Refine PSDI details as an OPTIMADE provider

[alex-belozerov]

This PR refines basic information about PSDI as an OPTIMADE provider and adds a link to the index meta-db.

[ml-evs]

Looks good to me, though the build failures on our end are a bit worrying. I can try to take a look soon, but likely we don't need to hold this up

[ml-evs]

Build should work now, with some luck...

[ml-evs]

Build should work now, with some luck...

Ok it worked, but was very slow as the cache was cold (and the wait-for-it is a bit buggy as the preview doesn't get pulled down between builds anymore (I think)) -- if there's no objections I'll just merge this and we can see how things go next time we make a change.

[Bud-Macaulay]

Hey all, this commit seems to cause issues on some of our tools.
Looks like there is some sort of HSTS+308 redirect issue on /links to /links/ (or something similar that i dont have time to investigate)
We can hide this entry on our end for now but it would be nice if this can get fixed so we can show the PSDI database,

I've linked the relevant issue
aiidalab/aiidalab-widgets-base#731 (comment)

I imagine fixing this will not neccessarily require changes to this commit - so feel free to ping me when solved.

[alex-belozerov]

Hey all, this commit seems to cause issues on some of our tools. Looks like there is some sort of HSTS+308 redirect issue on /links to /links/ (or something similar that i dont have time to investigate) We can hide this entry on our end for now but it would be nice if this can get fixed so we can show the PSDI database,

I've linked the relevant issue aiidalab/aiidalab-widgets-base#731 (comment)

I imagine fixing this will not neccessarily require changes to this commit - so feel free to ping me when solved.

Hi, I have fixed that redirect (I hope). @Bud-Macaulay can you please check it again on your end?

[ml-evs]

Hey all, this commit seems to cause issues on some of our tools. Looks like there is some sort of HSTS+308 redirect issue on /links to /links/ (or something similar that i dont have time to investigate) We can hide this entry on our end for now but it would be nice if this can get fixed so we can show the PSDI database,
I've linked the relevant issue aiidalab/aiidalab-widgets-base#731 (comment)
I imagine fixing this will not neccessarily require changes to this commit - so feel free to ping me when solved.

Hi, I have fixed that redirect (I hope). @Bud-Macaulay can you please check it again on your end?

Hi @alex-belozerov, I think from https://github.com/Materials-Consortia/optimade-python-tools/actions/runs/22563241262/job/65444816839?pr=2356 it seems that this is still causing lots of redirects (I reran the tests ~an hour ago). We will fix the underlying issue with the client not catching this in Materials-Consortia/optimade-python-tools#2357 anyway.

[Bud-Macaulay]

Okay seems fixed now on my end!

Quick unrelated note: We currently can't render psdi entries as structures. I presume these data sets will use smiles/inchi as the primary way of identifying structures and have no plans of adding the other fields (sites/positions)?

[alex-belozerov]

Okay seems fixed now on my end!

@Bud-Macaulay, thanks for confirming - we have fixed the redirect from our Nginx to Github Pages. Interestingly, curl handled it fine, but the python libraries went into an infinite http-https loop.

Quick unrelated note: We currently can't render psdi entries as structures. I presume these data sets will use smiles/inchi as the primary way of identifying structures and have no plans of adding the other fields (sites/positions)?

You are right - the pchprop database will not have sites/positions in the near future.