Allright. That is changed. But, does the grid_search not call the target function? Best Troels 2014-06-06 15:36 GMT+02:00 Edward d'Auvergne <edward@xxxxxxxxxxxxx>:
Ok, I just checked for the problem and it is the parameter values of: R20 = [10.1, 10.2, 10.3, 100.1, 100.2, 100.3, 20.1, 20.2, 20.3, 200.1, 200.2, 200.3, 30.1, 30.2, 30.3, 300.1, 300.2, 300.3, 40.1, 40.2, 40.3, 400.1, 400.2, 400.3] If you look at the specific_analyses.relax_disp.parameters.linear_constraints() function, you will see that the R20 constraints are: 0 <= R2 <= 200 0 <= R2A <= 200 0 <= R2B <= 200 You have quite a few R20 values above 200, and that is the constraint violation. When there is a violation, the point is skipped and, as the grid search only has one point, then the grid search will do nothing. The value of 200 is reasonable as with such a relaxation rate, you will no longer observe any peaks in your spectra. That is the reason for this value, and it prevents optimisation from sending the R20 values to unreasonably high values. Maybe you should divide all your R20 values by 10? Regards, Edward On 6 June 2014 15:28, Troels Emtekær Linnet <tlinnet@xxxxxxxxxxxxx> wrote:Hi Ed. All r2, r2a and r2b values should be between 5 - 20. Thanks. But, how does the grid_search skip the target function? Best Troels 2014-06-06 15:25 GMT+02:00 Edward d'Auvergne <edward@xxxxxxxxxxxxx>:Note that the indices of ci match the b vector, so you can look at the assemble_linear_constraints() function and work out which constraint is being violated. Regards, Edward On 6 June 2014 15:23, Edward d'Auvergne <edward@xxxxxxxxxxxxx> wrote:Here, all negative ci values indicate a constraint violation, as: ci = A.x - b, where A and b are the constraint matrices and x is the parameter vector. The linear constraints used by minfx are defined as: A.x - b >= 0. So you will need to determine which parameters are violating the constraints and change the input. Regards, Edward On 6 June 2014 15:20, Troels Emtekær Linnet <tlinnet@xxxxxxxxxxxxx> wrote:Ah "Constraint violated, skipping grid point." Fitting to the spin block [':1@N', ':2@N', ':3@N', ':4@N'] ---------------------------------------------------------- Unconstrained grid search size: 1 (constraints may decrease thissize).Grid search ~~~~~~~~~~~ Linear constraint matrices. A: array([[ 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,0.,0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [-1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., -1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., -1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., -1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., -1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., -1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., -1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., -1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., -1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., -1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., -1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., -1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., -1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., -1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., -1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., -1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., -1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., -1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., -1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., -1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., -1., 0., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., -1., 0., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., -1., 0., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., -1., 0., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., -1., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1.], [ 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., -1.]]) b: array([ 0. , -20. , 0. , -20. , 0. , -20. , 0. ,-20., 0. , -20. , 0. , -20. , 0. , -20. , 0. , -20.,0. , -20. , 0. , -20. , 0. , -20. , 0. , -20.,0. , -20. , 0. , -20. , 0. , -20. , 0. , -20.,0. , -20. , 0. , -20. , 0. , -20. , 0. , -20.,0. , -20. , 0. , -20. , 0. , -20. , 0. , -20.,0. , 0. , 0. , 0. , -1. , 0.5, 0. , -200.])Searching through 1 grid nodes. k: 0 xk: [ 1.01, 1.02, 1.03, 10.01, 10.02, 10.03, 2.01, 2.02, 2.03, 20.01, 20.02, 20.03, 3.01, 3.02, 3.03, 30.01, 30.02, 30.03, 4.01, 4.02, 4.03, 40.01, 40.02, 40.03, 1, 2, 3, 4, 0.9, 0.1] Constraint violated, skipping grid point. ci: array([ 1.010000000000000e+00, 1.899000000000000e+01, 1.020000000000000e+00, 1.898000000000000e+01, 1.030000000000000e+00, 1.897000000000000e+01, 1.001000000000000e+01, 9.990000000000000e+00, 1.002000000000000e+01, 9.980000000000000e+00, 1.003000000000000e+01, 9.970000000000001e+00, 2.010000000000000e+00, 1.799000000000000e+01, 2.020000000000000e+00, 1.798000000000000e+01, 2.030000000000000e+00, 1.797000000000000e+01, 2.001000000000000e+01, -9.999999999998010e-03, 2.002000000000000e+01, -1.999999999999957e-02, 2.003000000000000e+01, -3.000000000000114e-02, 3.010000000000000e+00, 1.699000000000000e+01, 3.020000000000000e+00, 1.698000000000000e+01, 3.030000000000000e+00, 1.697000000000000e+01, 3.001000000000000e+01, -1.001000000000000e+01, 3.002000000000000e+01, -1.002000000000000e+01, 3.003000000000000e+01, -1.003000000000000e+01, 4.010000000000000e+00, 1.599000000000000e+01, 4.020000000000000e+00, 1.598000000000000e+01, 4.029999999999999e+00, 1.597000000000000e+01, 4.001000000000001e+01, -2.001000000000001e+01, 4.002000000000000e+01, -2.002000000000000e+01, 4.003000000000000e+01, -2.003000000000000e+01, 1.000000000000000e+00, 2.000000000000000e+00, 3.000000000000000e+00, 4.000000000000000e+00, 9.999999999999998e-02, 4.000000000000000e-01, 1.000000000000000e-01, 1.999000000000000e+02]) Optimised parameter values: r2a (SQ CPMG - 500.00000000 MHz) 10.100000000000000 r2a (SQ CPMG - 600.00000000 MHz) 10.199999999999999 r2a (SQ CPMG - 700.00000000 MHz) 10.300000000000001 r2b (SQ CPMG - 500.00000000 MHz) 100.099999999999994 r2b (SQ CPMG - 600.00000000 MHz) 100.199999999999989 r2b (SQ CPMG - 700.00000000 MHz) 100.299999999999997 r2a (SQ CPMG - 500.00000000 MHz) 20.100000000000001 r2a (SQ CPMG - 600.00000000 MHz) 20.199999999999999 r2a (SQ CPMG - 700.00000000 MHz) 20.300000000000004 r2b (SQ CPMG - 500.00000000 MHz) 200.099999999999966 r2b (SQ CPMG - 