Adding display_name field on Elasticsearch mapping for exa… (#186)

* [AMD-1157] Adding schema_table field on Elasticsearch mapping for exact match

* Use display_name

* Update

databuilder/models/table_elasticsearch_document.py

@@ -20,12 +20,14 @@ class TableESDocument(ElasticsearchDocument):
                  total_usage,  # type: int
                  unique_usage,  # type: int
                  tags,  # type: List[str]
+                 display_name=None,  # type: Optional[str]
                  ):
         # type: (...) -> None
         self.database = database
         self.cluster = cluster
         self.schema_name = schema_name
         self.name = name
+        self.display_name = display_name if display_name else '{schema}.{table}'.format(schema=schema_name, table=name)
         self.key = key
         self.description = description
         # todo: use last_updated_timestamp to match the record in metadata

databuilder/publisher/elasticsearch_publisher.py

@@ -60,6 +60,9 @@ class ElasticsearchPublisher(Publisher):
                     }
                   }
                 },
+                "display_name": {
+                  "type": "keyword"
+                },
                 "last_updated_epoch": {
                   "type": "date",
                   "format": "epoch_second"

setup.py

@@ -2,7 +2,7 @@ import os
 from setuptools import setup, find_packages
 
 
-__version__ = '1.5.0'
+__version__ = '1.5.1'
 
 
 requirements_path = os.path.join(os.path.dirname(os.path.realpath(__file__)), 'requirements.txt')

tests/unit/extractor/test_neo4j_extractor.py

@@ -104,6 +104,7 @@ class TestNeo4jExtractor(unittest.TestCase):
                                cluster='test_cluster',
                                schema_name='test_schema',
                                name='test_table_name',
+                               display_name='test_schema.test_table_name',
                                key='test_table_key',
                                description='test_table_description',
                                last_updated_epoch=123456789,

tests/unit/loader/test_file_system_elasticsearch_json_loader.py

@@ -108,7 +108,7 @@ class TestFSElasticsearchJSONLoader(unittest.TestCase):
             ('{"key": "test_table_key", "column_descriptions": ["test_comment1", "test_comment2"], '
              '"schema_name": "test_schema", "database": "test_database", "cluster": "test_cluster", '
              '"column_names": ["test_col1", "test_col2"], "name": "test_table", '
-             '"last_updated_epoch": 123456789,'
+             '"last_updated_epoch": 123456789, "display_name": "test_schema.test_table", '
              '"description": "test_description", "unique_usage": 5, "total_usage": 10, '
              '"tags": ["test_tag1", "test_tag2"]}')
         ]
@@ -146,7 +146,7 @@ class TestFSElasticsearchJSONLoader(unittest.TestCase):
             ('{"key": "test_table_key", "column_descriptions": ["test_comment1", "test_comment2"], '
              '"schema_name": "test_schema", "database": "test_database", "cluster": "test_cluster", '
              '"column_names": ["test_col1", "test_col2"], "name": "test_table", '
-             '"last_updated_epoch": 123456789,'
+             '"last_updated_epoch": 123456789, "display_name": "test_schema.test_table", '
              '"description": "test_description", "unique_usage": 5, "total_usage": 10, '
              '"tags": ["test_tag1", "test_tag2"]}')
         ] * 5

tests/unit/models/test_table_elasticsearch_document.py

@@ -28,6 +28,7 @@ class TestTableElasticsearchDocument(unittest.TestCase):
                                   "cluster": "test_cluster",
                                   "schema_name": "test_schema",
                                   "name": "test_table",
+                                  "display_name": "test_schema.test_table",
                                   "key": "test_table_key",
                                   "last_updated_epoch": 123456789,
                                   "description": "test_table_description",