Parser improvements

[jClugstor]

Checklist

• [x ] Appropriate tests were added
• [x ] Any code changes were done in a way that does not break public API
• [ x] All documentation related to code changes were updated
• [ x] The new code follows the
  contributor guidelines, in particular the SciML Style Guide and
  COLPRAC.
• [x ] Any new documentation only uses public API

Additional context

Add any other context about the problem here.

This PR adds proper support for parsing in to an AST, and evaluating that AST to get a ModelingToolkit system. This means that the parsing no longer relies on parsing Julia expressions in to symbolics expression, adds proper support for all BaseModelica operators, etc.

These changes were necessary to make implementing all of the BaseModelica specification possible in the future.

[ChrisRackauckas]

Precompilation failure?

[jClugstor]

Yeah, not sure why, looks like MTK can't precompile

ERROR: LoadError: `unsorted_arguments(x)` is deprecated, use `arguments(x)` instead.

[ChrisRackauckas]

https://github.com/JuliaSymbolics/SymbolicUtils.jl/blob/d009ca8e986b12982dce13794218ea6b632da617/src/SymbolicUtils.jl#L23

[jClugstor]

I found that, but isn't that just where the deprecation is set? How would I turn using istree in to an error, isn't that what depwarn=error does? Wouldn't the warning be coming from wherever istree is called, which is what I want to find?

[ChrisRackauckas]

just delete that line. Make it impossible to get a deprecation

[jClugstor]

Oh yeah I tried that, all the tests pass with out any errors.

[jClugstor]

I know what's happening. Basically the ambiguities test for aqua creates a new Julia instance, and it uses Base.julia_cmd() to do it. So any options that were used to start the original Julia instance would be applied to it.

# Ambiguity test is run inside a clean process.
    # https://github.com/JuliaLang/julia/issues/28804
    code = """
    $(Base.load_path_setup_code())
    using Aqua
    Aqua.test_ambiguities_impl(
        $packages_repr,
        $options_repr,
        $exclude_repr,
        $skipdetails,
    ) || exit(1)
    """
    cmd = Base.julia_cmd()
    if something(color, Base.JLOptions().color == 1)
        cmd = `$cmd --color=yes`
    end
    cmd = `$cmd --startup-file=no -e $code`

When I precompile MTK in a Julia session with depwarn=error I get a warning about how istree is deprecated, even though MTK doesn't use istree anymore. I get a warning, not an error.

I think when Aqua gets to the point of using ModelingToolkit in BaseModelica, the istree warning is triggering an error.

I'll try to find where this warning is coming from and what's causing it.

[ChrisRackauckas]

Just throw a testset over the tests for now so it runs them all. See ModelingToolkit's test form for that.

[jClugstor]

They're all in testsets and safetestsets. Just the Aqua tests fail.

[ChrisRackauckas]

testset over that

[ChrisRackauckas]

Just do exactly as the ModelingToolkit tests...

[jClugstor]

Actions was down for a bit. Yeah, the GROUP stuff, right?

[ChrisRackauckas]

https://github.com/SciML/ModelingToolkit.jl/blob/master/test/runtests.jl#L19

[jClugstor]

It decides to work now. Simply lovely

[ChrisRackauckas]

You're not running anything if you haven't set any test groups.

[ChrisRackauckas]

https://github.com/SciML/ModelingToolkit.jl/blob/master/.github/workflows/ci.yml#L59-L62

[jClugstor]

Is that not what I did here? b85b7c2#diff-f88e7ee60a80885f57e6c7794c2dd3d3604ffed08c7c2bde856b1669186e8444R32

[jClugstor]

It's actually running the tests now, I had to add with: group: "${{ matrix.group }}" to make sure it actually runs the groups :-|

[jClugstor]

Is it ok to merge this?

[ChrisRackauckas]

Run the formatter.