COSMIT refactor LibSVM wrapper for safety and readability

* don't pass fit-time parameters to prediction functions
* do memory deallocation in a finally clause
* trim docstrings and comments

TODO:
* fit needs the finally clause as well
* sparse LibSVM should get the same treatment

(test_svc_clone_with_callable_kernel might fail sometimes. That appears
to be unrelated.)

Author: larsmans

sklearn/svm/base.py

@@ -59,8 +59,10 @@ class BaseLibSVM(six.with_metaclass(ABCMeta, BaseEstimator)):
 
     Parameter documentation is in the derived `SVC` class.
     """
-    # see ./classes.py for SVC class.
 
+    # The order of these must match the integer values in LibSVM.
+    # XXX These are actually the same in the dense case. Need to factor
+    # this out.
     _sparse_kernels = ["linear", "poly", "rbf", "sigmoid", "precomputed"]
 
     @abstractmethod
@@ -291,13 +293,10 @@ def _dense_predict(self, X):
 
         return libsvm.predict(
             X, self.support_, self.support_vectors_, self.n_support_,
-            self.dual_coef_, self._intercept_,
-            self._label, self.probA_, self.probB_,
-            svm_type=svm_type,
-            kernel=kernel, C=C, nu=self.nu,
-            probability=self.probability, degree=self.degree,
-            shrinking=self.shrinking, tol=self.tol, cache_size=self.cache_size,
-            coef0=self.coef0, gamma=self._gamma, epsilon=self.epsilon)
+            self.dual_coef_, self._intercept_, self._label,
+            self.probA_, self.probB_, svm_type=svm_type, kernel=kernel,
+            degree=self.degree, coef0=self.coef0, gamma=self._gamma,
+            cache_size=self.cache_size)
 
     def _sparse_predict(self, X):
         X = sp.csr_matrix(X, dtype=np.float64)
@@ -354,8 +353,6 @@ def decision_function(self, X):
         X = self._validate_for_predict(X)
         X = self._compute_kernel(X)
 
-        C = 0.0  # C is not useful here
-
         kernel = self.kernel
         if callable(kernel):
             kernel = 'precomputed'
@@ -365,10 +362,8 @@ def decision_function(self, X):
             self.dual_coef_, self._intercept_, self._label,
             self.probA_, self.probB_,
             svm_type=LIBSVM_IMPL.index(self._impl),
-            kernel=kernel, C=C, nu=self.nu,
-            probability=self.probability, degree=self.degree,
-            shrinking=self.shrinking, tol=self.tol, cache_size=self.cache_size,
-            coef0=self.coef0, gamma=self._gamma, epsilon=self.epsilon)
+            kernel=kernel, degree=self.degree, cache_size=self.cache_size,
+            coef0=self.coef0, gamma=self._gamma)
 
         # In binary case, we need to flip the sign of coef, intercept and
         # decision function.
@@ -539,10 +534,8 @@ def _dense_predict_proba(self, X):
             X, self.support_, self.support_vectors_, self.n_support_,
             self.dual_coef_, self._intercept_, self._label,
             self.probA_, self.probB_,
-            svm_type=svm_type, kernel=kernel, C=C, nu=self.nu,
-            probability=self.probability, degree=self.degree,
-            shrinking=self.shrinking, tol=self.tol, cache_size=self.cache_size,
-            coef0=self.coef0, gamma=self._gamma, epsilon=self.epsilon)
+            svm_type=svm_type, kernel=kernel, degree=self.degree,
+            cache_size=self.cache_size, coef0=self.coef0, gamma=self._gamma)
 
         return pprob