Fix ping service

[jmcmurry]

todo	status	pattern	failing URI
make note	doesn't seem to aspire to being resolvable in bioportal	http://purl.bioontology.org/ontology/SBO/SBO:$id	http://purl.bioontology.org/ontology/SBO/SBO:0000262
make note	Doesn't seem to aspire to being resolvable in bioportal	http://purl.bioontology.org/ontology/MA/MA:$id	http://purl.bioontology.org/ontology/MA/MA:0002502
track down error	This is the wrong pattern for amigo; where is it from? Should be http://amigo2.berkeleybop.org/amigo/term/GO:$id	http://amigo2.berkeleybop.org/cgi-bin/amigo2/amigo/term/GO:$id	http://amigo2.berkeleybop.org/cgi-bin/amigo2/amigo/term/GO:0006915
track down error	I think this is an isDeprecated issue	http://www.obofoundry.org/ro/#OBO_REL:is_a$id	http://www.obofoundry.org/ro/#OBO_REL:is_a
try again	This is I believe correct but server temporarily down; fix timeline unknown	http://compbio.charite.de/hpoweb/showterm?id=HP:$id	http://compbio.charite.de/hpoweb/showterm?id=HP:0000118
try again	seems to be working now; perhaps was transient outage?	http://pir.georgetown.edu/cgi-bin/pro/entry_pro?id=PR:$id	http://pir.georgetown.edu/cgi-bin/pro/entry_pro?id=PR:000000024

[putmantime]

I failed to apply the same obsolete=True filter on the ping module that I was using to create the list of agents. I have now applied that an both RO and the deprecated amigo2 services are no longer being pinged, and no longer in the list..

[jmcmurry]

Great. The unresolvable purls in bioportal are the only real sticky wickets then.

[putmantime]

Here is the updated list; still not getting a response from HP and PR.

[
  {
    "URIexample": "http://purl.bioontology.org/ontology/MA/MA:0002502",
    "URIpattern": "http://purl.bioontology.org/ontology/MA/MA:$id"
  },
  {
    "URIexample": "http://purl.bioontology.org/ontology/SBO/SBO:0000262",
    "URIpattern": "http://purl.bioontology.org/ontology/SBO/SBO:$id"
  },
  {
    "URIexample": "http://pir.georgetown.edu/cgi-bin/pro/entry_pro?id=PR:000000024",
    "URIpattern": "http://pir.georgetown.edu/cgi-bin/pro/entry_pro?id=PR:$id"
  },
  {
    "URIexample": "http://compbio.charite.de/hpoweb/showterm?id=HP:0000118",
    "URIpattern": "http://compbio.charite.de/hpoweb/showterm?id=HP:$id"
  }
]

[putmantime]

But your right, http://pir.georgetown.edu/cgi-bin/pro/entry_pro?id=PR:000000024 is resolving when i paste in my browser. Not sure why it keeps coming back 404 when I do it from the python script

[jmcmurry]

Super, thanks, none of these are failing due to poorly composed URLS, so that's the real thing to avoid. Not sure what to do about bioportal. Perhaps actually remove them from resulting merged set?

[putmantime]

Another case where I should have studied the xml better. Both have a state=down attribute. I can filter those out as well

[jmcmurry]

Well, those can be transient. I wouldn't necessarily filter them out. We can preserve the state=down flag in our merged record if we want, but we're essentially doing our own test, which is more up-to-date by definition.

[putmantime]

Ah, now I understand what the reliability scores mean. Well, would it be better to capture that from id.org, or generate our own?

[jmcmurry]

I think it is more important to do our own because it tests the uptime while also exposing situations in which we have composed the URL poorly, whether due to omitted prefix, duplicated prefix, casing issues, etc.