From bd34376119b57c891fafef8497245232176d80eb Mon Sep 17 00:00:00 2001 From: Jean-Philippe ARGAUD Date: Mon, 22 Jan 2018 11:42:35 +0100 Subject: [PATCH] Update algorithmic test --- ...erification_des_Assimilation_Algorithms.py | 35 +++++++++++-------- 1 file changed, 20 insertions(+), 15 deletions(-) diff --git a/test/test6901/Verification_des_Assimilation_Algorithms.py b/test/test6901/Verification_des_Assimilation_Algorithms.py index be533a8..bf6ded8 100644 --- a/test/test6901/Verification_des_Assimilation_Algorithms.py +++ b/test/test6901/Verification_des_Assimilation_Algorithms.py @@ -22,9 +22,10 @@ "Verification de la disponibilite de l'ensemble des algorithmes" # ============================================================================== -import adaoBuilder, numpy +import adaoBuilder, numpy, sys def test1(): """Verification de la disponibilite de l'ensemble des algorithmes\n(Utilisation d'un operateur matriciel)""" + print(test1.__doc__) Xa = {} for algo in ("3DVAR", "Blue", "ExtendedBlue", "LinearLeastSquares", "NonLinearLeastSquares", "DerivativeFreeOptimization"): print("") @@ -43,13 +44,13 @@ def test1(): Xa[algo] = adaopy.get("Analysis")[-1] del adaopy # - for algo in ("ExtendedKalmanFilter", "KalmanFilter", "UnscentedKalmanFilter", "4DVAR"): + for algo in ("ExtendedKalmanFilter", "KalmanFilter", "UnscentedKalmanFilter", "EnsembleKalmanFilter", "4DVAR"): print("") msg = "Algorithme en test : %s"%algo print(msg+"\n"+"-"*len(msg)) # adaopy = adaoBuilder.New() - adaopy.setAlgorithmParameters(Algorithm=algo, Parameters={"EpsilonMinimumExponent":-10, }) + adaopy.setAlgorithmParameters(Algorithm=algo, Parameters={"EpsilonMinimumExponent":-10, "SetSeed":1000}) adaopy.setBackground (Vector = [0,1,2]) adaopy.setBackgroundError (ScalarSparseMatrix = 1.) adaopy.setObservation (Vector = [0.5,1.5,2.5]) @@ -68,7 +69,7 @@ def test1(): print(msg+"\n"+"-"*len(msg)) # adaopy = adaoBuilder.New() - adaopy.setAlgorithmParameters(Algorithm=algo, Parameters={"BoxBounds":3*[[-1,3]], "SetSeed":1000, }) + adaopy.setAlgorithmParameters(Algorithm=algo, Parameters={"BoxBounds":3*[[-1,3]], "SetSeed":1000}) adaopy.setBackground (Vector = [0,1,2]) adaopy.setBackgroundError (ScalarSparseMatrix = 1.) adaopy.setObservation (Vector = [0.5,1.5,2.5]) @@ -99,9 +100,10 @@ def test1(): print("") msg = "Tests des ecarts attendus :" print(msg+"\n"+"="*len(msg)) - verify_similarity_of_algo_results(("3DVAR", "Blue", "ExtendedBlue", "4DVAR", "DerivativeFreeOptimization"), Xa) - verify_similarity_of_algo_results(("LinearLeastSquares", "NonLinearLeastSquares"), Xa) - verify_similarity_of_algo_results(("ExtendedKalmanFilter", "KalmanFilter", "UnscentedKalmanFilter"), Xa) + verify_similarity_of_algo_results(("3DVAR", "Blue", "ExtendedBlue", "4DVAR", "DerivativeFreeOptimization"), Xa, 5.e-5) + verify_similarity_of_algo_results(("LinearLeastSquares", "NonLinearLeastSquares"), Xa, 5.e-7) + verify_similarity_of_algo_results(("KalmanFilter", "ExtendedKalmanFilter", "UnscentedKalmanFilter"), Xa, 1.e-14) + verify_similarity_of_algo_results(("KalmanFilter", "EnsembleKalmanFilter"), Xa, 5.e-2) print(" Les resultats obtenus sont corrects.") print("") # @@ -109,6 +111,7 @@ def test1(): def test2(): """Verification de la disponibilite de l'ensemble des algorithmes\n(Utilisation d'un operateur fonctionnel)""" + print(test2.