Derek furst/remove offending fields

src/app_neo4j_queries.py

@@ -574,13 +574,25 @@ def get_provenance(neo4j_driver, uuid, depth):
     max_level_str = ''
     if depth is not None and len(str(depth)) > 0:
         max_level_str = f"maxLevel: {depth}, "
-
+    delimiter = ', '
+    entity_properties = ['group_uuid', 'created_by_user_displayname', 'created_by_user_sub', 'created_by_user_email', 'uuid', 'dataset_type', 'hubmap_id', 'entity_type', 'status', 'created_timestamp']
+    activity_properties = ['hubmap_id', 'created_by_user_displayname', 'created_by_user_sub', 'created_by_user_email', 'creation_action', 'uuid', 'status', 'created_timestamp']
+    entity_cypher_pairs_list = [f'{prop}: node.{prop}' for prop in entity_properties]
+    activity_cypher_pairs_list = [f'{prop}: node.{prop}' for prop in activity_properties]
+    entity_cypher_pairs_string = delimiter.join(entity_cypher_pairs_list)
+    activity_cypher_pairs_string = delimiter.join(activity_cypher_pairs_list)
     # More info on apoc.path.subgraphAll() procedure: https://neo4j.com/labs/apoc/4.0/graph-querying/expand-subgraph/
     query = (f"MATCH (n:Entity) "
              f"WHERE n.uuid = '{uuid}' "
              f"CALL apoc.path.subgraphAll(n, {{ {max_level_str} relationshipFilter:'<ACTIVITY_INPUT|<ACTIVITY_OUTPUT', labelFilter:'-Lab' }}) "
              f"YIELD nodes, relationships "
-             f"WITH [node in nodes | node {{ .*, label:labels(node)[0] }} ] as nodes, "
+             f"WITH [node in nodes | "
+             f"  CASE "
+             f"    WHEN 'Activity' IN labels(node) THEN node {{{activity_cypher_pairs_string} ,  label: labels(node)[0] }} "
+             f"    WHEN 'Entity' IN labels(node) THEN node {{{entity_cypher_pairs_string} , label: labels(node)[0] }} "
+             f"    ELSE node {{ label: labels(node)[0] }} "
+             f"  END "
+             f"] as nodes, "
              f"[rel in relationships | rel {{ .*, fromNode: {{ label:labels(startNode(rel))[0], uuid:startNode(rel).uuid }}, toNode: {{ label:labels(endNode(rel))[0], uuid:endNode(rel).uuid }}, rel_data: {{ type: type(rel) }} }} ] as rels "
              f"WITH {{ nodes:nodes, relationships:rels }} as json "
              f"RETURN json")

src/requirements.txt

@@ -1,6 +1,6 @@
 Flask==3.0.3
 neo4j==5.20.0
-prov==2.0.0
+prov==2.0.1
 Werkzeug==3.0.3
 deepdiff==7.0.1
 

src/schema/schema_constants.py

@@ -22,7 +22,7 @@ class SchemaConstants(object):
 
     DOI_BASE_URL = 'https://doi.org/'
 
-    DATASETS_OMITTED_FIELDS = ['ingest_metadata', 'metadata', 'files']
+    OMITTED_FIELDS = ['ingest_metadata', 'files']
 
 # Define an enumeration to classify an entity's visibility, which can be combined with
 # authorization info when verify operations on a request.

src/schema/schema_neo4j_queries.py

@@ -164,7 +164,7 @@ def filter_ancestors_by_type(neo4j_driver, direct_ancestor_uuids, entity_type):
 """
 def get_children(neo4j_driver, uuid, property_key = None):
     results = []
-
+    fields_to_omit = SchemaConstants.OMITTED_FIELDS
     if property_key:
         query = (f"MATCH (e:Entity)-[:ACTIVITY_INPUT]->(:Activity)-[:ACTIVITY_OUTPUT]->(child:Entity) "
                  # The target entity can't be a Lab
@@ -193,7 +193,10 @@ def get_children(neo4j_driver, uuid, property_key = None):
             else:
                 # Convert the list of nodes to a list of dicts
                 results = nodes_to_dicts(record[record_field_name])
-
+                if fields_to_omit:
+                    for node_dict in results:
+                        for field in fields_to_omit:
+                            node_dict.pop(field, None)
     return results
 