600.00000000 MHz) 200.199999999999989 r2b (SQ CPMG - 700.00000000 MHz) 200.300000000000011 r2a (SQ CPMG - 500.00000000 MHz) 30.100000000000001 r2a (SQ CPMG - 600.00000000 MHz) 30.199999999999999 r2a (SQ CPMG - 700.00000000 MHz) 30.300000000000004 r2b (SQ CPMG - 500.00000000 MHz) 300.100000000000023 r2b (SQ CPMG - 600.00000000 MHz) 300.199999999999989 r2b (SQ CPMG - 700.00000000 MHz) 300.300000000000011 r2a (SQ CPMG - 500.00000000 MHz) 40.099999999999994 r2a (SQ CPMG - 600.00000000 MHz) 40.200000000000003 r2a (SQ CPMG - 700.00000000 MHz) 40.299999999999997 r2b (SQ CPMG - 500.00000000 MHz) 400.100000000000023 r2b (SQ CPMG - 600.00000000 MHz) 400.199999999999932 r2b (SQ CPMG - 700.00000000 MHz) 400.300000000000011 dw 1.000000000000000 dw 2.000000000000000 dw 3.000000000000000 dw 4.000000000000000 pA 0.900000000000000 kex 1000.000000000000000 ('CR72 full', 'Ala', ':1@N', 'r2a', 'SQ CPMG - 600.00000000 MHz',10.2,10.2) ('CR72 full', 'Ala', ':1@N', 'r2a', 'SQ CPMG - 500.00000000 MHz',10.1,10.1) ('CR72 full', 'Ala', ':1@N', 'r2a', 'SQ CPMG - 700.00000000 MHz',10.3,10.3) ('CR72 full', 'Ala', ':1@N', 'r2b', 'SQ CPMG - 600.00000000 MHz', 100.19999999999999, 100.2) ('CR72 full', 'Ala', ':1@N', 'r2b', 'SQ CPMG - 500.00000000 MHz', 100.1, 100.1) ('CR72 full', 'Ala', ':1@N', 'r2b', 'SQ CPMG - 700.00000000 MHz', 100.3, 100.3) ('CR72 full', 'Ala', ':2@N', 'r2a', 'SQ CPMG - 600.00000000 MHz',20.2,20.2) ('CR72 full', 'Ala', ':2@N', 'r2a', 'SQ CPMG - 500.00000000 MHz',20.1,20.1) ('CR72 full', 'Ala', ':2@N', 'r2a', 'SQ CPMG - 700.00000000 MHz', 20.300000000000004, 20.3) ('CR72 full', 'Ala', ':2@N', 'r2b', 'SQ CPMG - 600.00000000 MHz', 200.2, 200.2) ('CR72 full', 'Ala', ':2@N', 'r2b', 'SQ CPMG - 500.00000000 MHz', 200.09999999999997, 200.1) ('CR72 full', 'Ala', ':2@N', 'r2b', 'SQ CPMG - 700.00000000 MHz', 200.3, 200.3) ('CR72 full', 'Ala', ':3@N', 'r2a', 'SQ CPMG - 600.00000000 MHz',30.2,30.2) ('CR72 full', 'Ala', ':3@N', 'r2a', 'SQ CPMG - 500.00000000 MHz',30.1,30.1) ('CR72 full', 'Ala', ':3@N', 'r2a', 'SQ CPMG - 700.00000000 MHz', 30.300000000000004, 30.3) ('CR72 full', 'Ala', ':3@N', 'r2b', 'SQ CPMG - 600.00000000 MHz', 300.2, 300.2) ('CR72 full', 'Ala', ':3@N', 'r2b', 'SQ CPMG - 500.00000000 MHz', 300.1, 300.1) ('CR72 full', 'Ala', ':3@N', 'r2b', 'SQ CPMG - 700.00000000 MHz', 300.3, 300.3) ('CR72 full', 'Ala', ':4@N', 'r2a', 'SQ CPMG - 600.00000000 MHz',40.2,40.2) ('CR72 full', 'Ala', ':4@N', 'r2a', 'SQ CPMG - 500.00000000 MHz', 40.099999999999994, 40.1) ('CR72 full', 'Ala', ':4@N', 'r2a', 'SQ CPMG - 700.00000000 MHz',40.3,40.3) ('CR72 full', 'Ala', ':4@N', 'r2b', 'SQ CPMG - 600.00000000 MHz', 400.19999999999993, 400.2) ('CR72 full', 'Ala', ':4@N', 'r2b', 'SQ CPMG - 500.00000000 MHz', 400.1, 400.1) ('CR72 full', 'Ala', ':4@N', 'r2b', 'SQ CPMG - 700.00000000 MHz', 400.3, 400.3) 2014-06-06 15:14 GMT+02:00 Edward d'Auvergne <edward@xxxxxxxxxxxxx>:Try setting "verbosity=10" in the grid_search user function call. This might reveal the problem. Regards, Edward On 6 June 2014 15:12, Troels Emtekær Linnet <tlinnet@xxxxxxxxxxxxx> wrote:I put in a print command in the target function to see R20A and R20B. This is not showing when doing grid search, but only when doing minimisation? Best Troels 