
if r.extended_channels is not None:
data[:,-1] = y2 / r.extended_channels[:,2]
else:
if DEBUG:
print "Warning: No normalization output for this region!"
data[:,-1] = y2
except FloatingPointError:
data[:,-5] = specs.preedge_calculate(x,y)
data[:,-4] = 1.0
data[:,-3] = specs.shirley_calculate(x,y - data[:,-5])
data[:,-2] = y - data[:,-5] - data[:,-3]
data[:,-1] = y - data[:,-5] - data[:,-3]
else:
blank = zeros((r.values_per_curve))
data[:,-5] = blank
data[:,-4] = blank
data[:,-3] = blank
data[:,-2] = blank
data[:,-1] = blank
hdr_string += "Preedge Scale_Division Shirley Counts-Preedge-Shirley Counts-Preedge-Shirley/I0"
# As before, there are possible filename conflicts because there is
# no requirement that SPECS region names be unique.
have_name = False
name_counter = 0
while not have_name:
if name_counter == 0:
r_write_name = r.name + ".xy"
else:
r_write_name = r.name + "-" + str(name_counter) + ".xy"
if not os.path.exists(r_write_name):
have_name = True
name_counter = 0
else:
# Darn, the directory prob already exists.
name_counter += 1
try:
savetxt(r_write_name, data,header=hdr_string)
except TypeError:
# Numpy version not recent enough for the header option.
print ("""Your numpy version (%s) is not recent enough to support the\nheader option in savetxt: upgrade to a version >= 1.7 to get\nthis functionality.""" % (npyversion))
savetxt(r_write_name, data)
os.chdir(unpack_path)
os.chdir(cwd)
print "===================================================================="
print "= ="
print "= Finished processing ="
print "= ="
print "===================================================================="
def loadxml(self):
xmlfile = tkFileDialog.askopenfilename(
title="Choose a SPECS .xml file.", initialdir=os.getcwd())
if xmlfile != None:
xmlcontents = specs.SPECS(xmlfile)
# Make a folder: call it "xmlfile"-unpacked
base_path = os.path.split(xmlfile)[0]
cwd = os.getcwd()
# Remove whitespace either side of the name. It's important. :o)
unpack_path = os.path.join(base_path,
xmlfile + "-unpacked").strip()
os.mkdir(unpack_path)
os.chdir(unpack_path)
for g in xmlcontents.groups:
if DEBUG:
print "=================== OUTPUT: %s ====================" % (
g.name)
# We want to make a folder matching the group name, but we might not
# be able to because it might already exist (group names are not
# required to be unique in SPECS). So we have to fiddle here a bit.
have_name = False
name_counter = 0
while not have_name:
if name_counter == 0:
tmpname = g.name.strip()
else:
tmpname = (g.name + "-" + str(name_counter)).strip()
if not os.path.exists(tmpname):
os.mkdir(tmpname)
os.chdir(tmpname)
have_name = True
name_counter = 0
else:
# Darn, the directory prob already exists.
name_counter += 1
for r in g.regions:
if DEBUG:
print "Constructing output for region %s..." % (r.name)
# Decide how many outputs based on whether we have extended channels
# and so on.
have_nine_channels = False
num_outputs = 2 # Bare output is energy and counts
if r.channel_counts.shape[1] > 1:
have_nine_channels = True
num_outputs += r.channel_counts.shape[1]
chan_out_idx = 2 + r.channel_counts.shape[1]
if r.extended_channels is not None:
# Output all extended channels plus counts / I0.
ex_out_sidx = num_outputs
num_outputs += r.extended_channels.shape[1] + 1
ex_out_fidx = ex_out_sidx + 9
# For now, automatically add the shirley and linear subtractions.
# Eventually this will be an option.
num_outputs += 5 # Linear preedge, background
# scaling factor, shirley, counts - L - S and
# (counts - L - S) / I0.
data = zeros((r.values_per_curve, num_outputs))
hdr_string = ""
if r.scan_mode == "FixedAnalyzerTransmission":
data[:, 0] = r.binding_axis
x = r.binding_axis
hdr_string += "Binding_energy_(eV) "
elif r.scan_mode == "ConstantFinalState":
data[:, 0] = r.excitation_axis
x = r.excitation_axis
hdr_string += "Photon_energy_(eV) "
else:
data[:, 0] = r.kinetic_axis
x = r.kinetic_axis
hdr_string += "Kinetic_energy_(eV) "
data[:, 1] = r.counts
y = r.counts
hdr_string += "Counts "
if have_nine_channels:
data[:, 2:chan_out_idx] = r.channel_counts
for i in range(2, chan_out_idx):
hdr_string += "Channeltron_#%d " % (i - 1)
if r.extended_channels is not None:
data[:, ex_out_sidx:ex_out_fidx] = r.extended_channels
for i in range(1, ex_out_fidx - ex_out_sidx + 1):
hdr_string += "Extended_Channel_#%d " % (i)
# Need to check for divide-by-zero here
try:
data[:,
ex_out_fidx] = r.counts / r.extended_channels[:,
2]
# Removed the normalization here because we want un-normalized y for the
# pre-edge fitting.
#y = r.counts / r.extended_channels[:,2]
hdr_string += "Counts/I0 "
except FloatingPointError:
print "Divide by zero in I0 channel: skipping I0 normalization."
data[:, ex_out_fidx] = r.counts
#y = r.counts
hdr_string += "Counts "
else:
print "Extended channels are not present: No I0 normalization."
# The order of operations here is:
#
# 1. Calculate the linear pre-edge (gives L).
# 2. Subtract the linear pre-edge (With next step gives y1).
# 3. Divide by an intensity offset to normalize against changes
# in the overall intensity of the background, eg. from analyzer.
# We take this offset to be the value of the linear pre-edge at the
# lowest energy in the region. This could be more accurate but it
# is pretty good. This gives y1.
# (We skip if the background is zero or contains zeros)
# 4. Calculate Shirley background using this new spectrum (gives S).
# 5. Subtract the Shirley background (gives y2).
# 6. Divide by I0 if present. We don't want to take y to be the normalized
# spectrum to start with because the linear pre-edge fit is not robust
# in those circumstances. Note this means we get TWO normalized spectra
# in the file - one column before all the automatic stuff and one at the very
# last column.
# We really want to only do this stuff for XPS scans: if the analyzer mode is
# not in FixedAnalyzerTransmission these extra columns will be zeroed. Strictly
# speaking, we should also allow this for FRR mode (FixedRetardingRatio) but
# since the input routine doesn't account for FRR properly yet there isn't
# much point.
# Have to check for a zero-division here. If we encounter one,
# don't normalize against the pre-edge or divide by I0.
print "Scan mode is: ", r.scan_mode
if r.scan_mode == "FixedAnalyzerTransmission":
print "FixedAnalyzerTransmission mode detected: attempting pre-edge and Shirley subtraction."
try:
L = specs.preedge_calculate(x, y)
y1 = (y - L) / L[x.argmin()]
S = specs.shirley_calculate(x, y1)
y2 = y1 - S
data[:, -5] = L
data[:, -4] = L[x.argmin()]
data[:, -3] = S
data[:, -2] = y2
if r.extended_channels is not None:
data[:, -1] = y2 / r.extended_channels[:, 2]
else:
if DEBUG:
print "Warning: No normalization output for this region!"
data[:, -1] = y2
except FloatingPointError:
data[:, -5] = specs.preedge_calculate(x, y)
data[:, -4] = 1.0
data[:, -3] = specs.shirley_calculate(
x, y - data[:, -5])
data[:, -2] = y - data[:, -5] - data[:, -3]
data[:, -1] = y - data[:, -5] - data[:, -3]
else:
blank = zeros((r.values_per_curve))
data[:, -5] = blank
data[:, -4] = blank
data[:, -3] = blank
data[:, -2] = blank
data[:, -1] = blank
hdr_string += "Preedge Scale_Division Shirley Counts-Preedge-Shirley Counts-Preedge-Shirley/I0"
# As before, there are possible filename conflicts because there is
# no requirement that SPECS region names be unique.
have_name = False
name_counter = 0
while not have_name:
if name_counter == 0:
r_write_name = r.name + ".xy"
else:
r_write_name = r.name + "-" + str(
name_counter) + ".xy"
if not os.path.exists(r_write_name):
have_name = True
name_counter = 0
else:
# Darn, the directory prob already exists.
name_counter += 1
try:
savetxt(r_write_name, data, header=hdr_string)
except TypeError:
# Numpy version not recent enough for the header option.
print(
"""Your numpy version (%s) is not recent enough to support the\nheader option in savetxt: upgrade to a version >= 1.7 to get\nthis functionality."""
% (npyversion))
savetxt(r_write_name, data)
os.chdir(unpack_path)
os.chdir(cwd)
print "===================================================================="
print "= ="
print "= Finished processing ="
print "= ="
print "===================================================================="
