Merge branch 'dev' of ssh://github.com/dream-olfaction/olfaction-prediction into dev

Author: rgerkin

opc_python/gerkin/fit2.py

@@ -52,11 +52,11 @@ def rfc_maker(n_estimators=n_estimators,max_features=max_features,
                    n_jobs=-1, random_state=seed, **kwargs)
 
     n_descriptors = len(descriptors) 
-    p = ProgressBar(n_descriptors)
+    p = ProgressBar(n_descriptors * 2)
     rfcs = {x:{} for x in ('mean','std')}
     for d,descriptor in enumerate(descriptors*2):
-        p.animate(d,"Fitting %s" % descriptor)
         kind = 'std' if d >= len(descriptors) else 'mean'
+        p.animate(d,"Fitting %s %s" % (descriptor, kind))
         rfcs[kind][descriptor] = rfc_maker(n_estimators=n_estimators,
                                         max_features=max_features[d],
                                         min_samples_leaf=min_samples_leaf[d],
@@ -438,4 +438,4 @@ def feature_sweep(X,Y,n_estimators=50,n_splits=25,
                 # Compute the correlation coefficient.
                 rs[d,i,j] = np.corrcoef(predicted,observed.loc[test])[1,0] 
     p.animate(None,'Finished') 
-    return rs
+    return rs

opc_python/utils/loading.py

@@ -284,9 +284,9 @@ def get_CID_dilutions(kind, target_dilution=None, cached=True):
             data = list(set(data))
         elif kind == 'training-norep':
             training = set(get_CID_dilutions('training',
-                                             target_dilution=target_dilution))
+                                             target_dilution=target_dilution, cached=cached))
             replicated = set(get_CID_dilutions('replicated',
-                                               target_dilution=target_dilution))
+                                               target_dilution=target_dilution, cached=cached))
             data = list(training.difference(replicated))
     data = sorted(data)
     return data