Wrong list of parents for a numeric variable
Issue #333
resolved
In the network ID-decide-test.pgmx, if I open the potential for "Health state" and click "Reorder variables", the list should only include the parents of this node, but it also includes, by mistake, the child variable.
If I change the potential type from "Exact" to "UnivariateDistr" and click "Reorder variables", the list contains "pseudoVariableDistributionName", which should not be there.
An undesirable consequence of this mistake is that reordering the variables leads to missing some of them.
Comments (3)
-
-
This commit is aimed at showing the correct list of variables to reorder. It is still pending to make the values (and the list of variables) appear reordered correctly in the potential.
-
- changed status to resolved
This commit fixes this issue.
- Log in to comment
The current implementation of UnivariateDistr was in alpha state when it was decided to discontinue it and replace it with a new implementation of the tables (@Manuel Luque ). As far as I know, while that implementation is being carried out, Exact potential should be used instead (although “Reorder variables“ does not work with Exact either).
Shall I fix UnivariateDistr or Exact?. Both will be replaced anyway, but I think Exact covers the current necessities of OpenMarkov, is more evolved and its layout is cleaner.