mypy: Fix strict-optional in analytics.

analytics/lib/counts.py

@@ -143,6 +143,7 @@ def do_fill_count_stat_at_hour(stat, end_time):
     start_time = end_time - stat.interval
     if not isinstance(stat, LoggingCountStat):
         timer = time.time()
+        assert(stat.data_collector.pull_function is not None)
         rows_added = stat.data_collector.pull_function(stat.property, start_time, end_time)
         logger.info("%s run pull_function (%dms/%sr)" %
                     (stat.property, (time.time()-timer)*1000, rows_added))

analytics/management/commands/populate_analytics_db.py

@@ -17,7 +17,7 @@ from zerver.models import Realm, UserProfile, Stream, Message, Client, \
 from datetime import datetime, timedelta
 
 from six.moves import zip
-from typing import Any, Dict, List, Optional, Text, Type, Union
+from typing import Any, Dict, List, Optional, Text, Type, Union, Mapping
 
 class Command(BaseCommand):
     help = """Populates analytics tables with randomly generated data."""
@@ -60,7 +60,7 @@ class Command(BaseCommand):
         shylock = self.create_user('shylock@analytics.ds', 'Shylock', True, installation_time, realm)
 
         def insert_fixture_data(stat, fixture_data, table):
-            # type: (CountStat, Dict[Optional[str], List[int]], Type[BaseCount]) -> None
+            # type: (CountStat, Mapping[Optional[str], List[int]], Type[BaseCount]) -> None
             end_times = time_range(last_end_time, last_end_time, stat.frequency,
                                    len(list(fixture_data.values())[0]))
             if table == RealmCount:
@@ -76,13 +76,14 @@ class Command(BaseCommand):
         stat = COUNT_STATS['realm_active_humans::day']
         realm_data = {
             None: self.generate_fixture_data(stat, .1, .03, 3, .5, 3, partial_sum=True),
-        } # type: Dict[Optional[str], List[int]]
+        } # type: Mapping[Optional[str], List[int]]
         insert_fixture_data(stat, realm_data, RealmCount)
         FillState.objects.create(property=stat.property, end_time=last_end_time,
                                  state=FillState.DONE)
 
         stat = COUNT_STATS['messages_sent:is_bot:hour']
-        user_data = {'false': self.generate_fixture_data(stat, 2, 1, 1.5, .6, 8, holiday_rate=.1)}
+        user_data = {'false': self.generate_fixture_data(
+            stat, 2, 1, 1.5, .6, 8, holiday_rate=.1)}  # type: Mapping[Optional[str], List[int]]
         insert_fixture_data(stat, user_data, UserCount)
         realm_data = {'false': self.generate_fixture_data(stat, 35, 15, 6, .6, 4),
                       'true': self.generate_fixture_data(stat, 15, 15, 3, .4, 2)}

analytics/tests/test_counts.py

@@ -201,7 +201,7 @@ class TestProcessCountStat(AnalyticsTestCase):
         with self.assertRaises(ValueError):
             process_count_stat(stat, installation_epoch() + 65*self.MINUTE)
         with self.assertRaises(ValueError):
-            process_count_stat(stat, installation_epoch().replace(tzinfo=None) + self.HOUR)
+            process_count_stat(stat, installation_epoch().replace(tzinfo=None) + self.HOUR)  # type: ignore # https://github.com/python/typeshed/pull/1347
 
     # This tests the LoggingCountStat branch of the code in do_delete_counts_at_hour.
     # It is important that do_delete_counts_at_hour not delete any of the collected

analytics/tests/test_views.py

@@ -19,7 +19,7 @@ import mock
 import ujson
 
 from six.moves import range
-from typing import List, Dict
+from typing import List, Dict, Optional
 
 class TestStatsEndpoint(ZulipTestCase):
     def test_stats(self):
@@ -47,7 +47,7 @@ class TestGetChartData(ZulipTestCase):
         return [0, 0, i, 0]
 
     def insert_data(self, stat, realm_subgroups, user_subgroups):
-        # type: (CountStat, List[str], List[str]) -> None
+        # type: (CountStat, List[Optional[str]], List[str]) -> None
         if stat.frequency == CountStat.HOUR:
             insert_time = self.end_times_hour[2]
             fill_time = self.end_times_hour[-1]

analytics/views.py

@@ -183,7 +183,7 @@ def rewrite_client_arrays(value_arrays):
     return mapped_arrays
 
 def get_time_series_by_subgroup(stat, table, key_id, end_times, subgroup_to_label, include_empty_subgroups):
-    # type: (CountStat, Type[BaseCount], Optional[int], List[datetime], Dict[Optional[str], str], bool) -> Dict[str, List[int]]
+    # type: (CountStat, Type[BaseCount], int, List[datetime], Dict[Optional[str], str], bool) -> Dict[str, List[int]]
     queryset = table_filtered_to_id(table, key_id).filter(property=stat.property) \
                                                   .values_list('subgroup', 'end_time', 'value')
     value_dicts = defaultdict(lambda: defaultdict(int)) # type: Dict[Optional[str], Dict[datetime, int]]