load_bigquery_stats: Create empty table with schema. (#2642)

* load_bigquery_stats: Create empty table with schema.

Previously the schema was only set on load. If an engine isn't running
(and therefore no stats are output), the schema on the table is never
set, leading to exceptions in /fuzz-strategy-selection.

Fixes #2629.

* fix tests

* fix import

Author: oliverchang

src/appengine/handlers/cron/load_bigquery_stats.py

@@ -77,7 +77,7 @@ def _create_dataset_if_needed(self, bigquery, dataset_id):
 
     return self._execute_insert_request(dataset_insert)
 
-  def _create_table_if_needed(self, bigquery, dataset_id, table_id):
+  def _create_table_if_needed(self, bigquery, dataset_id, table_id, schema):
     """Create a new table if needed."""
     project_id = utils.get_application_id()
     table_body = {
@@ -91,6 +91,9 @@ def _create_table_if_needed(self, bigquery, dataset_id, table_id):
         },
     }
 
+    if schema is not None:
+      table_body['schema'] = schema
+
     table_insert = bigquery.tables().insert(
         projectId=project_id, datasetId=dataset_id, body=table_body)
     return self._execute_insert_request(table_insert)
@@ -110,14 +113,16 @@ def _load_data(self, bigquery, fuzzer):
     for kind in STATS_KINDS:
       kind_name = kind.__name__
       table_id = kind_name
-      if not self._create_table_if_needed(bigquery, dataset_id, table_id):
-        continue
 
       if kind == fuzzer_stats.TestcaseRun:
         schema = fuzzer_stats_schema.get(fuzzer)
       else:
         schema = kind.SCHEMA
 
+      if not self._create_table_if_needed(bigquery, dataset_id, table_id,
+                                          schema):
+        continue
+
       gcs_path = fuzzer_stats.get_gcs_stats_path(kind_name, fuzzer, timestamp)
       load = {
           'destinationTable': {

src/clusterfuzz/_internal/tests/appengine/handlers/cron/load_bigquery_stats_test.py

@@ -20,6 +20,7 @@
 import webtest
 
 from clusterfuzz._internal.datastore import data_types
+from clusterfuzz._internal.metrics import fuzzer_stats
 from clusterfuzz._internal.tests.test_libs import helpers as test_helpers
 from clusterfuzz._internal.tests.test_libs import test_utils
 from handlers.cron import load_bigquery_stats
@@ -76,6 +77,7 @@ def test_execute(self):
                     'tableId': 'JobRun',
                     'datasetId': 'fuzzer_stats',
                 },
+                'schema': fuzzer_stats.JobRun.SCHEMA,
             },
             datasetId='fuzzer_stats',
             projectId='test-clusterfuzz'),