MAINT: use column's missing value.

STY: fix flake8 failures.

MAINT: fixes for compatibility with py2.

MAINT: fix import error.

MAINT: use dict.items() for compatibility with py3.

tests/pipeline/test_buyback_auth.py

@@ -119,7 +119,7 @@ def zip_with_dates(index_dates, dts):
     return pd.Series(pd.to_datetime(dts), index=index_dates)
 
 
-class BuybackAuthLoaderCommonTest:
+class BuybackAuthLoaderCommonTest(object):
     """
     Tests for loading the buyback authorization announcement data.
     """
@@ -294,8 +294,8 @@ class CashBuybackAuthLoaderTestCase(TestCase, BuybackAuthLoaderCommonTest):
             #  until 15?
             A: zip_with_floats_dates(
                 ['NaN'] * num_days_between(dates, None, '2014-01-14') +
-               [10] * num_days_between_dates('2014-01-15', '2014-01-19') +
-               [20] * num_days_between_dates('2014-01-20', None)
+                [10] * num_days_between_dates('2014-01-15', '2014-01-19') +
+                [20] * num_days_between_dates('2014-01-20', None)
             ),
             B: zip_with_floats_dates(
                 ['NaN'] * num_days_between_dates(None, '2014-01-14') +

tests/pipeline/test_earnings.py

@@ -393,9 +393,11 @@ class EarningsCalendarLoaderInferTimestampTestCase(TestCase):
         )
         assert_series_equal(
             loader.events_by_sid[1].loc[:, ANNOUNCEMENT_FIELD_NAME],
-            pd.Series(index=announcement_dates[1].loc[:, TS_FIELD_NAME],
-                      data=np.array(
-                          announcement_dates[1].loc[:, ANNOUNCEMENT_FIELD_NAME]
-                      ),
-                      name=ANNOUNCEMENT_FIELD_NAME)
+            pd.Series(
+                index=announcement_dates[1].loc[:, TS_FIELD_NAME],
+                data=np.array(
+                    announcement_dates[1].loc[:, ANNOUNCEMENT_FIELD_NAME]
+                ),
+                name=ANNOUNCEMENT_FIELD_NAME
+            )
         )

zipline/pipeline/loaders/blaze/events.py

@@ -119,4 +119,4 @@ class BlazeEventsCalendarLoader(PipelineLoader):
         ).load_adjusted_array(columns, dates, assets, mask)
 
     def prepare_data(self, raw, gb):
-        return {sid: raw.loc[group] for sid, group in gb.groups.iteritems()}
+        return {sid: raw.loc[group] for sid, group in gb.groups.items()}

zipline/pipeline/loaders/buyback_auth.py

@@ -6,7 +6,7 @@ from ..data.buyback_auth import (
     CashBuybackAuthorizations,
     ShareBuybackAuthorizations
 )
-from events import EventsLoader
+from .events import EventsLoader
 from zipline.utils.memoize import lazyval
 
 

zipline/pipeline/loaders/earnings.py

@@ -3,7 +3,7 @@ Reference implementation for EarningsCalendar loaders.
 """
 
 from ..data.earnings import EarningsCalendar
-from events import EventsLoader
+from .events import EventsLoader
 from zipline.utils.memoize import lazyval
 
 ANNOUNCEMENT_FIELD_NAME = "announcement_date"

zipline/pipeline/loaders/events.py

@@ -1,5 +1,3 @@
-from abc import abstractmethod
-
 import numpy as np
 import pandas as pd
 from six import iteritems
@@ -65,7 +63,7 @@ class EventsLoader(PipelineLoader):
                     raise ValueError(
                         "Got DatetimeIndex for sid %d.\n"
                         "Pass `infer_timestamps=True` to use the first date in"
-                        " `all_dates` as implicit timestamp."% k
+                        " `all_dates` as implicit timestamp." % k
                     )
                 self.events_by_sid[k] = pd.DataFrame(v)
                 v.index = [dates[0]] * len(v)
@@ -76,7 +74,7 @@ class EventsLoader(PipelineLoader):
                         raise ValueError(
                             "Got DataFrame without a '%s' column for sid %d.\n"
                             "Pass `infer_timestamps=True` to use the first "
-                            "date in `all_dates` as implicit timestamp."%
+                            "date in `all_dates` as implicit timestamp." %
                             (TS_FIELD_NAME, k)
                         )
                     self.events_by_sid[k] = v = v.copy()
@@ -85,7 +83,7 @@ class EventsLoader(PipelineLoader):
                     self.events_by_sid[k] = v.set_index(TS_FIELD_NAME)
             else:
                 raise ValueError("Data for sid %s must be in DataFrame, "
-                                 "Series, or DatetimeIndex."% k)
+                                 "Series, or DatetimeIndex." % k)
             # Once data is in a DF, make sure columns are correct.
             cols_except_ts = (set(v.columns.values) -
                               {TS_FIELD_NAME} -
@@ -150,9 +148,7 @@ class EventsLoader(PipelineLoader):
                 event_date_field_name,
                 value_field_name,
                 previous_value_field.dtype,
-                # TODO: need to get actual name/method to use to get missing
-                # value
-                None
+                previous_value_field.missing_value
             ),
             adjustments=None,
         )

zipline/pipeline/loaders/utils.py

@@ -1,7 +1,6 @@
 import datetime
 
 import numpy as np
-from numpy import NaN
 import pandas as pd
 from six import iteritems
 from six.moves import zip
@@ -125,14 +124,9 @@ def previous_value(date_index, events_by_sid, event_date_field, value_field,
     next_date_frame
     """
     sids = list(events_by_sid)
-    # TODO: generalize; need to use dtype of column and missing value for that
-    # column; so pass
-    # in the pipeline column's attributes for these (replace NaN and dtype
-    # below)
     out = np.full(
         (len(date_index), len(sids)),
-        # TODO; replace with missing_value
-        NaN,
+        missing_value,
         dtype=value_field_dtype
     )
     d_n = date_index[-1].asm8