potiuk commented on a change in pull request #12466:
URL: https://github.com/apache/airflow/pull/12466#discussion_r527275948
##########
File path: airflow/providers/apache/cassandra/provider.yaml
##########
@@ -33,10 +35,15 @@ integrations:
sensors:
- integration-name: Apache Cassandra
python-modules:
- - airflow.providers.apache.cassandra.sensors.record
- - airflow.providers.apache.cassandra.sensors.table
+ - .sensors.record
+ - .sensors.table
hooks:
- integration-name: Apache Cassandra
python-modules:
- - airflow.providers.apache.cassandra.hooks.cassandra
+ - .hooks.cassandra
+
+connections:
+ - connection-type: cassandra
+ hook-class: .hooks.cassandra.CassandraHook
+ connection-id-parameter-name: cassandra_conn_id
Review comment:
There are many other things that are not yet defined here (other PRs are
addressing it) - extra links and connection in forms in javascript. If you look
at those, you will understand that those cannot be as easily incorporated in
the classes.
I prefer to keep "provider delivered functionality" in one place rather than
some of it in .yaml, some of this in some connection and some of this in other
classes.
Also, our proposal is far from complete - it does not actually solve the
problem of discovering those Hooks (unless you have not posted it. Even with
the entry_points mechanism, you will have to somehow at least keep the list of
hooks in each provider (there are at least 4 in the google provider). That
would mean that you would have to not only keep the list of hooks but also
remember to update the "extra" information in the hooks - this will have to be
(by its nature) split between two places.
What was your idea here? where did you want to keep the list of connections
to load?
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
