Open yasu-sh opened 1 year ago
I'm guessing thisI will be more consistent in R for py-tetrad. Several bugs were fixed concerning the handling of knowledge.
I'm guessing thisI will be more consistent in R for py-tetrad. Several bugs were fixed concerning the handling of knowledge.
Sure. This issue would become one of good reasons to switch from r-causal to py-tetrad.
This should be consistent with the new r-tetrad (part of py-tetrad). Many knowledge bugs were fixed, and many bugs for various algorithms. The downside is that r-causal is still relatively new; we are revising the API of Tetrad and adding documentation for Python and R users, so there may be some changes as we clarify the organization of that code. But we are quickly converging to a solution and will keep the py-tetrad and r-tetrad code current.
I should say on a personal note that the only one of our immediate team supporting Tetrad in R right now is me, and trying to bring r-causal up to date at the same time as I'm bringing r-tetrad up to date is more than I can handle right now. However, if someone wanted to step in and help with it, that might be helpful. There are more recent tetrad jars, new version of causal-command, and they would have to be updated and any out of data code in r-causal would need to be adjusted for the new versions.
This is a small note for known issue:
Checked Java's internal object after setting prior knowledge in r-causal. Looks omitted 1 of 2 Forbidden Directed Edge: Between column1 And column12
I guess indicated message"Between / And" is invalid and "From / To" is good for understanding. Reason: the consistency of TETRAD GUI behavior.