Chuck Hastings

This PR defines a uniform random walk implementation using the neighborhood sampling functions. This will be refactored once the new sampling primitive (https://github.com/rapidsai/cugraph/issues/2580) is implemented, but should provide a stronger...

Use neighborhood sampling as the algorithm for a round of performance tuning on this primitive. Blocked by https://github.com/rapidsai/cugraph/issues/2581

Update the Random Walks implementation to use the new selection primitive. Blocked by https://github.com/rapidsai/cugraph/issues/2580

Modify Neighborhood sampling implementation to use the new selection primitive. Blocked by https://github.com/rapidsai/cugraph/issues/2580

Create Selection primitive

1

Create new Selection primitive for use by Neighbor sampling and Random walks.

Provide support in python for creating an edge with both edge id and weight. Blocked by https://github.com/rapidsai/cugraph/issues/2576 to start work. Blocked by https://github.com/rapidsai/cugraph/issues/2575 to finish work.

Create C++ API for creating graph with multiple edge properties. Only edge id and weight will be supported initially, but the C++ API could include support for more complex structures.

Add support for both edge id and weight in C API for graph creation. Support for more complex edge properties will be addressed later.

Update pylibcugraph interface for graph creation to support edge id and weight. More extensive property support will be considered later.

Initial focus will be adding edge id. Some functionality will exist to support arbitrary edge property structures, but fully implementing will require restructuring code to be header only. Full edge...