Papers displayed incorrectly in the results table at /phenotypes/

[abarysh]

Example: https://yeastphenome.org/phenotypes/?query=sulfur
Papers are displayed as: "Eide DJ; Yu D" instead of first~last, year.

[vsoch]

I'll work on this next once we finish with #130. I'll first try updating the listing of the phenotypes papers to try and be more like conditions. If for some reason that doesn't work, we can consider adding a new field to use consistently across queries (as discussed in #119).

[vsoch]

I've done some testing and I've reproduced why I changed the query to be as it is now. To recap, the "working" query for conditions (done with quotes because we haven't caught an issue yet, which might still be there) looks like this:

        return (
            apps.get_model("papers", "Paper")
            .objects.filter(
                Q(dataset__conditionset=self) | Q(dataset__control_conditionset=self)
            )
            .exclude(latest_data_status__status__name="not relevant")
            .distinct()
        )

The equivalent query for an observable looks like this, and note that this was being used for a short time:

    def papers(self):
        return (
            apps.get_model("papers", "Paper")
            .objects.filter(dataset__phenotype__observable=self)
            .exclude(latest_data_status__status__name="not relevant")
            .distinct()
            .order_by("first_author")
        )

However, when using this function, if you look at the sorting, it is off - note that a J is coming between D and C (this is what I saw, and why I changed it to be as it is now):

And if we remove the order_by it's just totally wrong (nothing seems to be sorted). The sorting strategy is basically identical, here for papers:

            agg_field = "phenotype__dataset__paper__first_author"
            queryset = queryset.annotate(
                paper_list=StringAgg(
                    agg_field, delimiter="; ", distinct=True, ordering=agg_field
                )
            )

and conditions:

            agg_field = "condition__conditionset__dataset__paper__first_author"
            queryset = queryset.annotate(
                paper_list=StringAgg(
                    agg_field, delimiter="; ", distinct=True, ordering=agg_field
                )
            )

So we have two options I think. I can do one or the other above, depending on which you think is less wrong - the current approach doesn't have the last name and year, and the second approach above has sorting which is slightly off. The second option would be to create a field in papers that has the first, last, and year of the author, and just sort by and then return that. If there is some issue with doing a sort on the aggregated field, then we might still see an issue here, I'm not sure. Let me know your thoughts!

[abarysh]

Ok, I think we should go ahead with a dedicated field in papers. It will simplify many things down the road. Could you create an issue for this and add it to the queue? It should be a fairly simple change: new field (e.g., systematic_name), read-only for the admin user, populated automatically at "Save" (just like the systematic names in datasets and conditionsets).

[vsoch]

I don't think conditions is sorting correctly either - here is an example - see that A comes after B and H.