2014-06-06 15:10 GMT+02:00 Troels Emtekær Linnet <tlinnet@xxxxxxxxxxxxx>:I tried: ## Experiments # Exp 1 sfrq_1 = 500.0*1E6 r20_key_1 = generate_r20_key(exp_type=EXP_TYPE_CPMG_SQ, frq=sfrq_1) time_T2_1 = 0.05 ncycs_1 = range(2,22,2) # Here you define the direct R2eff errors (rad/s), asbeingadded or subtracted for the created R2eff point in the correspondingncyccpmg frequence. #r2eff_errs_1 = [0.05, -0.05, 0.05, -0.05, 0.05, -0.05, 0.05, -0.05, 0.05, -0.05, 0.05, -0.05, 0.05, -0.05, 0.05] r2eff_errs_1 = [0.0] * len(ncycs_1) exp_1 = [sfrq_1, time_T2_1, ncycs_1, r2eff_errs_1] sfrq_2 = 600.0*1E6 r20_key_2 = generate_r20_key(exp_type=EXP_TYPE_CPMG_SQ, frq=sfrq_2) time_T2_2 = 0.06 ncycs_2 = range(2,22,2) # Here you define the direct R2eff errors (rad/s), asbeingadded or subtracted for the created R2eff point in the correspondingncyccpmg frequence. #r2eff_errs_2 = [0.05, -0.05, 0.05, -0.05, 0.05, -0.05, 0.05, -0.05, 0.05, -0.05, 0.05, -0.05, 0.05, -0.05, 0.05, -0.05, 0.05] r2eff_errs_2 = [0.0] * len(ncycs_2) exp_2 = [sfrq_2, time_T2_2, ncycs_2, r2eff_errs_2] sfrq_3 = 700.0*1E6 r20_key_3 = generate_r20_key(exp_type=EXP_TYPE_CPMG_SQ, frq=sfrq_3) time_T2_3 = 0.07 ncycs_3 = range(2,22,2) # Here you define the direct R2eff errors (rad/s), asbeingadded or subtracted for the created R2eff point in the correspondingncyccpmg frequence. #r2eff_errs_2 = [0.05, -0.05, 0.05, -0.05, 0.05, -0.05, 0.05, -0.05, 0.05, -0.05, 0.05, -0.05, 0.05, -0.05, 0.05, -0.05, 0.05] r2eff_errs_3 = [0.0] * len(ncycs_3) exp_3 = [sfrq_3, time_T2_3, ncycs_3, r2eff_errs_3] # Collect all exps exps = [exp_1, exp_2, exp_3] R20 = [10.1, 10.2, 10.3, 100.1, 100.2, 100.3, 20.1, 20.2, 20.3, 200.1, 200.2, 200.3, 30.1, 30.2, 30.3, 300.1, 300.2, 300.3, 40.1, 40.2, 40.3, 400.1, 400.2, 400.3] #R20 = [10.1, 10.2, 10.3, 100.1, 100.2, 100.3, 20.1,20.2,20.3, 200.1, 200.2, 200.3] dw_arr = [1.0, 2.0, 3.0, 4.0] #dw_arr = [1.0, 2.0] pA_arr = [0.9] kex_arr = [1000.] spins = [ ['Ala', 1, 'N', {'r2a': {r20_key_1: R20[0], r20_key_2: R20[1], r20_key_3: R20[2]}, 'r2b': {r20_key_1: R20[3], r20_key_2: R20[4], r20_key_3: R20[5]}, 'kex': kex_arr[0], 'pA': pA_arr[0], 'dw': dw_arr[0]}], ['Ala', 2, 'N', {'r2a': {r20_key_1: R20[6], r20_key_2: R20[7], r20_key_3: R20[8]}, 'r2b': {r20_key_1: R20[9], r20_key_2: R20[10], r20_key_3: R20[11]}, 'kex': kex_arr[0], 'pA': pA_arr[0], 'dw': dw_arr[1]}], ['Ala', 3, 'N', {'r2a': {r20_key_1: R20[12], r20_key_2: R20[13], r20_key_3: R20[14]}, 'r2b': {r20_key_1: R20[15], r20_key_2: R20[16], r20_key_3: R20[17]}, 'kex': kex_arr[0], 'pA': pA_arr[0], 'dw': dw_arr[2]}], ['Ala', 4, 'N', {'r2a': {r20_key_1: R20[18], r20_key_2: R20[19], r20_key_3: R20[20]}, 'r2b': {r20_key_1: R20[21], r20_key_2: R20[22], r20_key_3: R20[23]}, 'kex': kex_arr[0], 'pA': pA_arr[0], 'dw': dw_arr[3]}], ] ------ ------------ relax> grid_search(lower=[10.1, 10.2, 10.3, 100.1, 100.2, 100.3, 20.1, 20.2, 20.3, 200.1, 200.2, 200.3, 30.1, 30.2, 30.3, 300.1, 300.2, 300.3, 40.1, 40.2, 40.3, 400.1, 400.2, 400.3, 1.0, 2.0, 3.0, 4.0, 0.9, 1000.0], upper=[10.1, 10.2, 10.3, 100.1, 100.2, 100.3, 20.1, 20.2, 20.3, 200.1, 200.2, 200.3, 30.1, 30.2, 30.3, 300.1, 300.2, 300.3, 40.1, 40.2, 40.3, 400.1, 400.2, 400.3, 1.0, 2.0, 3.0, 4.0, 0.9, 1000.0], inc=1, constraints=True, verbosity=1) Fitting