utils.paths module for query_paths endpoint

Merges #33
Refs #31

Author: dhimmel

dj_hetmech/settings.py

@@ -20,7 +20,7 @@
 # Read secrets from YAML file
 path = os.path.join(BASE_DIR, 'dj_hetmech', 'secrets.yml')
 with open(path) as read_file:
-    secrets = yaml.load(read_file)
+    secrets = yaml.full_load(read_file)
 
 # Quick-start development settings - unsuitable for production
 # See https://docs.djangoproject.com/en/2.1/howto/deployment/checklist/

dj_hetmech_app/management/commands/path_query.py

@@ -0,0 +1,28 @@
+import json
+
+import hetio.readwrite
+from django.core.management.base import BaseCommand
+
+from dj_hetmech_app.utils.paths import (
+    get_neo4j_node_info,
+    get_neo4j_rel_info,
+    get_paths,
+)
+
+
+class Command(BaseCommand):
+
+    help = 'Call dj_hetmech_app.utils.paths.get_paths for prototyping purposes.'
+
+    def handle(self, *args, **options):
+        metagraph_url = 'https://github.com/hetio/hetionet/raw/727411d4796ba2858bc8e684ba407047c1615390/hetnet/json/hetionet-v1.0-metagraph.json'
+        metagraph = hetio.readwrite.read_metagraph(metagraph_url)
+        json_obj = get_paths(
+            metagraph=metagraph,
+            metapath='CbGiGaD',
+            source_identifier='DB01156',  # Bupropion
+            target_identifier='DOID:0050742',  # nicotine dependency
+            limit=100,
+        )
+        json_str = json.dumps(json_obj, indent=2)
+        print(json_str)

dj_hetmech_app/utils.py dj_hetmech_app/utils/__init__.py

@@ -231,3 +231,10 @@ def get_hetionet_metagraph():
 def metapath_from_abbrev(abbreviation):
     metagraph = get_hetionet_metagraph()
     return metagraph.metapath_from_abbrev(abbreviation)
+
+
+@functools.lru_cache()
+def get_neo4j_driver():
+    from neo4j import GraphDatabase
+    driver = GraphDatabase.driver('bolt://neo4j.het.io')
+    return driver

dj_hetmech_app/utils/paths.py

@@ -0,0 +1,107 @@
+import json
+import functools
+
+import hetio.neo4j
+
+from dj_hetmech_app.utils import (
+    metapath_from_abbrev,
+    get_neo4j_driver,
+)
+
+
+def get_paths(metagraph, metapath, source_identifier, target_identifier, limit=None):
+    """
+    Return JSON-serializable object with paths between two nodes for a given metapath.
+    """
+    metapath = metagraph.get_metapath(metapath)
+    query = hetio.neo4j.construct_pdp_query(metapath, property='identifier', path_style='id_lists')
+    if limit is not None:
+        query += f'\nLIMIT {limit}'
+    driver = get_neo4j_driver()
+    neo4j_params = {
+        'source': source_identifier, 
+        'target': target_identifier,
+        'w': 0.5,
+    }
+    with driver.session() as session:
+        results = session.run(query, neo4j_params)
+        results = [dict(record) for record in results]
+    neo4j_node_ids = set()
+    neo4j_rel_ids = set()
+    paths_obj = []
+    for record in results:
+        row = {
+            'metapath': metapath.abbrev,
+        }
+        row.update(record)
+        neo4j_node_ids.update(row['node_ids'])
+        neo4j_rel_ids.update(row['rel_ids'])
+        paths_obj.append(row)
+
+    node_id_to_info = get_neo4j_node_info(neo4j_node_ids)
+    rel_id_to_info = get_neo4j_rel_info(neo4j_rel_ids)
+    json_obj = {
+        'query': {
+            'source': source_identifier,
+            'target': target_identifier,
+            'metapath': metapath.abbrev,
+            'metapath_id': [edge.get_id() for edge in metapath],
+            'limit': limit,
+        },
+        'paths': paths_obj,
+        'nodes': node_id_to_info,
+        'relationships': rel_id_to_info,
+    }
+    return json_obj
+
+
+cypher_node_query = '''\
+MATCH (node)
+WHERE id(node) IN $node_ids
+RETURN
+  id(node) AS neo4j_id,
+  node.identifier AS identifier,
+  head(labels(node)) AS node_label,
+  properties(node) AS data
+ORDER BY neo4j_id
+'''
+
+
+def get_neo4j_node_info(node_ids):
+    """
+    Return information on nodes corresponding to the input neo4j node ids.
+    """
+    node_ids = sorted(node_ids)
+    driver = get_neo4j_driver()
+    with driver.session() as session:
+        results = session.run(cypher_node_query, node_ids=node_ids)
+        results = [dict(record) for record in results]
+    id_to_info = {x['neo4j_id']: x for x in results}
+    return id_to_info
+
+
+cypher_rel_query = '''\
+MATCH ()-[rel]-()
+WHERE id(rel) in $rel_ids
+RETURN
+  id(rel) AS neo4j_id,
+  type(rel) AS rel_type,
+  id(startNode(rel)) AS source_neo4j_id,
+  id(endNode(rel)) AS target_neo4j_id,
+  properties(rel) AS data
+ORDER BY neo4j_id
+'''
+
+
+def get_neo4j_rel_info(rel_ids):
+    """
+    Return information on relationships corresponding to the
+    input neo4j relationship ids.
+    """
+    rel_ids = sorted(rel_ids)
+    driver = get_neo4j_driver()
+    with driver.session() as session:
+        results = session.run(cypher_rel_query, rel_ids=rel_ids)
+        results = [dict(record) for record in results]
+    id_to_info = {x['neo4j_id']: x for x in results}
+    return id_to_info