search

monarch-initiative / monarch-legacy

Monarch web application and API
BSD 3-Clause "New" or "Revised" License
42 stars 37 forks source link

Kallmann Syndrome link leads to 404 #984

Open harryhoch opened 8 years ago

harryhoch commented 8 years ago

Kallmann syndrome - http://beta.monarchinitiative.org/disease/MESH%3AD017436

Source link - http://www.nlm.nih.gov/cgi/mesh/2005/MB_cgi?mode=&term=D017436, leads to a 404.

jmcmurry commented 8 years ago

I don't see any one obvious way to fix this issue, but it is much broader than this specific ID D017436. There are implications for: a) what is the correct namespace for MESH in curie_map.yml, AND b) ensuring that the map is the source of all CURIE to URI expansion across the Monarch App, not just in dipper sourcecode.

1) If we expand the CURIE according to the current dipper yml mapping, we get http://purl.obolibrary.org/obo/MESH_D017436 which doesn't solve the problem; we get directed to a page with unexpected behavior: `This XML file does not appear to have any style information associated with it. The document tree is shown below.

Ontology not specified or not supported` 3) As Harry discovered, http://www.nlm.nih.gov/cgi/mesh/2005/MB_cgi?mode=&term=D017436 yields a 404. However if we use 2015 as the year instead of 2005 and a slightly different query syntax, we get the expected behavior. http://www.nlm.nih.gov/cgi/mesh/2015/MB_cgi?exact&field=uid&term=D017436 ![screen shot 2015-10-07 at 13 09 29](https://cloud.githubusercontent.com/assets/3658863/10349734/b9fae158-6cf4-11e5-9cfa-79e11261f686.png) 4) Using identifiers.org resolver (http://identifiers.org/mesh/D017436) we get redirected to a blank page which happens to live in the 2012 MESH namespace. ![screen shot 2015-10-07 at 13 04 56](https://cloud.githubusercontent.com/assets/3658863/10349598/229d57aa-6cf4-11e5-9740-6d186088ce02.png) So not sure what is going on there... As a stop gap, we could update the dipper curie map with the functioning resolution namespace for MeSH (http://www.nlm.nih.gov/cgi/mesh/2015/MB_cgi?exact&field=uid&term=), however, they might well change their API. Moreover, I'm not sure it will necessarily work with every term (if there is one that has been deprecated, for instance). In either case, I do think we need to update the source of the mappings to the curie_map.yml, that way we can fix issues like this in a single place rather than distributed across the application. Please advise.
harryhoch commented 8 years ago

So there would appear to be an issue here with presentation of the links. Autocomplete comes up with links for several other Kallmann variants, going to places like ontobee, OMIM, orphanet, etc. It might be good to try to change this link to make this variation a bit clearer.