Google dataset search support

[lmarini]

Add required jsonld schema.org fields to pages for indexing with google data search.

[MBcode]

I can start to look at taking the mappings from v1 that would apply here, but
also want to think a little about how schema.org fields that come from class attributes
might interop with user uploaded metadata attached to the same object

[MBcode]

Have branch: First attributes don't even need mapping after a json dump (from: get_dataset_metadata), just the addition of a 1st: "@context" : { "@vocab" : "https://schema.org/" }, element; then will have to coordinate to figure out how to get this within a <script type="application/ld+json"> w/in the dataset html; that is usr-metdata, not the internal class metadata, that we see in routers/datasets, so writing a form of to_jsonld for this w/new mapping

[MBcode]

Have a datasetout2jsonld and wanted to know if I should put it as a route, before coordinating w/the front end?

[lmarini]

Yes, route would be great. Thank you. From: mike bobak ***@***.***> Date: Wednesday, October 5, 2022 at 4:59 PM To: clowder-framework/clowder2 ***@***.***> Cc: Marini, Luigi ***@***.***>, Author ***@***.***> Subject: Re: [clowder-framework/clowder2] Google dataset search support (Issue #63) Have a datasetout2jsonld<https://urldefense.com/v3/__https:/github.com/clowder-framework/clowder2/blob/63-Google-dataset-search-support/backend/app/routers/datasets.py*L223__;Iw!!DZ3fjg!-QklULvlE2a_mhOQBVSa19BlL7aurW35OPtA0V5joRkIVhb_EVTuwFpPc3285dLN4aJdq4jXXA1NXK56IJuT3O6S5g$> and wanted to know if I should put it as a route, before coordinating w/the front end? — Reply to this email directly, view it on GitHub<https://urldefense.com/v3/__https:/github.com/clowder-framework/clowder2/issues/63*issuecomment-1269024703__;Iw!!DZ3fjg!-QklULvlE2a_mhOQBVSa19BlL7aurW35OPtA0V5joRkIVhb_EVTuwFpPc3285dLN4aJdq4jXXA1NXK56IJtqBB7HSg$>, or unsubscribe<https://urldefense.com/v3/__https:/github.com/notifications/unsubscribe-auth/AAESNLLYFSV6H6FWGUOAQH3WBX225ANCNFSM55IVQM7A__;!!DZ3fjg!-QklULvlE2a_mhOQBVSa19BlL7aurW35OPtA0V5joRkIVhb_EVTuwFpPc3285dLN4aJdq4jXXA1NXK56IJuGFHlh9w$>. You are receiving this because you authored the thread.Message ID: ***@***.***>

[MBcode]

cool, I'll continue were I left a stub for it, and update here/tnx!

[MBcode]

get_dataset_jsonld route, off of get_dataset now, so can get just one for the frontend rendering of that one dataset page, but the problem is this method gets a dict from mongo that has objids & datatimes which are not serializable, as originally needed
So for now I'm dropping those keys till I can get around that, here; just needed: dso= DatasetOut.from_mongo(dataset) to be able to dso.json() and easily remap all the keys to schema.org

[MBcode]

Have jsonld script route will look/ask about getting into the dataset page;
just have to figure out how I should change the signature of the route, maybe end w/ /jsonld ? for this non-usr metadata. Once we agree on this I'll make a PR.
Talked w/Chen maybe Aruna can integrate

[MBcode]

I have the working start of the sitemap which I could get it with the making the ld+json script; but think it best to take advice of having someone who knows typscript have it inserted. So this could be 2 or 3 PRs, depending on if I should include the sitemap generation now too, lmk

[lmarini]

Can you add it to the backend as an endpoint? The frontend will have to retrieve it from there anyways.

[MBcode]

I have the /jsonld route in, and will add the sitemap, but need a public setting to be able to crawl it
Will check dependencies for routes, /sitemap.xml also has to be open
If /summary.jsonld is open then sitemap could point to it for now

[MBcode]

will take out of draft PR once I can open these routes, but still getting: ‘Cannot GET /sitemap.xml’ &similar for summary.jsonld route, even w/o the dependencies= line in main.py