search

monarch-initiative / phenogrid

The phenogrid widget
13 stars 14 forks source link

CORS error causes phenogrid to churn with no result #125

Closed jmcmurry closed 9 years ago

jmcmurry commented 9 years ago

On Mac OSX Mavericks and Chrome

Looking at the source of the problem, I get a 413 (Full head) The first item in the list is a CORS one: J.ajaxTransport.G.cors.t.crossDomain.send Please advise; does this have anything to do with having changed the SciGraph URLs in configuration this afternoon?

screen shot 2015-07-29 at 23 03 19

For thoroughness, I should also mention that the same page doesn't load at all on stage, (http://stage.monarchinitiative.org/disease/DOID%3A5723 ) I'm guessing that is expected at present, but please confirm.

harryhoch commented 9 years ago

Interesting. 413 HTTP code is a "Request Entity" too large. This page has 663 phenotypes. Ouch. Any thoughts on what we can do here to handle this sort of request? Is Ringo at all configurable? @cmungall ,@kltm

yuanzhou commented 9 years ago

The current version still uses HTTP GET to send out those phenotypes, 663 phenotypes may cause the URL too long. Kent has changed it to POST, but I haven't been able to merge his PR due to some local testing errors.

cmungall commented 9 years ago

Well it's really 663 phenotype associations, the same phenotype may be there twice if it's associated with different leaf diseases. I assume they are being uniqified before sent?

On 30 Jul 2015, at 4:49, Harry Hochheiser wrote:

Interesting. 413 HTTP code is a "Request Entity" too large. This page has 663 phenotypes. Ouch. Any thoughts on what we can do here to handle this sort of request? Is Ringo at all configurable? @cmungall ,@kltm

Reply to this email directly or view it on GitHub: https://github.com/monarch-initiative/phenogrid/issues/125#issuecomment-126286713

harryhoch commented 9 years ago

@cmungall, Phenogrid does no filtering of input phenotypes before sending the response to /simsearch. We could add a filter in phenogrid, or in the /simsearch call - where's the best spot?

yuanzhou commented 9 years ago

One feedback from today's meeting about filtering the phenotype input is to check duplicated phenotype IDs.

yuanzhou commented 9 years ago

Issue fixed after using the ajax POST. @harryhoch

yuanzhou commented 9 years ago

Already added check to remove any duplicated phenotypes before sending them out to simsearch.