Maybe the -in/-child-of and -under/-descendant-of predicates could be overloaded to allow either a section path or a metadata filter expression (same as that in -has).
OR, new predicates could be defined: say, -parent-has and -ancestor-has. However, the filtering is based on the keybind not the section. Using two -in or -under predicates with -parent-has or -ancestor-has respectively wouldn't allow the metadata predicates to distinguish between the two -in or the two -under predicates.
As for implementation, maybe there could be a dictionary mapping section paths to the metadata of the section which the path represents.
Maybe the
-in/-child-ofand-under/-descendant-ofpredicates could be overloaded to allow either a section path or a metadata filter expression (same as that in-has).OR, new predicates could be defined: say,
-parent-hasand-ancestor-has. However, the filtering is based on the keybind not the section. Using two-inor-underpredicates with-parent-hasor-ancestor-hasrespectively wouldn't allow the metadata predicates to distinguish between the two-inor the two-underpredicates.As for implementation, maybe there could be a dictionary mapping section paths to the metadata of the section which the path represents.
Depends on: #11 Related: #14 #16