metadata.yaml formatting

bigquery_etl/cli/query.py

@@ -750,7 +750,7 @@ def _validate_schema(query_file):
 
     if not query_schema.compatible(table_schema):
         click.echo(
-            f"Schema for query in {query_file_path} "
+            f"ERROR: Schema for query in {query_file_path} "
             f"incompatible with schema deployed for "
             f"{project_name}.{dataset_name}.{table_name}",
             err=True,

bigquery_etl/metadata/parse_metadata.py

@@ -207,8 +207,26 @@ class Metadata:
                 if label_value == "":
                     metadata_dict["labels"][label_key] = True
 
+        # Use literal yaml representation for descriptions
+        class Literal(str):
+            pass
+
+        def literal_presenter(dumper, data):
+            return dumper.represent_scalar("tag:yaml.org,2002:str", data, style="|")
+
+        yaml.add_representer(Literal, literal_presenter)
+
+        if "description" in metadata_dict:
+            metadata_dict["description"] = Literal(metadata_dict["description"])
+
         converter = cattr.Converter()
-        file.write_text(yaml.dump(converter.unstructure(metadata_dict)))
+        file.write_text(
+            yaml.dump(
+                converter.unstructure(metadata_dict),
+                default_flow_style=False,
+                sort_keys=False,
+            )
+        )
 
     def is_public_bigquery(self):
         """Return true if the public_bigquery flag is set."""

sql/moz-fx-data-shared-prod/telemetry_derived/clients_daily_v6/metadata.yaml

@@ -1,35 +1,29 @@
-bigquery:
-  clustering:
-    fields:
-    - normalized_channel
-    - sample_id
-  time_partitioning:
-    field: submission_date
-    require_partition_filter: true
-    type: day
-description: 'A daily aggregation over `main` pings from each Firefox
-
+friendly_name: Clients Daily
+description: |-
+  A daily aggregation over `main` pings from each Firefox
   desktop client, partitioned by day.
 
-
   It should normally be accessed through the user-facing view
-
   `telemetry.clients_daily`.
-
   Note that by end of Q1 2021, that view start referencing the downstream table
-
   `clients_last_seen_joined_v1` which merges in fields based on the `event`
-
   ping.
-
-  See https://github.com/mozilla/bigquery-etl/issues/1761'
-friendly_name: Clients Daily
+  See https://github.com/mozilla/bigquery-etl/issues/1761
+owners:
+- dthorn@mozilla.com
 labels:
   application: firefox
   schedule: daily
-owners:
-- dthorn@mozilla.com
 scheduling:
   allow_field_addition_on_date: '2020-11-13'
   dag_name: bqetl_main_summary
   start_date: '2019-11-05'
+bigquery:
+  time_partitioning:
+    field: submission_date
+    type: day
+    require_partition_filter: true
+  clustering:
+    fields:
+    - normalized_channel
+    - sample_id