MAINT: Make beta calculation robust to missing values.

Risk calculations are robust to nans, except for
beta which calls numpy with the complete list of
algorithm_returns. If nans are present the result
of covar will be nan.

This is fixed by filtering out nans in
algorithm_returns.

zipline/finance/risk/cumulative.py

@@ -441,8 +441,13 @@ def calculate_beta(self):
         if len(self.annualized_mean_returns) < 2:
             return 0.0
 
-        returns_matrix = np.vstack([self.algorithm_returns,
-                                    self.benchmark_returns])
+        # Drop nans if there are gaps in the data
+        algorithm_returns = self.algorithm_returns.dropna()
+        benchmark_returns = \
+            self.benchmark_returns.loc[algorithm_returns.index]
+
+        returns_matrix = np.vstack([algorithm_returns,
+                                    benchmark_returns])
         C = np.cov(returns_matrix, ddof=1)
         algorithm_covariance = C[0][1]
         benchmark_variance = C[1][1]