fixed bug in benchmark script

do_benchmark.py

@@ -16,11 +16,8 @@ NAME = args.name
 FIRST_ID = args.first_id
 LAST_ID = args.last_id
 
-MODEL = args.model
 N_BINS = args.nbins
 NOISE_VAR = args.noise_var
-ITERATION_LIMIT = args.iteration_limit
-TOL_REL_GRADNORM = args.tol_rel_gradnorm
 BENCHMARK = args.benchmark
 VERBOSITY = args.verbosity
 
@@ -35,15 +32,12 @@ if CONFIG is not None:
     with open('./model_configurations.txt') as f:
         configs = eval(f.read())
     parameters = configs[CONFIG]
-    MODEL = parameters.get('model', MODEL)
     N_BINS = parameters.get('nbins', N_BINS)
     NOISE_VAR = parameters.get('noise_var', NOISE_VAR)
     POWER_SPECTRUM_BETA_STR = parameters.get(
             'power_spectrum_beta', POWER_SPECTRUM_BETA_STR)
     POWER_SPECTRUM_F_STR = parameters.get(
             'power_spectrum_f', POWER_SPECTRUM_F_STR)
-    ITERATION_LIMIT = parameters.get('iteration_limit', ITERATION_LIMIT)
-    TOL_REL_GRADNORM = parameters.get('tol_rel_gradnorm', TOL_REL_GRADNORM)
 
 if LAST_ID is None:
     LAST_ID = get_benchmark_default_length(BENCHMARK)
@@ -55,14 +49,17 @@ print(
         'power spectrum beta: {}\n'
         'power spectrum f: {}\n'
         'rho: {}\n'
+        'scale_max: {}\n'
         'storing results with suffix {}'.format(
             BENCHMARK, FIRST_ID, LAST_ID, N_BINS,
             NOISE_VAR,
             POWER_SPECTRUM_BETA_STR,
             POWER_SPECTRUM_F_STR,
             RHO,
+            SCALE_MAX,
             NAME))
 
+np.random.seed(1)
 POWER_SPECTRUM_BETA = lambda q: eval(POWER_SPECTRUM_BETA_STR)
 POWER_SPECTRUM_F = lambda q: eval(POWER_SPECTRUM_F_STR)
 scale = (0, SCALE_MAX)
@@ -75,7 +72,8 @@ if os.path.isfile(prediction_file):
         prediction_file = './benchmark_predictions/{}_{}_{}.txt'.format(
                 BENCHMARK, NAME, c)
 
-accuracy = 0
+np.random.seed(1)
+
 sum_of_weights = 0
 weighted_correct = 0
 
@@ -88,13 +86,6 @@ for i in range(FIRST_ID-1, LAST_ID):
     scaler = MinMaxScaler(scale)
     x, y = scaler.fit_transform(np.array((x, y)).T).T
 
-    minimizer = nifty5.RelaxedNewton(controller=nifty5.GradientNormController(
-            tol_rel_gradnorm=TOL_REL_GRADNORM,
-            iteration_limit=ITERATION_LIMIT,
-            convergence_level=5,
-            ))
-
-    if MODEL == 1:
     bcm = bayesian_causal_model.cause_model_shallow.CausalModelShallow(
         N_bins=N_BINS,
         noise_var=NOISE_VAR,
@@ -102,21 +93,12 @@ for i in range(FIRST_ID-1, LAST_ID):
         power_spectrum_beta=POWER_SPECTRUM_BETA,
         power_spectrum_f=POWER_SPECTRUM_F,
         )
-    elif MODEL == 2:
-        bcm = bayesian_causal_model_nifty.cause_model_shallow.CausalModelShallow(
-            N_bins=N_BINS,
-            noise_var=NOISE_VAR,
-            rho=RHO,
-            power_spectrum_beta=POWER_SPECTRUM_BETA,
-            power_spectrum_f=POWER_SPECTRUM_F,
-            minimizer=minimizer,
-            )
 
     bcm.set_data(x, y)
 
-    H1 = bcm.get_evidence(direction=1, verbosity=1)
-    H2 = bcm.get_evidence(direction=-1, verbosity=1)
-    predicted_direction = 1 if int(H1 < H2) else 0
+    H1 = bcm.get_evidence(direction=1, verbosity=VERBOSITY - 1)
+    H2 = bcm.get_evidence(direction=-1, verbosity=VERBOSITY - 1)
+    predicted_direction = 1 if int(H1 < H2) else -1
 
     if predicted_direction == true_direction:
         fore = colorama.Fore.GREEN
@@ -126,6 +108,7 @@ for i in range(FIRST_ID-1, LAST_ID):
     sum_of_weights += weight
     accuracy = weighted_correct / sum_of_weights
 
+    if VERBOSITY > 0:
         print(
                 'dataset {}, {} true direction: {}, predicted direction {}\n'
                 'H1: {:.2e},\n H2: {:.2e},\n{}'
@@ -147,7 +130,6 @@ print('accuracy: {:.2f}'.format(accuracy))
 
 benchmark_information = {
         'benchmark': BENCHMARK,
-        'model': MODEL,
         'n_bins': N_BINS,
         'noise_var': NOISE_VAR,
         'rho': RHO,