Set ssl and sslverifcation to catalog or db credentials

[Soufraz]

Description

Connect into a trino/presto db using catalog and credential file.
In datagrip and dbeaver I can connect only with two params:
SSL=True
SSLVerification=NONE
I am trying to find out how to fill this params in catalog definition or credentials file.
When I add to con url the pipeline not even start to run.

Context

Id like to get data using catalogs through trino/presto. I installed sqlalchemy-trino and change the con url to trino://
But to connect in database I need to set ssl params even in datagrip or dbeaver.
Then I am trying to find out how to pass this params in credentials con.
I tried ssl, ssl_context, ssl_verification and a few more but any param that I add I get an error of not recognized params.

kedro.io.core.DataSetError: Failed while loading data from data set SQLQueryDataSet(load_args={}, sql=BIG_QUERY_HERE.
HTTPSConnectionPool(host='presto.address.com', port=8446): Max retries exceeded with url: /v1/statement (Caused by ConnectTimeoutError(<urllib3.connection.HTTPSConnection object at 0x7f183810d520>, 'Connection to presto.address.com timed out. (connect timeout=30.0)'))

Your Environment

sqlalchemy==1.4.15
sqlalchemy-trino==0.3.0
psycopg2-binary

• Kedro version used: 0.17.3
• Python version used: 3.8.5
• Operating system and version: Manjaro 21

[merelcht]

Hi @Soufraz, thanks for reaching out!
Could you paste the catalog entry you tried?

[Soufraz]

# credentials
trino_connection:
  con: trino://user:pass@presto.address.com:8446/hive

# catalog
attachments:
  type: pandas.SQLQueryDataSet
  credentials: trino_connection
  sql: 'query here'

Params Ive tried:

load_args:
  ssl:
    fake_flag_to_enable_tls: True
    verification: None

[merelcht]

Under the hood we use pandas read_sql_query to pass on the load_args. You can find here which arguments are supported: https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.read_sql_query.html

[Soufraz]

I cant only connect in this way on trino and I didnt found how to pass things like _http_session.verify and auth in kedro catalogs.
Is there a way to create this connection in code and use as a connection in some catalog?

from trino import dbapi, auth
conn = dbapi.connect(
    host=host,
    port=port,
    user=username,
    catalog=dialect,
    schema=database_name,
    http_scheme='https',
    auth=auth.BasicAuthentication(username, password),
)
conn._http_session.verify = False

[Soufraz]

For now I built a custom dataset. Thanks.

[datajoely]

@Soufraz we're always happy to accept PRs if you'd like to contribute a sql.PrestDataSet?

[Soufraz]

I can try. Hahaha. I'll send until the end of this week.