Author: bugman
Date: Fri May 3 21:55:32 2013
New Revision: 19660
URL: http://svn.gna.org/viewcvs/relax?rev=19660&view=rev
Log:
Started to redesign the relaxation dispersion target function class.
The input data is now expected to be R2eff/R1rho data and all mentions of
exponential curves have
been eliminated. The func_exp_fit() target function has been deleted as it
is not used - as now
the _minimise_r2eff() method in the dispersion specific analysis class is
used instead. And the
func_fast_2site() target function has been renamed to func_LM63().
Modified:
branches/relax_disp/target_functions/relax_disp.py
Modified: branches/relax_disp/target_functions/relax_disp.py
URL:
http://svn.gna.org/viewcvs/relax/branches/relax_disp/target_functions/relax_disp.py?rev=19660&r1=19659&r2=19660&view=diff
==============================================================================
--- branches/relax_disp/target_functions/relax_disp.py (original)
+++ branches/relax_disp/target_functions/relax_disp.py Fri May 3 21:55:32
2013
@@ -27,14 +27,14 @@
from numpy import dot, float64, zeros
# relax module imports.
-from lib.curve_fit.exponential import exponential_2param_neg
-from lib.dispersion.equations import fast_2site
+from lib.dispersion.equations import r2eff_LM63
from lib.errors import RelaxError
from target_functions.chi2 import chi2
+from specific_analyses.relax_disp.variables import MODEL_CR72, MODEL_LM63,
MODEL_R2EFF
class Dispersion:
- def __init__(self, model=None, num_params=None, num_spins=None,
num_exp_curves=None, num_times=None, values=None, errors=None,
cpmg_frqs=None, spin_lock_nu1=None, relax_times=None, scaling_matrix=None):
+ def __init__(self, model=None, num_params=None, num_spins=None,
num_frq=None, num_disp_points=None, values=None, errors=None, cpmg_frqs=None,
spin_lock_nu1=None, scaling_matrix=None):
"""Relaxation dispersion target functions for optimisation.
Models
@@ -42,9 +42,8 @@
The following models are currently supported:
- - 'exp_fit': Simple fitting of the exponential curves with
parameters {R2eff, I0},
- - 'fast 2-site': The 2-site fast exchange equation with
parameters {R2eff, I0, R2, Rex, kex},
- - 'slow 2-site': The 2-site slow exchange equation with
parameters {R2eff, I0, R2A, kA, dw}.
+ - 'LM63': The Luz and Meiboom (1963) 2-site fast exchange model.
+ - 'CR72': The Carver and Richards (1972) 2-site model for all
time scales.
@keyword model: The relaxation dispersion model to fit.
@@ -53,38 +52,35 @@
@type num_param: int
@keyword num_spins: The number of spins in the cluster.
@type num_spins: int
- @keyword num_exp_curves: The number of exponential curves.
- @type num_exp_curves: int
- @keyword num_times: The number of relaxation times.
- @type num_times: int
- @keyword values: The peak intensities. The first
dimension is that of the spin cluster (each element corresponds to a
different spin in the block), the second dimension is the spectrometer field
strength, the third is the exponential curves, and the fourth are the
relaxation times along the exponential curve.
- @type values: numpy rank-4 float array
- @keyword errors: The peak intensity errors. The four
dimensions must correspond to those of the values argument.
- @type errors: numpy rank-4 float array
- @keyword cpmg_frqs: The CPMG frequencies in Hertz for each
separate exponential curve. This will be ignored for R1rho experiments.
+ @keyword num_frq: The number of spectrometer field
strengths.
+ @type num_frq: int
+ @keyword num_disp_points: The number of points on the dispersion
curve.
+ @type num_disp_points: int
+ @keyword values: The R2eff/R1rho values. The first
dimension is that of the spin cluster (each element corresponds to a
different spin in the block), the second dimension is the spectrometer field
strength, and the third is the dispersion points.
+ @type values: numpy rank-3 float array
+ @keyword errors: The R2eff/R1rho errors. The three
dimensions must correspond to those of the values argument.
+ @type errors: numpy rank-3 float array
+ @keyword cpmg_frqs: The CPMG frequencies in Hertz for each
separate dispersion point. This will be ignored for R1rho experiments.
@type cpmg_frqs: numpy rank-1 float array
- @keyword spin_lock_nu1: The spin-lock field strengths in Hertz
for each separate exponential curve. This will be ignored for CPMG
experiments.
