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
Describe the bug
Currently, the results of start_k are added at depth 0, while initial_root_ids the neighbors of initial_root_ids are treated as depth 0 (making the initial roots effectively depth -1). This leads to confusing explanations of how the traversals behave.
Expected behavior
Initial roots should be roots (depth 0) and returned.
Ideally, the docs for traverse and atraverse should also clarify that initial_root_ids and start_k are combined, therefore start_k=0 should be used if you only want initial roots. Possibly a warning too?
The text was updated successfully, but these errors were encountered:
I think this should be an option. I can see where both cases could be desirable.
When initial roots are desired to be part of the output, then they should be added at depth=0. When they are not, it should continue to work as it does now (depth==-1).
Describe the bug
Currently, the results of
start_k
are added at depth 0, whileinitial_root_ids
the neighbors ofinitial_root_ids
are treated as depth 0 (making the initial roots effectively depth -1). This leads to confusing explanations of how the traversals behave.To Reproduce
Read the docs at https://datastax.github.io/graph-rag/get-started/traversal/
Expected behavior
Initial roots should be roots (depth 0) and returned.
Ideally, the docs for
traverse
andatraverse
should also clarify thatinitial_root_ids
andstart_k
are combined, thereforestart_k=0
should be used if you only want initial roots. Possibly a warning too?The text was updated successfully, but these errors were encountered: