american-art / PUAM

Princeton University Art Museum
Other
4 stars 4 forks source link

(!) all triples should be recorded using branded URLs #30

Open workergnome opened 7 years ago

workergnome commented 7 years ago

URIs without the data subdomain still exist in the triplestore. There are a bunch of MMOs with URIs like

http://americanartcollaborative.org/puam/object/10048

as well as

http://data.americanartcollaborative.org/puam/object/10048

Are the un-prefixed ones typos, or are they legacy data?

cathryng commented 7 years ago

i thought the data/ prefix was done globally by ISI? are these duping or did some just get missed?

workergnome commented 7 years ago

Not sure. This isn't something in your data, @cathryng, but something that would happen in the mapping.

I suspect that ether there's a typo somewhere in the mapping, where someone left of thedata. prefix, or this was done early in the mapping process and just haven't been deleted.

caknoblock commented 7 years ago

The URI prefix is set for each data file and I think someone just set one or more of the prefixes wrong. This is easily fixed.

On Jan 9, 2017, at 12:59 PM, David Newbury notifications@github.com wrote:

Not sure. This isn't something in your data, @cathryng https://github.com/cathryng, but something that would happen in the mapping.

I suspect that ether there's a typo somewhere in the mapping, where someone left of thedata. prefix, or this was done early in the mapping process and just haven't been deleted.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/american-art/PUAM/issues/30#issuecomment-271405965, or mute the thread https://github.com/notifications/unsubscribe-auth/ABB-qWy8L15YXNBVEudmfZI1E5OsQKkrks5rQp_KgaJpZM4Ld4wR.

{"api_version":"1.0","publisher":{"api_key":"05dde50f1d1a384dd78767c55493e4bb","name":"GitHub"},"entity":{"external_key":"github/american-art/PUAM","title":"american-art/PUAM","subtitle":"GitHub repository","main_image_url":"https://cloud.githubusercontent.com/assets/143418/17495839/a5054eac-5d88-11e6-95fc-7290892c7bb5.png","avatar_image_url":"https://cloud.githubusercontent.com/assets/143418/15842166/7c72db34-2c0b-11e6-9aed-b52498112777.png","action":{"name":"Open in GitHub","url":"https://github.com/american-art/PUAM"}},"updates":{"snippets":[{"icon":"PERSON","message":"@workergnome in #30: Not sure. This isn't something in your data, @cathryng, but something that would happen in the mapping. \r\n\r\nI suspect that ether there's a typo somewhere in the mapping, where someone left of thedata. prefix, or this was done early in the mapping process and just haven't been deleted.\r\n"}],"action":{"name":"View Issue","url":"https://github.com/american-art/PUAM/issues/30#issuecomment-271405965"}}}

VladimirAlexiev commented 7 years ago

There is also some data recorded against branded (pure) URLs like http://data.crystalbridges.org/object/108.

I personally prefer this (and I think the museums should). But certainly having 3 varieties of URLs will make everyone's life unpleasant.

workergnome commented 7 years ago

@VladimirAlexiev: the branded URLs are intentional—art this point, Craig has them set up to redirect to the AAC endpoint, but this will allow migration to self-hosting once complete.

VladimirAlexiev commented 7 years ago

Yes, that's what I understood, and https://github.com/american-art/semantic-hosting/issues/4 describes how they can stay that way in the browser (rather than being AAC-ified) using Apache ModRewrite.

This issue is that semantic data is recorded against 3 varieties of URLs: all of it should be against branded URLs. I'll change the title

workergnome commented 7 years ago

One of the reasons that we're seeing this is because not all museums have set up the redirects already. I'm not sure how that progress is happening—that's @caknoblock's domain, but I know that some institutions have already set up their redirects and can use the branded URLs, and some of them wanted more time, so are using the http://data.americanartcollaborative.org/puam/object/10048 urls until the redirects can be set up. the http://americanartcollaborative.org/puam/object/10048 are typos or problems and should be fixed and deleted.

VladimirAlexiev commented 7 years ago

When you generate and store 10M triples, you better do it against stable URLs. IMHO the RDF conversion shouldn't depend on redirection details that change monthly. http://data.artmuseum.princeton.edu doesn't resolve at present, but so what: you can still use this URL in the AAC repo, and the browse app will work. When they get around to setting a redirect, that URL will lead to the AAC server.

cathryng commented 7 years ago

our redirect is now live - at the url mentioned above

workergnome commented 7 years ago

Great. @caknoblock, flagging this for update once people are back.

VladimirAlexiev commented 7 years ago

The general issue is https://github.com/american-art/semantic-hosting/issues/4