Protein count on Condensate List Page and Detal Page doesnt match

Pointed out by Anna, for "Paraspeckle (9606)"; the protein count on the list page shows "65", whereas on the detail page is "69"

So, count 69 is right. The size of the list of proteins in Paraspeckle (9606) is 69 and not 65.

I checked the API call made by the frontend and it is {{host}}/condensates?size=100000&is_experimental=false

The protein list size here is also 69 and the attribute "protein_count" is also 69 in the API response.

Also, a secondary comment:

This call doesn't need to fetch all the fields, if you ask for selected fields the response could be lesser in size (don't need protein_confidence_score, protein_pubmed_ids etc here)

The protein_count is 69 but the actual size of protein list is 65. Could you check why this discrepancy happened? Doesn't it have to be same as the actual protein(proteins) list length?

Also, a secondary comment:

This call doesn't need to fetch all the fields, if you ask for selected fields the response could be lesser in size (don't need protein_confidence_score, protein_pubmed_ids etc here)

Sure, I will put the fields too.

The size of proteins list is 69 but the size of protein_pubmed_ids dict is 65 actually

The reason of this is to list unique gene names from 69 proteins. Can you check whether you still keep the redundant duplicates in your protein list or not?

https://git.mpi-cbg.de/dd-code-team/dd-code-api/-/blob/develop/src/models/condensates.py#L50-53

        serialized_record['proteins'] = ProteinModel().fetch_by_unique_keys(
            unique_key_values=serialized_record.get('proteins'),
            fields=['gene_name', 'name']
        )

Sorry, I didn't get your point? No, the lists don't have any duplicates. I also re-checked this specific protein list.

The function fetch_by_unique_keys fetches multiple records based on the unique ID values passed for the model. Docstring says that https://git.mpi-cbg.de/dd-code-team/dd-code-api/-/blob/develop/src/models/model.py#L12-19

Fetch multiple records by unique key identifier.

Now I understand it. There are proteins having null for gene_name and the list skipped those proteins. protein_count will be used instead. Thanks!

mentioned in commit 065cf3ca

closed

This bug is still showing up on Dev server (after hard refresh also). And I noted it in some other cases also for synthetic condensates.

Was this deployed on Dev?

The field uses protein_count and the server api returns 69.

Yes that is correct and the list page still shows "65"

It's done.