Add default cypher query for user/dashboard entity to search extractor

databuilder/extractor/neo4j_search_data_extractor.py

@@ -15,10 +15,12 @@ class Neo4jSearchDataExtractor(Extractor):
     Use Neo4jExtractor extractor class
     """
     CYPHER_QUERY_CONFIG_KEY = 'cypher_query'
+    ENTITY_TYPE = 'entity_type'
 
-    DEFAULT_NEO4J_CYPHER_QUERY = textwrap.dedent(
+    DEFAULT_NEO4J_TABLE_CYPHER_QUERY = textwrap.dedent(
         """
-        MATCH (db:Database)<-[:CLUSTER_OF]-(cluster:Cluster)<-[:SCHEMA_OF]-(schema:Schema)<-[:TABLE_OF]-(table:Table)
+        MATCH (db:Database)<-[:CLUSTER_OF]-(cluster:Cluster)
+        <-[:SCHEMA_OF]-(schema:Schema)<-[:TABLE_OF]-(table:Table)
         {publish_tag_filter}
         OPTIONAL MATCH (table)-[:DESCRIPTION]->(table_description:Description)
         OPTIONAL MATCH (table)-[:TAGGED_BY]->(tags:Tag) WHERE tags.tag_type='default'
@@ -46,19 +48,68 @@ class Neo4jSearchDataExtractor(Extractor):
         """
     )
 
+    DEFAULT_NEO4J_USER_CYPHER_QUERY = textwrap.dedent(
+        """
+        MATCH (user:User)
+        OPTIONAL MATCH (user)-[read:READ]->(a)
+        OPTIONAL MATCH (user)-[own:OWNER_OF]->(b)
+        OPTIONAL MATCH (user)-[follow:FOLLOWED_BY]->(c)
+        OPTIONAL MATCH (user)-[manage_by:MANAGE_BY]->(manager)
+        {publish_tag_filter}
+        with user, a, b, c, read, own, follow, manager
+        where user.full_name is not null
+        return user.email as email, user.first_name as first_name, user.last_name as last_name,
+        user.full_name as full_name, user.github_username as github_username, user.team_name as team_name,
+        user.employee_type as employee_type, manager.email as manager_email,
+        user.slack_id as slack_id, user.is_active as is_active,
+        REDUCE(sum_r = 0, r in COLLECT(DISTINCT read)| sum_r + r.read_count) AS total_read,
+        count(distinct b) as total_own,
+        count(distinct c) AS total_follow
+        order by user.email
+        """
+    )
+
+    # todo: 1. change total_read once we have the usage;
+    #  2. add more fields once we have in the graph; 3. change mode to generic once add more support for dashboard
+    DEFAULT_NEO4J_DASHBOARD_CYPHER_QUERY = textwrap.dedent(
+        """
+        MATCH (db:Dashboard)
+        OPTIONAL MATCH (db)-[:DASHBOARD_OF]->(dbg:Dashboardgroup)
+        OPTIONAL MATCH (db)-[:DESCRIPTION]->(db_descr:Description)
+        OPTIONAL MATCH (dbg)-[:DESCRIPTION]->(dbg_descr:Description)
+        {publish_tag_filter}
+        with db, dbg, db_descr, dbg_descr
+        where db.name is not null
+        return dbg.name as dashboard_group, db.name as dashboard_name,
+        coalesce(db_descr.description, '') as description,
+        coalesce(dbg.description, '') as dashboard_group_description,
+        'mode' as product,
+        1 AS total_usage
+        order by dbg.name
+        """
+    )
+
+    # todo: we will add more once we add more entities
+    DEFAULT_QUERY_BY_ENTITY = {
+        'table': DEFAULT_NEO4J_TABLE_CYPHER_QUERY,
+        'user': DEFAULT_NEO4J_USER_CYPHER_QUERY,
+        'dashboard': DEFAULT_NEO4J_DASHBOARD_CYPHER_QUERY
+    }
+
     def init(self, conf):
         # type: (ConfigTree) -> None
         """
         Initialize Neo4jExtractor object from configuration and use that for extraction
         """
         self.conf = conf
-
         # extract cypher query from conf, if specified, else use default query
         if Neo4jSearchDataExtractor.CYPHER_QUERY_CONFIG_KEY in conf:
             self.cypher_query = conf.get_string(Neo4jSearchDataExtractor.CYPHER_QUERY_CONFIG_KEY)
         else:
+            entity_type = conf.get_string(Neo4jSearchDataExtractor.ENTITY_TYPE, default='table').lower()
+            default_query = Neo4jSearchDataExtractor.DEFAULT_QUERY_BY_ENTITY[entity_type]
             self.cypher_query = self._add_publish_tag_filter(conf.get_string(JOB_PUBLISH_TAG, ''),
-                                                             Neo4jSearchDataExtractor.DEFAULT_NEO4J_CYPHER_QUERY)
+                                                             cypher_query=default_query)
 
         self.neo4j_extractor = Neo4jExtractor()
         # write the cypher query in configs in Neo4jExtractor scope
@@ -98,5 +149,4 @@ def _add_publish_tag_filter(self, publish_tag, cypher_query):
             publish_tag_filter = ''
         else:
             publish_tag_filter = """WHERE table.published_tag = '{}'""".format(publish_tag)
-
         return cypher_query.format(publish_tag_filter=publish_tag_filter)

tests/unit/extractor/test_neo4j_search_data_extractor.py

@@ -1,4 +1,9 @@
 import unittest
+from mock import patch
+
+from pyhocon import ConfigFactory
+from databuilder import Scoped
+from databuilder.extractor.neo4j_extractor import Neo4jExtractor
 from databuilder.extractor.neo4j_search_data_extractor import Neo4jSearchDataExtractor
 
 
@@ -18,6 +23,25 @@ def test_not_adding_filter(self):
 
         self.assertEqual(actual, """MATCH (table:Table)  RETURN table""")
 
+    def test_default_search_query(self):
+        # type: (Any) -> None
+        with patch.object(Neo4jExtractor, '_get_driver'):
+            extractor = Neo4jSearchDataExtractor()
+            conf = ConfigFactory.from_dict({
+                'extractor.search_data.extractor.neo4j.{}'.format(Neo4jExtractor.GRAPH_URL_CONFIG_KEY):
+                    'test-endpoint',
+                'extractor.search_data.extractor.neo4j.{}'.format(Neo4jExtractor.NEO4J_AUTH_USER):
+                    'test-user',
+                'extractor.search_data.extractor.neo4j.{}'.format(Neo4jExtractor.NEO4J_AUTH_PW):
+                    'test-passwd',
+                'extractor.search_data.{}'.format(Neo4jSearchDataExtractor.ENTITY_TYPE):
+                    'dashboard',
+            })
+            extractor.init(Scoped.get_scoped_conf(conf=conf,
+                                                  scope=extractor.get_scope()))
+            self.assertEqual(extractor.cypher_query, Neo4jSearchDataExtractor
+                             .DEFAULT_NEO4J_DASHBOARD_CYPHER_QUERY.format(publish_tag_filter=''))
+
 
 if __name__ == '__main__':
     unittest.main()