In my following block of code, sometime a DataFrame is created, sometime it is not.
It seems like the issue is due to the fact that neo4j is racing to execute the command. I have tried to split the run commands into multiple session like the code below, I have also tried to include all of them into 1 session. Yet the consistency of the number of time something is return is still low. How can I stop this eager operation?
set_label_query = """
MATCH (s:Startup)
WHERE "{vertical_original}" IN s.Verticals
WITH s
MATCH (s)<-[:INVESTOR_INVESTED_IN]-(i:Investor)
WITH s, i
MATCH(i)<-[:MADE_LP_COMMITMENT_TO_VC]-(l:Limited_Partner)
SET i:{vertical}, l:{vertical}, s:{vertical}
RETURN COUNT(i)
;
"""
create_gds_project_query = '''
CALL gds.graph.project(
'climate_cleantech_undirected',
['{vertical}', 'Limited_Partner', 'Investor', 'Startup'],
{{INVESTOR_INVESTED_IN: {{orientation: 'UNDIRECTED'}},
MADE_LP_COMMITMENT_TO_VC: {{orientation: 'UNDIRECTED'}}
}}
);
'''
create_rank_query = '''
CALL gds.pageRank.stream('climate_cleantech_undirected', {{
nodeLabels:['{vertical}'] ,
maxIterations: 20,
dampingFactor: 0.85
}})
YIELD nodeId, score
WITH gds.util.asNode(nodeId) AS node, score
WHERE 'Investor' IN labels(node)
RETURN node.Name, node.Website, score
ORDER BY score DESC;
'''
remove_graph_query = "CALL gds.graph.drop('climate_cleantech_undirected', false)"
with neo4j_driver.session() as session:
with session.begin_transaction() as tx:
tx.run(set_label_query.format(vertical=vertical))
tx.run(create_gds_project_query.format(vertical=vertical))
result_data = tx.run(create_rank_query.format(vertical=vertical)).data()
df = pd.DataFrame(result_data)
print(df)
tx.commit()
print('execute 2')
with neo4j_driver.session() as session:
with session.begin_transaction() as tx:
tx.run(remove_label_query.format(vertical=vertical))
tx.run(remove_graph_query)
tx.commit()
result when something is return:
