Rating summary is asking for Rating entities for implicit feedback data about lenskit

[ishank-zefo]

I changed my config to
bind ItemScorer to ItemItemScorer within (UserVectorNormalizer) { bind VectorNormalizer to UnitVectorNormalizer } bind RatingVectorPDAO to EntityCountRatingVectorPDAO set InteractionEntityType to EntityType.forName("PAGEVIEW") bind (BaselineScorer, ItemScorer) to UserMeanItemScorer bind (UserMeanBaseline, ItemScorer) to ItemMeanRatingItemScorer

So this particular issue is solved but during recommendation but still at the time of creating recommendations following query gets executed
dao.query(CommonTypes.RATING) .withAttribute(CommonAttributes.USER_ID, user) .valueSet(CommonAttributes.ITEM_ID);
in TopNItemRecommender.java which is again causing issues.

[mdekstrand]

Thanks for raising this - I apologize for not noticing it earlier.

TopNRecommender does need to be updated.

[mdekstrand]

The solution here is to make TopNItemRecommender use InteractionEntityType. We also need a test for this.

[memmm]

Is there a possible configuration for implicit data that doesn't make use of TopNItemRecommender? Or can a fix for it be expected?

[mdekstrand]

You can implement your own ItemRecommender, if you want.

Future LensKit development is in LKPY.