to the spin block [':1@N', ':2@N', ':3@N', ':4@N'] ---------------------------------------------------------- Unconstrained grid search size: 1 (constraints may decrease this size). Grid search ~~~~~~~~~~~ Searching through 1 grid nodes. Optimised parameter values: r2a (SQ CPMG - 500.00000000 MHz) 10.100000000000000 r2a (SQ CPMG - 600.00000000 MHz) 10.199999999999999 r2a (SQ CPMG - 700.00000000 MHz) 10.300000000000001 r2b (SQ CPMG - 500.00000000 MHz) 100.099999999999994 r2b (SQ CPMG - 600.00000000 MHz) 100.199999999999989 r2b (SQ CPMG - 700.00000000 MHz) 100.299999999999997 r2a (SQ CPMG - 500.00000000 MHz) 20.100000000000001 r2a (SQ CPMG - 600.00000000 MHz) 20.199999999999999 r2a (SQ CPMG - 700.00000000 MHz) 20.300000000000004 r2b (SQ CPMG - 500.00000000 MHz) 200.099999999999966 r2b (SQ CPMG - 600.00000000 MHz) 200.199999999999989 r2b (SQ CPMG - 700.00000000 MHz) 200.300000000000011 r2a (SQ CPMG - 500.00000000 MHz) 30.100000000000001 r2a (SQ CPMG - 600.00000000 MHz) 30.199999999999999 r2a (SQ CPMG - 700.00000000 MHz) 30.300000000000004 r2b (SQ CPMG - 500.00000000 MHz) 300.100000000000023 r2b (SQ CPMG - 600.00000000 MHz) 300.199999999999989 r2b (SQ CPMG - 700.00000000 MHz) 300.300000000000011 r2a (SQ CPMG - 500.00000000 MHz) 40.099999999999994 r2a (SQ CPMG - 600.00000000 MHz) 40.200000000000003 r2a (SQ CPMG - 700.00000000 MHz) 40.299999999999997 r2b (SQ CPMG - 500.00000000 MHz) 400.100000000000023 r2b (SQ CPMG - 600.00000000 MHz) 400.199999999999932 r2b (SQ CPMG - 700.00000000 MHz) 400.300000000000011 dw 1.000000000000000 dw 2.000000000000000 dw 3.000000000000000 dw 4.000000000000000 pA 0.900000000000000 kex 1000.000000000000000 ('CR72 full', 'Ala', ':1@N', 'r2a', 'SQ CPMG - 600.00000000MHz',10.2, 10.2) ('CR72 full', 'Ala', ':1@N', 'r2a', 'SQ CPMG - 500.00000000MHz',10.1, 10.1) ('CR72 full', 'Ala', ':1@N', 'r2a', 'SQ CPMG - 700.00000000MHz',10.3, 10.3) ('CR72 full', 'Ala', ':1@N', 'r2b', 'SQ CPMG - 600.00000000MHz',100.19999999999999, 100.2) ('CR72 full', 'Ala', ':1@N', 'r2b', 'SQ CPMG - 500.00000000MHz',100.1, 100.1) ('CR72 full', 'Ala', ':1@N', 'r2b', 'SQ CPMG - 700.00000000MHz',100.3, 100.3) ('CR72 full', 'Ala', ':2@N', 'r2a', 'SQ CPMG - 600.00000000MHz',20.2, 20.2) ('CR72 full', 'Ala', ':2@N', 'r2a', 'SQ CPMG - 500.00000000MHz',20.1, 20.1) ('CR72 full', 'Ala', ':2@N', 'r2a', 'SQ CPMG - 700.00000000MHz',20.300000000000004, 20.3) ('CR72 full', 'Ala', ':2@N', 'r2b', 'SQ CPMG - 600.00000000MHz',200.2, 200.2) ('CR72 full', 'Ala', ':2@N', 'r2b', 'SQ CPMG - 500.00000000MHz',200.09999999999997, 200.1) ('CR72 full', 'Ala', ':2@N', 'r2b', 'SQ CPMG - 700.00000000MHz',200.3, 200.3) ('CR72 full', 'Ala', ':3@N', 'r2a', 'SQ CPMG - 600.00000000MHz',30.2, 30.2) ('CR72 full', 'Ala', ':3@N', 'r2a', 'SQ CPMG - 500.00000000MHz',30.1, 30.1) ('CR72 full', 'Ala', ':3@N', 'r2a', 'SQ CPMG - 700.00000000MHz',30.300000000000004, 30.3) ('CR72 full', 'Ala', ':3@N', 'r2b', 'SQ CPMG - 600.00000000MHz',300.2, 300.2) ('CR72 full', 'Ala', ':3@N', 'r2b', 'SQ CPMG - 500.00000000MHz',300.1, 300.1) ('CR72 full', 'Ala', ':3@N', 'r2b', 'SQ CPMG - 700.00000000MHz',300.3, 300.3) ('CR72 full', 'Ala', ':4@N', 'r2a', 'SQ CPMG - 600.00000000MHz',40.2, 40.2) ('CR72 full', 'Ala', ':4@N', 'r2a', 'SQ CPMG - 500.00000000MHz',40.099999999999994, 40.1) ('CR72 full', 'Ala', ':4@N', 'r2a', 'SQ CPMG - 700.00000000MHz',40.3, 40.3) ('CR72 full', 'Ala', ':4@N', 'r2b', 'SQ CPMG - 600.00000000MHz',400.19999999999993, 400.2) ('CR72 full', 'Ala', ':4@N', 'r2b', 'SQ CPMG - 500.00000000MHz',400.1, 400.1) ('CR72 full', 'Ala', ':4@N', 'r2b', 'SQ CPMG - 700.00000000MHz',400.3, 400.3) -------- It jumps over the target function? This is a little weird? Best Troels 2014-06-06 15:02 GMT+02:00 Edward d'Auvergne <edward@xxxxxxxxxxxxx>:Hi, Some more information is needed, as it's not possible to tell where this stopped. A good idea would be to turn up the verbositylevelto see what minfx is doing. Did you call the grid_search user function? Or the minimise user function? Did the grid search saysomethinglike "Searching through 1 grid nodes"? Regards, Edward On 6 June 2014 14:48, Troels Emtekær Linnet <tlinnet@xxxxxxxxxxxxx> wrote:Hi Edward. When I try to do a grid search, it does not initalize func_CR72_full in the target function? If I make import sys sys.exit() It does not stop? Only if I do a minimise, it stops? It is in specific_analyses/relax_disp/optimisation.py line 745 Inserting print model.func.im_func.__name__ gives func_CR72_full How does it know how to unpack and calculate? Best Troels 2014-06-06 12:32 GMT+02:00 Edward d'Auvergne <edward@xxxxxxxxxxxxx>:Hi, That sounds good. Maybe it's best to have the number offieldsand number of spins set to something different and not to 2?Thatway the unpacking is stressed as much as possible and there cannotbe aaccidental swap of the field and spin dimensions being unnoticed by the test. This is not likely, but I've encountered enough weird and supposedly impossible situations in the development of relax that it would not surprise me. Cheers, Edward On 6 June 2014 12:27, Troels Emtekær Linnet <tlinnet@xxxxxxxxxxxxx> wrote:Check. I am generating R2eff data for 3 fields, and 3 spins, for full model. I will put the data intest_suite/shared_data/dispersion/bug_22146_unpacking_r2a_r2b_clusterBest Troels 2014-06-06 12:11 GMT+02:00 Edward d'Auvergne <edward@xxxxxxxxxxxxx>:Hi, Right, you have the 2 parameters in the self.num_spins*2 part. And I forgot about the parameters being different for eachfield.It would be good to then have a multi-field and multi-spin cluster system test to really make sure that relax operates correctly, especially with the data going into the target function and the subsequently unpacking the results into the relax data store. For example someone might modify the loop_parameters() function - this concept could be migrated into the specific API and converted into a common mechanism for all the analysis types as it is quite powerful - and they may not know that the change they just made broke code in the target_functions.relax_disp module. Cheers, Edward On 6 June 2014 12:05, Troels Emtekær Linnet <tlinnet@xxxxxxxxxxxxx> wrote:Hi Ed. The implementations needs: R20 = params[:self.end_index[1]].reshape(self.num_spins*2, self.num_frq) R20A = R20[::2].flatten() R20B = R20[1::2].flatten() 2014-06-06 11:55 GMT+02:00 Edward d'Auvergne <edward@xxxxxxxxxxxxx>:The different unpacking implementations can be tested with the