Remove PSDEF in favor of PSL

[AntonReinhard]

The docs building still fails, because unfortunately the documentation still has a cyclic dependency to QEDcore and QEDprocesses. I think the best way forward is to remove the dependency from the literate.jl tutorials by using self-defined custom types and removing or changing the jldoctests that rely on QEDcore or processes. Those should probably be moved to their respective packages.

[AntonReinhard]

The docs should build for now, but they're not very pretty yet. They now use each other's custom definitions instead of QEDcore or QEDprocesses. For this I had to add a FourMomentum and a AbstractModel tutorial, both of which are very bare bones for the moment.

I also had to disable some of the jldoctests where I'm not sure if they're doable nicely here. I just changed them to julia codeblocks instead of jldoctest for now. I think the functionality jldoctests should generally be moved to where it's implemented, i.e. QEDcore or QEDprocesses.

Other than that I think this is the way to go, because otherwise our cyclic dependencies problem will still persist.

@szabo137 thoughts?

[szabo137]

The docs should build for now, but they're not very pretty yet. They now use each other's custom definitions instead of QEDcore or QEDprocesses. For this I had to add a FourMomentum and a AbstractModel tutorial, both of which are very bare bones for the moment.

I also had to disable some of the jldoctests where I'm not sure if they're doable nicely here. I just changed them to julia codeblocks instead of jldoctest for now. I think the functionality jldoctests should generally be moved to where it's implemented, i.e. QEDcore or QEDprocesses.

Other than that I think this is the way to go, because otherwise our cyclic dependencies problem will still persist.

@szabo137 thoughts?

I think for now, it is fine to leave it like that. Eventually, we will have TestImplementation ready to be a dependency of the docs (see #151 gor details). Then we could just use it to mock the functionality from QEDcore, which is not available anymore.

[AntonReinhard]

Okay sounds good, then I'll just go over it a little bit tomorrow and it should be ready. Then we'll have to fix the integration tests by adjusting QEDcore.

[szabo137]

In general it looks very good, many thanks for that!
I've just some comments and one request: please look up the PSL interface and use it properly in the construction of the PSPs.

[szabo137]

Okay sounds good, then I'll just go over it a little bit tomorrow and it should be ready. Then we'll have to fix the integration tests by adjusting QEDcore.

This is correct! I think after adjusting QEDcore, the other packages are pretty easy to fix.

I think it is our first multi-level fix, showing the greatness of our integration test suite 👯‍♂️ ( @SimeonEhrig stay tuned if we break something, again 🫶)

[AntonReinhard]

please look up the PSL interface and use it properly in the construction of the PSPs.

You're completely right. I mostly just removed the psdef and replaced occurrences with the psl, intending to see what needs to be changed when it breaks. But it seems it didn't break in those instances so I forgot to actually use the new interface ^^

[szabo137]

please look up the PSL interface and use it properly in the construction of the PSPs.

You're completely right. I mostly just removed the psdef and replaced occurrences with the psl, intending to see what needs to be changed when it breaks. But it seems it didn't break in those instances so I forgot to actually use the new interface ^^

This was also my guess ^^
However, I don't know, maybe this is the time to add some unit tests for the PSPs to check the construction from coordinates — TDD style.

[SimeonEhrig]

I think it is our first multi-level fix, showing the greatness of our integration test suite 👯‍♂️ ( @SimeonEhrig stay tuned if we break something, again 🫶)

Be aware, my weekends starts 5 minutes before you break my CI ^^

[szabo137]

From my side, this looks good now. Therefore, we can proceed fixing the integration tests by opening the respective PRs in the other packages.

[szabo137]

@AntonReinhard I removed the PSDEF from QEDcore: QEDjl-project/QEDcore.jl#94

Please retrigger the CI with this fix. Locally, the unittests in QEDjl-project/QEDcore.jl#94 pass with this version of QEDbase.

[AntonReinhard]

The formatter fails on the tutorials, because I did not indent the includes, which would have to be indented, but the reason for the indent is not visible in the docs, so in the docs it looks weird. And if I turn off the format for those lines with a format: off, then that comment is also visible in the docs. Do we just want to turn off the format check for the tutorials?