def fit_curve(ps_list):
#--input:ps_list, (x,y)
x = np.array([r[0] for r in ps_list])
y = np.array([r[1] for r in ps_list])
mod = GaussianModel()
pars = mod.guess(y, x=x)
bestresult = mod.fit(y, pars, x=x)
return (bestresult.best_fit,x,y,getfloat_attr(bestresult, 'chisqr'),bestresult)
def test_getfloat_attr(params, par, attr, expected):
"""Test getfloat_attr function."""
if par == 'aa1':
# add an attribute that is not None, float, or int
# This will never occur for Parameter values, but this function is
# also used on the MinimizerResult/ModelResult where it could happen.
params['aa1'].test = 20 + 5j
output = getfloat_attr(params[par], attr)
assert output == expected
if par == 'a':
assert len(output) == 10 # leading blank for pos values is stripped
elif par == 'b':
assert len(output) == 11
elif par == 'c1':
assert len(output) == 1
def export_modelresult(result, filename='fitresult.xdi',
datafile=None, ydata=None, yerr=None,
_larch=None, **kwargs):
"""
export an lmfit ModelResult to an XDI data file
Arguments
---------
result ModelResult, required
filename name of output file ['fitresult.xdi']
datafile name of data file [`None`]
ydata data array used for fit [`None`]
yerr data error array used for fit [`None`]
Notes
-----
keyword arguments should include independent variables
Example
-------
result = model.fit(ydata, params, x=x)
export_modelresult(result, 'fitresult_1.xdi', x=x,
datafile='XYData.txt')
"""
if not isinstance(result, ModelResult):
raise ValueError("export_fit needs a lmfit ModelReult")
header = ["XDI/1.1 Lmfit Result File"]
hadd = header.append
if datafile is not None:
hadd(" Datafile.name: %s " % datafile)
else:
hadd(" Datafile.name: <unknnown>")
ndata = len(result.best_fit)
columns = OrderedDict()
for aname in result.model.independent_vars:
val = kwargs.get(aname, None)
if val is not None and len(val) == ndata:
columns[aname] = val
if ydata is not None:
columns['ydata'] = ydata
if yerr is not None:
columns['yerr'] = yerr
columns['best_fit'] = result.best_fit
columns['init_fit'] = result.init_fit
delta_fit = result.eval_uncertainty(result.params, **kwargs)
columns['delta_fit'] = delta_fit
if len(result.model.components) > 1:
comps = result.eval_components(result.params, **kwargs)
for name, val in comps.items():
columns[name] = val
clabel = []
for i, cname in enumerate(columns):
hadd(" Column.%i: %s" % (i+1, cname))
clabel.append('%15s ' % cname)
hadd("Fit.Statistics: Start here")
hadd(" Fit.model_name: %s" % result.model.name)
hadd(" Fit.method: %s" % result.method)
hadd(" Fit.n_function_evals: %s" % getfloat_attr(result, 'nfev'))
hadd(" Fit.n_data_points: %s" % getfloat_attr(result, 'ndata'))
hadd(" Fit.n_variables: %s" % getfloat_attr(result, 'nvarys'))
hadd(" Fit.chi_square: %s" % getfloat_attr(result, 'chisqr', length=11))
hadd(" Fit.reduced_chi_square: %s" % getfloat_attr(result, 'redchi', length=11))
hadd(" Fit.akaike_info_crit: %s" % getfloat_attr(result, 'aic', length=11))
hadd(" Fit.bayesian_info_crit: %s" % getfloat_attr(result, 'bic', length=11))
hadd("Param.Statistics: Start here")
namelen = max([len(p) for p in result.params])
for name, par in result.params.items():
space = ' '*(namelen+1-len(name))
nout = "Param.%s:%s" % (name, space)
inval = '(init= ?)'
if par.init_value is not None:
inval = '(init=% .7g)' % par.init_value
try:
sval = gformat(par.value)
except (TypeError, ValueError):
sval = 'Non Numeric Value?'
if par.stderr is not None:
serr = gformat(par.stderr, length=9)
sval = '%s +/-%s' % (sval, serr)
if par.vary:
bounds = "[%s: %s]" % (gformat(par.min), gformat(par.max))
hadd(" %s %s %s %s" % (nout, sval, bounds, inval))
elif par.expr is not None:
hadd(" %s %s == '%s'" % (nout, sval, par.expr))
else:
hadd(" %s % .7g (fixed)" % (nout, par.value))
hadd("//////// Fit Report ////////")
for r in result.fit_report().split('\n'):
hadd(" %s" % r)
hadd("-" * 77)
hadd("".join(clabel)[1:])
header[0] = "XDI/1.1 Lmfit Result File %i header lines" % (len(header))
dtable = []
for key, dat in columns.items():
dtable.append(dat)
dtable = np.array(dtable).transpose()
datatable = []
for i in range(ndata):
col = dtable[i, :]*1.0
row = []
for cval in col:
try:
val = gformat(cval, length=15)
except:
val = repr(cval)
row.append(val)
datatable.append(" ".join(row))
datatable.append('')
with open(filename, 'w') as fh:
fh.write("\n".join(['#%s' % s for s in header]))
fh.write("\n")
fh.write("\n".join(datatable))
def export_modelresult(result, filename='fitresult.xdi',
datafile=None, ydata=None, yerr=None,
_larch=None, **kwargs):
"""
export an lmfit ModelResult to an XDI data file
Arguments
---------
result ModelResult, required
filename name of output file ['fitresult.xdi']
datafile name of data file [`None`]
ydata data array used for fit [`None`]
yerr data error array used for fit [`None`]
Notes
-----
keyword arguments should include independent variables
Example
-------
result = model.fit(ydata, params, x=x)
export_modelresult(result, 'fitresult_1.xdi', x=x,
datafile='XYData.txt')
"""
if not isinstance(result, ModelResult):
raise ValueError("export_fit needs a lmfit ModelReult")
header = ["XDI/1.1 Lmfit Result File"]
hadd = header.append
if datafile is not None:
hadd(" Datafile.name: %s " % datafile)
else:
hadd(" Datafile.name: <unknnown>")
ndata = len(result.best_fit)
columns = OrderedDict()
for aname in result.model.independent_vars:
val = kwargs.get(aname, None)
if val is not None and len(val) == ndata:
columns[aname] = val
if ydata is not None:
columns['ydata'] = ydata
if yerr is not None:
columns['yerr'] = yerr
columns['best_fit'] = result.best_fit
columns['init_fit'] = result.init_fit
delta_fit = 0.0*result.best_fit
if not any([p.stderr is None for p in result.params.values()]):
delta_fit = result.eval_uncertainty(result.params, **kwargs)
columns['delta_fit'] = delta_fit
if len(result.model.components) > 1:
comps = result.eval_components(result.params, **kwargs)
for name, val in comps.items():
columns[name] = val
clabel = []
for i, cname in enumerate(columns):
hadd(" Column.%i: %s" % (i+1, cname))
clabel.append('%15s ' % cname)
hadd("Fit.Statistics: Start here")
hadd(" Fit.model_name: %s" % result.model.name)
hadd(" Fit.method: %s" % result.method)
hadd(" Fit.n_function_evals: %s" % getfloat_attr(result, 'nfev'))
hadd(" Fit.n_data_points: %s" % getfloat_attr(result, 'ndata'))
hadd(" Fit.n_variables: %s" % getfloat_attr(result, 'nvarys'))
hadd(" Fit.chi_square: %s" % getfloat_attr(result, 'chisqr', length=11))
hadd(" Fit.reduced_chi_square: %s" % getfloat_attr(result, 'redchi', length=11))
hadd(" Fit.akaike_info_crit: %s" % getfloat_attr(result, 'aic', length=11))
hadd(" Fit.bayesian_info_crit: %s" % getfloat_attr(result, 'bic', length=11))
hadd("Param.Statistics: Start here")
namelen = max([len(p) for p in result.params])
for name, par in result.params.items():
space = ' '*(namelen+1-len(name))
nout = "Param.%s:%s" % (name, space)
inval = '(init= ?)'
if par.init_value is not None:
inval = '(init=% .7g)' % par.init_value
try:
sval = gformat(par.value)
except (TypeError, ValueError):
sval = 'Non Numeric Value?'
if par.stderr is not None:
serr = gformat(par.stderr, length=9)
sval = '%s +/-%s' % (sval, serr)
if par.vary:
bounds = "[%s: %s]" % (gformat(par.min), gformat(par.max))
hadd(" %s %s %s %s" % (nout, sval, bounds, inval))
elif par.expr is not None:
hadd(" %s %s == '%s'" % (nout, sval, par.expr))
else:
hadd(" %s % .7g (fixed)" % (nout, par.value))
hadd("//////// Fit Report ////////")
for r in result.fit_report().split('\n'):
hadd(" %s" % r)
hadd("-" * 77)
hadd("".join(clabel)[1:])
header[0] = "XDI/1.1 Lmfit Result File %i header lines" % (len(header))
dtable = []
for key, dat in columns.items():
dtable.append(dat)
dtable = np.array(dtable).transpose()
datatable = []
for i in range(ndata):
col = dtable[i, :]*1.0
row = []
for cval in col:
try:
val = gformat(cval, length=15)
except:
val = repr(cval)
row.append(val)
datatable.append(" ".join(row))
datatable.append('')
with open(filename, 'w') as fh:
fh.write("\n".join(['#%s' % s for s in header]))
fh.write("\n")
fh.write("\n".join(datatable))
