You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently we instantiate QueryArgs in compute/accumulate methods in Cython and pass them through to C++. Defaults are determined by whatever defaults are specified inside these compute methods. We should have a more well-defined resolution order where defaults are specified and used consistently.
The best approach will be to store a class attribute dictionary for all such classes that contains default query arguments. If users provide default values, they are used, but must be complete. We should provide some level of query mode inference, so that users could provide query arguments that contain just r_max, for example, and have that automatically be a ball query. We should keep the mode argument available to support any future query modes that might lead to conflicts with this resolution, see #372 for example.
The text was updated successfully, but these errors were encountered:
Currently we instantiate QueryArgs in compute/accumulate methods in Cython and pass them through to C++. Defaults are determined by whatever defaults are specified inside these compute methods. We should have a more well-defined resolution order where defaults are specified and used consistently.
The best approach will be to store a class attribute dictionary for all such classes that contains default query arguments. If users provide default values, they are used, but must be complete. We should provide some level of query mode inference, so that users could provide query arguments that contain just
r_max
, for example, and have that automatically be a ball query. We should keep the mode argument available to support any future query modes that might lead to conflicts with this resolution, see #372 for example.The text was updated successfully, but these errors were encountered: