Fix skopt (#31)

* fixed hyperopt2skopt

* updated optuna convert

docs/conf.py

@@ -40,7 +40,7 @@
 # The short X.Y version
 version = '0.4'
 # The full version, including alpha/beta/rc tags
-release = '0.4.0'
+release = '0.4.1'
 
 # -- General configuration ---------------------------------------------------
 

neptunecontrib/hpo/utils.py

@@ -14,6 +14,7 @@
 # limitations under the License.
 #
 
+import numpy as np
 import pandas as pd
 from scipy.optimize import OptimizeResult
 import skopt
@@ -64,7 +65,7 @@ def hyperopt2skopt(trials, space):
         results_.setdefault('x_iters', []).append(trial_params)
         results_.setdefault('func_vals', []).append(trial['result']['loss'])
     optimize_results = OptimizeResult()
-    optimize_results.x = list(trials.argmin.values())
+    optimize_results.x = [trials.argmin[name] for name in param_names]
     optimize_results.x_iters = results_['x_iters']
     optimize_results.fun = trials.best_trial['result']['loss']
     optimize_results.func_vals = results_['func_vals']
@@ -152,7 +153,7 @@ def optuna2skopt(results):
     """
 
     results_ = results['params']
-    results_['target'] = -1.0 * results['value']
+    results_['target'] = results['value']
     return df2result(results_,
                      metric_col='target',
                      param_cols=[col for col in results_.columns if col != 'target'])
@@ -231,6 +232,7 @@ def _convert_space_hop_skopt(space):
         elif method == 'uniform':
             dimensions.append(skopt.space.Real(low, high, name=name, prior='uniform'))
         elif method == 'loguniform':
+            low, high = np.exp(low), np.exp(high)
             dimensions.append(skopt.space.Real(low, high, name=name, prior='log-uniform'))
         else:
             raise NotImplementedError

neptunecontrib/monitoring/skopt.py

@@ -92,8 +92,8 @@ def send_runs(results, experiment=None):
 
     _exp = experiment if experiment else neptune
 
-    for loss, params in zip(results.func_vals, results.x_iters):
-        _exp.send_metric('run_score', y=loss)
+    for i, (loss, params) in enumerate(zip(results.func_vals, results.x_iters)):
+        _exp.send_metric('run_score', x=i, y=loss)
 
         named_params = _format_to_named_params(params, results)
         _exp.send_text('run_parameters', str(named_params))

setup.py

@@ -10,7 +10,7 @@ def main():
         requirements = [r.strip() for r in f]
         setup(
             name='neptune-contrib',
-            version='0.4.0',
+            version='0.4.1',
             description='Neptune Python library contributions',
             author='neptune.ml',
             author_email='[email protected]',