ARS regression test

ars-regression.py

from neo4j import GraphDatabase
from jinja2 import Template
import concurrent.futures
import sys, json, threading, requests, random

# update this query accordingly
QUERY = """{
  "message": {
    "query_graph": {
      "nodes": {
        "n00": {
          "id": "{{ disease }}",
          "category": "biolink:Disease"
        },
        "n01": {
          "category": "biolink:ChemicalSubstance"
        }
      },
      "edges": {
        "e00": {
          "subject": "n00",
          "object": "n01"
        }
      }
    }
  }
}
"""
ars_url = "https://ars.transltr.io/ars/api/submit"
num_threads = 2
template = Template(QUERY)

uri = "bolt://disease.ncats.io:80"
driver = GraphDatabase.driver(uri, auth=("neo4j", ""))

def runquery(disease, timeout=1000):
    query = template.render(disease=disease)
    headers = { 'Content-Type': 'application/json' }
    print('searching for %s...' % disease, file=sys.stderr)
    r = requests.post(ars_url, data=query, headers=headers, timeout=timeout)
    print('%d: %s' % (r.status_code, disease), file=sys.stderr)
    return r


def run_cypher(tx, query):
    diseases = []
    for row in tx.run(query):
        diseases.append(row['MONDO'])
    return diseases

def get_monogenic_diseases():
    CYPHER="""
MATCH(n:S_MONDO)--(d:DATA) WHERE d.id='MONDO:0000275'
WITH n,d MATCH (n)<-[e:R_subClassOf*1..10]-(m:S_MONDO)<-[:PAYLOAD]-(z:DATA)
WITH m,z MATCH p=(d:DATA)-[:PAYLOAD]->(:S_GARD)-[:R_exactMatch|:R_equivalentClass]-(m)
OPTIONAL MATCH q=(m)-[:R_exactMatch|:R_equivalentClass]-(:S_ORDO_ORPHANET)-[e:R_rel{name:'disease_associated_with_gene'}]->(:S_ORDO_ORPHANET)<-[:PAYLOAD]-(o:DATA)
RETURN DISTINCT d.id as `GARD_ID`,d.name as `GARD_Disease`,z.notation as MONDO,e.DisorderGeneAssociationType as `Disease_Gene_Association`,o.symbol as `Gene_Symbol`,o.label as `Gene_Name`
ORDER BY d.id
"""
    with driver.session() as session:
        return session.read_transaction(run_cypher, CYPHER)

def dos_attack(vector, threads=2):
    with concurrent.futures.ThreadPoolExecutor(max_workers=threads) as executor:
        with open('ars-regression-results.txt', 'w') as f:        
            for disease, r in zip(vector, executor.map(runquery, vector)):
                if 201 == r.status_code:
                    mesg = r.json()
                    print ('%s\t%d\t%s' % (
                        disease, r.status_code, mesg['pk']), file=f)
                else:
                    print ('%s\t%d\t%s' % (
                        disease, r.status_code, r.text), file=f)
    
if __name__ == "__main__":
    if len(sys.argv) == 1:
        print('Usage: %s SIZE\nwhere SIZE is the number of diseases' % sys.argv[0])
        sys.exit(1)
    size = int(sys.argv[1])
    attack_vector = get_monogenic_diseases()
    random.shuffle(attack_vector)
    dos_attack(attack_vector[0:size])