Closes #3

Updated use of bidict to take advantage of new API.

Author: capcarr

biosppy/biometrics.py

@@ -57,55 +57,6 @@ def __str__(self):
         return str("Combination of empty array.")
 
 
-class SubjectDict(bidict):
-    """Adaptation of bidirectional dictionary to return default values
-    on KeyError.
-
-    Attributes
-    ----------
-    LEFT : hashable
-        Left default token.
-    Right : hashable
-        Right default token.
-
-    """
-
-    LEFT = ''
-    RIGHT = ''
-
-    def __getitem__(self, keyorslice):
-        """Get an item; based on the bidict source."""
-
-        try:
-            start = keyorslice.start
-            stop = keyorslice.stop
-            step = keyorslice.step
-        except AttributeError:
-            # keyorslice is a key, e.g. b[key]
-            try:
-                return self._fwd[keyorslice]
-            except KeyError:
-                return self.RIGHT
-
-        # keyorslice is a slice
-        if (not ((start is None) ^ (stop is None))) or step is not None:
-            raise TypeError('Slice must only specify either start or stop')
-
-        if start is not None:
-            # forward lookup (by key), e.g. b[key:]
-            try:
-                return self._fwd[start]
-            except KeyError:
-                return self.RIGHT
-
-        # inverse lookup (by val), e.g. b[:val]
-        assert stop is not None
-        try:
-            return self._bwd[stop]
-        except KeyError:
-            return self.LEFT
-
-
 class BaseClassifier(object):
     """Base biometric classifier class.
 
@@ -133,7 +84,7 @@ class BaseClassifier(object):
     def __init__(self):
         # generic self things
         self.is_trained = False
-        self._subject2label = SubjectDict()
+        self._subject2label = bidict()
         self._nbSubjects = 0
         self._thresholds = {}
         self._autoThresholds = None
@@ -296,7 +247,7 @@ def list_subjects(self):
 
         """
 
-        subjects = [self._subject2label[:i] for i in xrange(self._nbSubjects)]
+        subjects = self._subject2label.keys()
 
         return subjects
 
@@ -448,7 +399,7 @@ def update_thresholds(self, fraction=1.):
 
         # gather data to test
         data = {}
-        for subject, label in self._subject2label.items():
+        for subject, label in self._subject2label.iteritems():
             # select a random fraction of the training data
             aux = self.io_load(label)
             indx = range(len(aux))
@@ -460,7 +411,7 @@ def update_thresholds(self, fraction=1.):
         _, res = self.evaluate(data, ths)
 
         # choose thresholds at EER
-        for subject, label in self._subject2label.items():
+        for subject, label in self._subject2label.iteritems():
             EER_auth = res['subject'][subject]['authentication']['rates']['EER']
             self.set_auth_thr(label, EER_auth[self.EER_IDX, 0], ready=True)
 
@@ -653,7 +604,7 @@ def identify(self, data, threshold=None):
         labels = self._identify(aux, threshold)
 
         # translate class labels
-        subjects = [self._subject2label[:item] for item in labels]
+        subjects = [self._subject2label.inv.get(item, '') for item in labels]
 
         return subjects
 
@@ -1216,7 +1167,7 @@ class SVM(BaseClassifier):
         Degree of the polynomial kernel function (‘poly’). Ignored by all other
         kernels.
     gamma : float, optional
-        Kernel coefficient for ‘rbf’, ‘poly’ and ‘sigmoid’. If gamma is 0.0
+        Kernel coefficient for ‘rbf’, ‘poly’ and ‘sigmoid’. If gamma is 'auto'
         then 1/n_features will be used instead.
     coef0 : float, optional
         Independent term in kernel function. It is only significant in ‘poly’
@@ -1246,7 +1197,7 @@ def __init__(self,
                  C=1.0,
                  kernel='linear',
                  degree=3,
-                 gamma=0.0,
+                 gamma='auto',
                  coef0=0.0,
                  shrinking=True,
                  tol=0.001,
@@ -1843,7 +1794,7 @@ def get_subject_results(results=None,
         Classifier thresholds.
     subjects : list
         Target subject classes.
-    subject_dict : SubjectDict
+    subject_dict : bidict
         Subject-label conversion dictionary.
     subject_idx : list
         Subject index.
@@ -1927,7 +1878,7 @@ def get_subject_results(results=None,
         misses = np.logical_not(np.logical_or(hits, rejects))
         nmisses = ns - (nhits + nrejects)
         missCounts = {
-            subject_dict[:ms]: np.sum(res == ms)
+            subject_dict.inv[ms]: np.sum(res == ms)
             for ms in np.unique(res[misses])
         }