Ingester: Fixing imports, calling store_latest after default store me…

…thod, adding --latest-table to cli.
planetlabs · Jun 27, 2024 · 03de50a · 03de50a
1 parent e0f34b0
commit 03de50a
Show file tree

Hide file tree

Showing 2 changed files with 12 additions and 10 deletions.
diff --git a/ingester/datalake_ingester/storage.py b/ingester/datalake_ingester/storage.py
@@ -78,6 +78,12 @@ def store_latest(self, record):
         expression_attribute_values = {
             ':new_start': {'N': str(record['metadata']['start'])}
         }
+
+        if record['metadata']['work_id'] is None:
+            work_id_value = {'BOOL': False}
+        else:
+            work_id_value = {'S': str(record['metadata']['work_id'])}
+
         record = {
             'what_where_key': {"S": record['metadata']['what']+':'+record['metadata']['where']},
             'time_index_key': {"S": record['time_index_key']},
@@ -108,9 +114,7 @@ def store_latest(self, record):
                     'where': {
                         'S': str(record['metadata']['where'])
                     },
-                    'work_id': {
-                        'S': str(record['metadata']['work_id'])
-                    }
+                    'work_id': work_id_value
                 }
             },
             'url': {"S": record['url']},

diff --git a/ingester/tests/test_ingester.py b/ingester/tests/test_ingester.py
@@ -36,19 +36,17 @@ def test_ingest_random(storage, dynamodb_records_table, random_s3_file_maker):
 def test_ingest_random_latest(storage, dynamodb_latest_table, random_s3_file_maker):
     storage.latest_table_name = 'latest'
     url, metadata = random_s3_file_maker()
-    if metadata['work_id'] == 'None':
-        metadata['work_id'] = None
     ingester = Ingester(storage)
     ingester.ingest(url)
     records = [dict(r) for r in dynamodb_latest_table.scan()]
-    def convert_metadata(metadata):
-        return {k: (decimal.Decimal(str(v)) if isinstance(v, (int, float)) else v) for k, v in metadata.items()}
-
-    converted_metadata = convert_metadata(metadata)
 
+    def convert_records(records):
+        return {k: (decimal.Decimal(str(v)) if isinstance(v, (int, float)) else v) for k, v in records[0].items()}
+
+    converted_records = convert_records(records)
     assert len(records) >= 1
     for r in records:
-        assert r['metadata'] == converted_metadata
+        assert r['metadata'] == converted_records['metadata']
 
 
 def test_ingest_no_end(storage, dynamodb_records_table, s3_file_from_metadata,