Fix minor issues with avro convention and missing avro_load call

mozaggregator/aggregator.py

@@ -85,7 +85,7 @@ def aggregate_metrics(
             channels,
             "normalized_app_name <> 'Fennec'"
         )
-        fennec_pings = dataset.load(
+        fennec_pings = dataset.load_avro(
             avro_prefix,
             "saved_session",
             submission_date,

mozaggregator/bigquery.py

@@ -144,7 +144,7 @@ class BigQueryDataset:
             filters.append(filter_clause)
 
         df = self.spark.read.format("avro").load(
-            "{prefix}/submission_date={submission_date}/telemetry_{doc_type}_{doc_version}".format(
+            "{prefix}/{submission_date}/{doc_type}_{doc_version}".format(
                 prefix=prefix,
                 submission_date=submission_date,
                 doc_type=doc_type,