+ @keyword spin_lock_nu1: The spin-lock field strengths in Hertz
for each separate dispersion point. This will be ignored for CPMG
experiments.
@type spin_lock_nu1: numpy rank-1 float array
- @keyword relax_times: The relaxation time points in seconds
for the exponential curve.
- @type relax_times: numpy rank-1 float array
@keyword scaling_matrix: The square and diagonal scaling matrix.
@type scaling_matrix: numpy rank-2 float array
"""
# Check the args.
- if model not in ['exp_fit', 'fast 2-site', 'slow 2-site']:
+ if model not in [MODEL_R2EFF, MODEL_LM63, MODEL_CR72]:
raise RelaxError("The model '%s' is unknown." % model)
# Store the arguments.
self.num_params = num_params
self.num_spins = num_spins
- self.num_exp_curves = num_exp_curves
- self.num_times = num_times
+ self.num_frq = num_frq
+ self.num_disp_points = num_disp_points
self.values = values
self.errors = errors
self.cpmg_frqs = cpmg_frqs
self.spin_lock_nu1 = spin_lock_nu1
- self.relax_times = relax_times
self.scaling_matrix = scaling_matrix
# Scaling initialisation.
@@ -92,18 +88,16 @@
if self.scaling_matrix != None:
self.scaling_flag = True
- # Create the structure for holding the back-calculated peak
intensities (matching the dimensions of the values structure so that external
code can access this data).
- self.back_calc = zeros((num_times, num_exp_curves, num_times),
float64)
+ # Create the structure for holding the back-calculated R2eff values
(matching the dimensions of the values structure).
+ self.back_calc = zeros((num_spins, num_frq, num_disp_points),
float64)
# Set up the model.
- if model == 'exp_fit':
- self.func = self.func_exp_fit
- elif model == 'fast 2-site':
- self.func = self.func_fast_2site
+ if model == MODEL_LM63:
+ self.func = self.func_LM63
- def func_exp_fit(self, params):
- """Target function for the simple exponential curve-fitting.
+ def func_LM63(self, params):
+ """Target function for the Luz and Meiboom (1963) fast 2-site
exchange model.
@param params: The vector of parameter values.
@type params: numpy rank-1 float array
@@ -115,41 +109,18 @@
if self.scaling_flag:
params = dot(params, self.scaling_matrix)
+ # Initialise.
+ chi2_sum = 0.0
+
# Loop over the spins.
- chi2_sum = 0.0
for spin_index in range(self.num_spins):
- # Loop over the exponential curves.
- for exp_index in range(self.num_exp_curves):
- # Unpack the exponential curve parameters.
- index = spin_index * 2 * self.num_exp_curves + exp_index *
self.num_exp_curves
- r2eff = params[index]
- i0 = params[index + 1]
+ # Loop over the spectrometer frequencies.
+ for frq_index in range(self.num_frq):
+ # Back calculate the R2eff values.
+ r2eff_LM63(params=params, cpmg_frqs=self.cpmg_frqs,
back_calc=self.back_calc[spin_index, frq_index],
num_disp_points=self.num_disp_points)
- # Back-calculate the points on the exponential curve.
- exponential_2param_neg(rate=r2eff, i0=i0,
x=self.relax_times, y=self.back_calc[spin_index, field_index, exp_index])
+ # Calculate and return the chi-squared value.
+ chi2_sum += chi2(values[spin_index, frq_index],
back_calc[spin_index, frq_index], sd[spin_index, frq_index])
- # Calculate the chi-squared value for this curve.
- chi2_sum += chi2(self.values[spin_index, field_index,
exp_index], self.back_calc[spin_index, field_index, exp_index],
self.errors[spin_index, field_index, exp_index])
-
- # Return the chi-squared value.
+ # Return the total chi-squared value.
return chi2_sum
-
-
- def func_fast_2site(self, params):
- """Target function for the fast 2-site exchange model.
-
- @param params: The vector of parameter values.
- @type params: numpy rank-1 float array
- @return: The chi-squared value.
- @rtype: float
- """
-
- # Scaling.
- if self.scaling_flag:
- params = dot(params, self.scaling_matrix)
-
- # Back calculated the effective transversal relaxation rates.
- fast_2site(params=params, cpmg_frqs=self.cpmg_frqs,
back_calc=self.back_calc, num_times=self.num_times)
-
- # Calculate and return the chi-squared value.
- return chi2(values, back_calc, sd)
r19660 - /branches/relax_disp/target_functions/relax_disp.py