timeit Python module to see which is fastest (http://thread.gmane.org/gmane.science.nmr.relax.devel/5937/focus=6010).Cheers, Edward On 6 June 2014 11:53, Edward d'Auvergne <edward@xxxxxxxxxxxxx> wrote:Hi, In this case, I think 'num_frq' should be fixed to 2. This dimension corresponds to the parameters R20A and R20B so it is always fixed to 2. Regards, Edward On 6 June 2014 11:51, Troels Emtekær Linnet <tlinnet@xxxxxxxxxxxxx> wrote:Hi. Another way is: ml = params[:end_index[1]].reshape(num_spins*2, num_frq) R20A = ml[::2].flatten() R20B = ml[1::2].flatten() Best Troels 2014-06-06 11:39 GMT+02:00 Troels Emtekær Linnet <tlinnet@xxxxxxxxxxxxx>:There is no doubt that it is the unpacking of the R20A and R20B in the target function. I was thinking of creating a function, which do the the unpacking. This unpacking function could then be tested with a unit test? What do you think? Where should I position such a function? Best Troels 2014-06-06 11:26 GMT+02:00 Edward d'Auvergne <edward@xxxxxxxxxxxxx>:Hi Troels, The best way to handle this is to first create a unit test of thespecific_analyses.relax_disp.parameters.disassemble_param_vector()where the problem is likely to be most easilyfound.I don't understand how this could be a problem as the assemble_param_vector() and disassemble_param_vector() functions both call the same loop_parameters() function for the ordering of the parameter values! Maybe the problem is in the unpacking of the parameter vector in the target functions themselves, for example in thefullB14 model: def func_B14_full(self, params): """Target function for the Baldwin (2014) 2-site exact solution model for all time scales. This assumes that pA > pB, and hence this must be implemented as a constraint. @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) # Unpack the parameter values. R20A = params[:self.end_index[0]] R20B = params[self.end_index[0]:self.end_index[1]] dw = params[self.end_index[1]:self.end_index[2]] pA = params[self.end_index[2]] kex = params[self.end_index[2]+1] # Calculate and return the chi-squared value. return self.calc_B14_chi2(R20A=R20A, R20B=R20B, dw=dw, pA=pA, kex=kex) This R20A and R20B unpacking might be the failure point as this may not match the loop_parameters() function - whichitmust! In any case, having a unit or system test catch theproblemwould be very useful for the stability of the dispersionanalysisin relax. A code example might be useful: R20_params = array([1, 2, 3, 4]) R20A, R20B = transpose(R20_params.reshape(2, 2) print(R20A) print(R20B) You should see that R20A is [1, 3], and R20B is[2,4]. This is how the parameters are handled in theloop_parameters()function which defines the parameter vector in all parts ofrelax.There might be a quicker way to unpack the parameters, but such an idea could be used for the target functions. Cheers, Edward On 6 June 2014 11:08, Troels E. Linnet <NO-REPLY.INVALID-ADDRESS@xxxxxxx> wrote:URL: <http://gna.org/bugs/?22146> Summary: Unpacking of R2A andR2Bis performed wrong for clustered "full" dispersion models Project: relax Submitted by: tlinnet Submitted on: Fri 06 Jun 201409:08:58AM UTC Category: relax's source code Specific analysis category: RelaxationdispersionPriority: 9 - Immediate