__doc__) Xa = {} M = numpy.matrix("1 0 0;0 2 0;0 0 3") def H(x): return M * numpy.asmatrix(numpy.ravel( x )).T @@ -131,13 +134,13 @@ def test2(): # M = numpy.matrix("1 0 0;0 2 0;0 0 3") def H(x): return M * numpy.asmatrix(numpy.ravel( x )).T - for algo in ("ExtendedKalmanFilter", "KalmanFilter", "UnscentedKalmanFilter", "4DVAR"): + for algo in ("ExtendedKalmanFilter", "KalmanFilter", "EnsembleKalmanFilter", "UnscentedKalmanFilter", "4DVAR"): print("") msg = "Algorithme en test : %s"%algo print(msg+"\n"+"-"*len(msg)) # adaopy = adaoBuilder.New() - adaopy.setAlgorithmParameters(Algorithm=algo, Parameters={"EpsilonMinimumExponent":-10, }) + adaopy.setAlgorithmParameters(Algorithm=algo, Parameters={"EpsilonMinimumExponent":-10, "SetSeed":1000}) adaopy.setBackground (Vector = [0,1,2]) adaopy.setBackgroundError (ScalarSparseMatrix = 1.) adaopy.setObservation (Vector = [0.5,1.5,2.5]) @@ -158,7 +161,7 @@ def test2(): print(msg+"\n"+"-"*len(msg)) # adaopy = adaoBuilder.New() - adaopy.setAlgorithmParameters(Algorithm=algo, Parameters={"BoxBounds":3*[[-1,3]], "SetSeed":1000, }) + adaopy.setAlgorithmParameters(Algorithm=algo, Parameters={"BoxBounds":3*[[-1,3]], "SetSeed":1000}) adaopy.setBackground (Vector = [0,1,2]) adaopy.setBackgroundError (ScalarSparseMatrix = 1.) adaopy.setObservation (Vector = [0.5,1.5,2.5]) @@ -172,8 +175,9 @@ def test2(): print("") msg = "Tests des ecarts attendus :" print(msg+"\n"+"="*len(msg)) - verify_similarity_of_algo_results(("3DVAR", "Blue", "ExtendedBlue", "4DVAR", "DerivativeFreeOptimization"), Xa) - verify_similarity_of_algo_results(("ExtendedKalmanFilter", "KalmanFilter", "UnscentedKalmanFilter"), Xa) + verify_similarity_of_algo_results(("3DVAR", "Blue", "ExtendedBlue", "4DVAR", "DerivativeFreeOptimization"), Xa, 5.e-5) + verify_similarity_of_algo_results(("KalmanFilter", "ExtendedKalmanFilter", "UnscentedKalmanFilter"), Xa, 1.e14) + verify_similarity_of_algo_results(("KalmanFilter", "EnsembleKalmanFilter"), Xa, 5.e-2) print(" Les resultats obtenus sont corrects.") print("") # @@ -184,13 +188,14 @@ def almost_equal_vectors(v1, v2, precision = 1.e-15, msg = ""): print(" Difference maximale %s: %.2e"%(msg, max(abs(v2 - v1)))) return max(abs(v2 - v1)) < precision -def verify_similarity_of_algo_results(serie = [], Xa = {}): +def verify_similarity_of_algo_results(serie = [], Xa = {}, precision = 1.e-15): print(" Comparaisons :") for algo1 in serie: for algo2 in serie: if algo1 is algo2: break - assert almost_equal_vectors( Xa[algo1], Xa[algo2], 5.e-5, "entre %s et %s "%(algo1, algo2) ) - print(" Algorithmes dont les resultats sont similaires : %s\n"%(serie,)) + assert almost_equal_vectors( Xa[algo1], Xa[algo2], precision, "entre %s et %s "%(algo1, algo2) ) + print(" Algorithmes dont les resultats sont similaires a %.0e : %s\n"%(precision, serie,)) + sys.stdout.flush() #=============================================================================== if __name__ == "__main__": -- 2.39.2