Fixes crashes when using numpy log on a batch transform data panel.

The recent change to the creation of the data panel ended up with
a panel with the dtype of 'object', which was causing numpy ufuncs
like `log` to crash out on an `AttributeError`.

This forces all frames in the panel to use a dtype of 'float',
we may want to look at seeting a dtype on a frame by frame basis,
e.g. 'volume' may more accurately be 'int'.

zipline/test_algorithms.py

@@ -267,6 +267,12 @@ def return_data(data, *args, **kwargs):
     return data
 
 
+@batch_transform
+def uses_ufunc(data, *args, **kwargs):
+    # ufuncs like np.log should not crash
+    return np.log(data)
+
+
 class BatchTransformAlgorithm(TradingAlgorithm):
     def initialize(self, *args, **kwargs):
         self.refresh_period = kwargs.pop('refresh_period', 1)
@@ -348,6 +354,12 @@ class BatchTransformAlgorithm(TradingAlgorithm):
             compute_only_full=False
         )
 
+        self.uses_ufunc = uses_ufunc(
+            refresh_period=self.refresh_period,
+            window_length=self.window_length,
+            clean_nans=False
+        )
+
         self.iter = 0
 
         self.set_slippage(FixedSlippage())
@@ -364,6 +376,7 @@ class BatchTransformAlgorithm(TradingAlgorithm):
             self.return_ticks.handle_data(data))
         self.history_return_not_full.append(
             self.return_not_full.handle_data(data))
+        self.uses_ufunc.handle_data(data)
 
         new_data = deepcopy(data)
         for sid in new_data:

zipline/transforms/utils.py

@@ -472,7 +472,8 @@ class BatchTransform(EventWindow):
         # user-overloaded get_value() method.
         data_dict = {tick['dt']: tick['data'] for tick in self.ticks}
         data = pd.Panel(data_dict, major_axis=self.field_names,
-                        minor_axis=self.sids)
+                        minor_axis=self.sids,
+                        dtype='float')
 
         if self.supplemental_data:
             # item will be a date stamp