Severity: 4 - Important Status: None Assigned to: None Originator Name: Originator Email: Open/Closed: Open Release: Repository: trunk Discussion Lock: Any Operating System: All systems_______________________________________________________Details: The unpacking of the R2A and R2B parameters inthe"full" model is performed wrong. This will happen performing a clusteredanalysis,using one of the "full" models. This bug affect all analysis performed running with a "full" model, with clustered residues. The bug is located in the target function: ./target_functions/relax_disp.py For all the "func_MODEL_full", the unpacking of: R20A = params[:self.end_index[0]] R20B =params[self.end_index[0]:self.end_index[1]]This is wrong, since the "params" list, is ordered: [spin, spin, spin, [dw], pA, kex], where spin = [nr_frq*r2a, nr_frq*r2b] This ordering happens in: ./specific_analysis/relax_disp/parameters.py in the loop_parameters.py A possible solutions i shown below. This alter the unpacking of the parameters. An example of profiling_cr72.py is attached. This can be downloaded, and run in base folderofrelax: ./profiling_cr72.py . This is with 3 frq, and 3 spins. The current implementations would unpack: ('R20A', array([ 2., 2., 2., 4., 4., 4., 12., 12., 12.]), 9) ('R20B', array([ 14., 14., 14., 22., 22., 22., 24., 24., 24.]), 9) R2A is 2, 12, 22 for the spins 0-3 R2B is, 4, 14, 24 for the spins 0-3 The suggested unpacking loop, unpacks to: ('R20A', array([ 2., 2., 2., 12., 12., 12., 22., 22., 22.]), 9) ('R20B', array([ 4., 4., 4., 14., 14., 14., 24., 24., 24.]), 9) ------- from numpy import array, concatenate, delete, index_exp import numpy p = array([ 1.000000000000000e+01, 1.000000000000000e+01, 1.100000000000000e+01 , 1.100000000000000e+01, 1.000000000000000e+01, 1.000000000000000e+01 , 1.100000000000000e+01, 1.100000000000000e+01, 1.000000000000000e+00 , 1.000000000000000e+00, 9.000000000000000e-01, 1.000000000000000e+03]) e = [4, 8, 10] # Now r2a = p[:e[0]] print r2a r2b = p[e[0]:e[1]] print r2b dw = p[e[1]:e[2]] print dw pA = p[e[2]] print pA kex = p[e[2]+1] print kex print "new" ns = 2 nf = 2 ml = p[:e[1]] R20A = array([]) R20B = array([]) for i in range(0, ns): # Array sorted per [spin, spin, spin], where spin = [nr_frq*r2a, nr_frq*r2b] spin_AB = ml[:nf*2] ml = delete(ml, numpy.s_[:nf*2]) R20A = concatenate([R20A, spin_AB[:nf] ]) R20B = concatenate([R20B, spin_AB[nf:] ]) print R20A print R20B print dw print pA print kex_______________________________________________________File Attachments:-------------------------------------------------------Date: Fri 06 Jun 2014 09:08:58 AM UTC Name: profiling_cr72.py Size: 17kB By: tlinnet <http://gna.org/bugs/download.php?file_id=20938_______________________________________________________Reply to this item at: <http://gna.org/bugs/?22146> _______________________________________________ Message sent via/by Gna! http://gna.org/ _______________________________________________ relax (http://www.nmr-relax.com) This is the relax-devel mailing list relax-devel@xxxxxxx To unsubscribe from this list, get a password reminder, or change your subscription options, visit the list information page at https://mail.gna.org/listinfo/relax-devel