Open andrewsu opened 1 year ago
Hopefully, with some of my upcoming optimizations, these limits can be raised substantially. I do think it makes sense to return a different result code for cases of the limit being hit.
I'm of the opinion that we'd rather get to a point where we can remove these limits altogether, if possible.
Here are two queries I ran based on the drug repurposing session at the Translator Relay in Feb 2023:
angiosarcoma -> [gene] -> [ChemicalEntity] (positive control: pembrolizumab) https://arax.ncats.io/?r=cad9aca8-bd39-4aa8-9c0a-0b55a7283578
Castleman disease -> [gene] -> [ChemicalEntity] (positive control: mTOR inhibitors like sirolimus) https://arax.ncats.io/?r=d26fb548-1bd5-4477-968a-602d78e479ef
In both queries, we return zero results with the following error:
QEdge undefined obtained xxx records, totalling xx for this query. This exceeds the per-query maximum of 60000. For stability purposes, this query is terminated. Please consider refining your query further.
Just want to do a quick evaluation later whether these limits are reasonable, whether partial results would have been useful, whether we should return a different result code (besides 200), etc.