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
Is your feature request related to a problem? Please describe.
In WN-LMF, the targets of <SenseRelation> child elements of <Sense> elements can be either a Sense or a Synset. Wn imports these into separate tables and makes them available via separate methods: Sense.get_related() and Sense.get_related_synsets() (currently, Sense.relations() only works for relations with sense targets). It might make sense to merge these into Sense.get_related() with a parameter to choose the target type:
The value of target_type is a string (or a StrEnum) with values "sense", "synset", and "both". Initially, target_type will default to "sense" and Sense.get_related_synsets() will call Sense.get_related(..., target_type="synset"), but later we could change the default to "both" and deprecate Sense.get_related_synsets().
Additionally, other relation methods like Sense.relations() can use this parameter to fill a feature gap.
Describe the solution you'd like
See above.
Describe alternatives you've considered
We could continue to have two functions for each (.get_related_synsets(), .synset_relations(), etc.). This seems unnecessary.
Additional context
Along with planning for the new .relation_map(), this would prevent two new functions when one could suffice.
The text was updated successfully, but these errors were encountered:
@fcbond Can you remind me why we have such a thing as Sense–Synset relations in WN-LMF? I looked through all senses in OMW, all versions of EWN/OEWN, OWN-PT, and Odenet, and couldn't find a single instance:
>>> len(wn.senses())
3407588
>>> sum(len(s.get_related_synsets()) for s in wn.senses())
0
I'm guessing there is some wordnet (maybe the Polish one?) that links senses to synsets with domain_topic, domain_region, exemplifies, or other relations, but maybe the feature has always just been aspirational?
If we don't foresee an immediate need for Sense–Synset relations, I'd suggest we remove support for them from Wn and maybe from WN-LMF. Trying to accommodate them makes the code much more complicated than it needs to be.
Is your feature request related to a problem? Please describe.
In WN-LMF, the targets of
<SenseRelation>
child elements of<Sense>
elements can be either aSense
or aSynset
. Wn imports these into separate tables and makes them available via separate methods:Sense.get_related()
andSense.get_related_synsets()
(currently,Sense.relations()
only works for relations with sense targets). It might make sense to merge these intoSense.get_related()
with a parameter to choose the target type:The value of
target_type
is a string (or aStrEnum
) with values "sense", "synset", and "both". Initially,target_type
will default to"sense"
andSense.get_related_synsets()
will callSense.get_related(..., target_type="synset")
, but later we could change the default to"both"
and deprecateSense.get_related_synsets()
.Additionally, other relation methods like
Sense.relations()
can use this parameter to fill a feature gap.Describe the solution you'd like
See above.
Describe alternatives you've considered
We could continue to have two functions for each (
.get_related_synsets()
,.synset_relations()
, etc.). This seems unnecessary.Additional context
Along with planning for the new
.relation_map()
, this would prevent two new functions when one could suffice.The text was updated successfully, but these errors were encountered: