Redirect NCBITaxon PURLs to official NCBI site

[jamesaoverton]

Resolves #1010 from obophenotype/ncbitaxon#117

I found this issue at the bottom of my inbox and decided to finally take care of it.

Note that there are non-numeric IDs for ranks and stuff in NCBITaxon, some using # fragments. I think that none of them ever resolved. Most were replaced by TAXRANK terms. If someone wants to do a forensic analysis of this, be my guest.

[matentzn]

🚀

[matentzn]

Actually there seem to be NEW pages now:

https://www.ncbi.nlm.nih.gov/datasets/taxonomy/9606/
https://www.ncbi.nlm.nih.gov/datasets/taxonomy/tree/?taxon=9606

Maybe worth considering?

[jamesaoverton]

Interesting. I'd never seen those before.

[lubianat]

just to mention that either look great & better than current :)

I tried to resolve a NCBITaxon term now and it took several minutes for the Ontobee page to resolve.

https://www.ncbi.nlm.nih.gov/datasets/taxonomy/9606/ looks really clean (both page and URI), so I'd vote for that

[jamesaoverton]

I would like to hear from @cmungall and @anitacaron before proceeding with this.

[anitacaron]

The new pages look great, but I've noticed that the definitions are not part of the set of files that we use to generate the OBO version; the information for the new page is coming from other files. The images are not the same either. I don't know whether this will affect the terms of the OBO release, like a discrepancy between the terms available or not.

[anitacaron]

What about this 'about' prefix?

- prefix: /about/
  replacement: http://www.ontobee.org/browser/rdf.php?o=NCBITaxon&iri=http://purl.obolibrary.org/obo/
  tests:
  - from: /about/NCBITaxon_1
    to: http://www.ontobee.org/browser/rdf.php?o=NCBITaxon&iri=http://purl.obolibrary.org/obo/NCBITaxon_1

[matentzn]

The new pages look great, but I've noticed that the definitions are not part of the set of files that we use to generate the OBO version; the information for the new page is coming from other files. The images are not the same either. I don't know whether this will affect the terms of the OBO release, like a discrepancy between the terms available or not.

Interesting @anitacaron - in this case, we should probably go with the reference to the old pages and do a thorough investigation first..

[matentzn]

Given @anitacaron observations of various discrepancies on the new site I think we need a more prinicpled investigation if we can link to that safely.

[jamesaoverton]

@anitacaron I figured that we would leave the /about/ unchanged.

What definitions are you referring to? Things like "Human (Homo sapiens) is a species of primate in the family Hominidae (great apes)." I bet those are generated automatically using a handful of patterns.

[anitacaron]

@jamesaoverton I think so too, but they're not sharing, at least I couldn't find the generated definition.

[jamesaoverton]

@anitacaron I think that you're saying that we should not use the new pages because they have automatically generated textual definitions, but I don't understand why that's a problem.

[anitacaron]

@jamesaoverton, the OBO release artefacts don't have the definitions shown on that page.

[jamesaoverton]

@anitacaron That's true. But why is it a problem?

[jamesaoverton]

I merged this. It changes NCBITaxon redirects from Ontobee to the old NCBI Taxonomy pages. I guess well change to the new NCBI Taxonomy pages at some point -- someone else can take care of that.