 
@@ -216,7 +219,7 @@ def get_children(neo4j_driver, uuid, property_key = None):
 """
 def get_parents(neo4j_driver, uuid, property_key = None):
     results = []
-
+    fields_to_omit = SchemaConstants.OMITTED_FIELDS
     if property_key:
         query = (f"MATCH (e:Entity)<-[:ACTIVITY_OUTPUT]-(:Activity)<-[:ACTIVITY_INPUT]-(parent:Entity) "
                  # Filter out the Lab entities
@@ -245,6 +248,10 @@ def get_parents(neo4j_driver, uuid, property_key = None):
             else:
                 # Convert the list of nodes to a list of dicts
                 results = nodes_to_dicts(record[record_field_name])
+                if fields_to_omit:
+                    for node_dict in results:
+                        for field in fields_to_omit:
+                            node_dict.pop(field, None)
 
     return results
 
@@ -380,7 +387,7 @@ def get_tuplets(neo4j_driver, uuid, property_key=None):
 """
 def get_ancestors(neo4j_driver, uuid, property_key = None):
     results = []
-
+    fields_to_omit = SchemaConstants.OMITTED_FIELDS
     if property_key:
         query = (f"MATCH (e:Entity)<-[:ACTIVITY_INPUT|ACTIVITY_OUTPUT*]-(ancestor:Entity) "
                  # Filter out the Lab entities
@@ -410,6 +417,11 @@ def get_ancestors(neo4j_driver, uuid, property_key = None):
                 # Convert the list of nodes to a list of dicts
                 results = nodes_to_dicts(record[record_field_name])
 
+                if fields_to_omit:
+                    for node_dict in results:
+                        for field in fields_to_omit:
+                            node_dict.pop(field, None)
+
     return results
 
 """
@@ -431,7 +443,7 @@ def get_ancestors(neo4j_driver, uuid, property_key = None):
 """
 def get_descendants(neo4j_driver, uuid, property_key = None):
     results = []
-
+    fields_to_omit = SchemaConstants.OMITTED_FIELDS
     if property_key:
         query = (f"MATCH (e:Entity)-[:ACTIVITY_INPUT|ACTIVITY_OUTPUT*]->(descendant:Entity) "
                  # The target entity can't be a Lab
@@ -460,7 +472,10 @@ def get_descendants(neo4j_driver, uuid, property_key = None):
             else:
                 # Convert the list of nodes to a list of dicts
                 results = nodes_to_dicts(record[record_field_name])
-
+                if fields_to_omit:
+                    for node_dict in results:
+                        for field in fields_to_omit:
+                            node_dict.pop(field, None)
     return results
 
 
@@ -1185,7 +1200,7 @@ def get_dataset_upload(neo4j_driver, uuid):
 def get_collection_datasets(neo4j_driver, uuid):
     results = []
 
-    fields_to_omit = SchemaConstants.DATASETS_OMITTED_FIELDS
+    fields_to_omit = SchemaConstants.OMITTED_FIELDS
     query = (f"MATCH (e:Dataset)-[:IN_COLLECTION]->(c:Collection) "
              f"WHERE c.uuid = '{uuid}' "
              f"RETURN COLLECT(apoc.create.vNode(labels(e), apoc.map.removeKeys(properties(e), {fields_to_omit}))) AS {record_field_name}")
@@ -1391,7 +1406,7 @@ def unlink_datasets_from_upload(neo4j_driver, upload_uuid, dataset_uuids_list):
 """
 def get_upload_datasets(neo4j_driver, uuid, property_key = None):
     results = []
-    fields_to_omit = SchemaConstants.DATASETS_OMITTED_FIELDS
+    fields_to_omit = SchemaConstants.OMITTED_FIELDS
     if property_key:
         query = (f"MATCH (e:Dataset)-[:IN_UPLOAD]->(s:Upload) "
                  f"WHERE s.uuid = '{uuid}' "