In volume calculation mode, only load atomic weight ratio from data files

[paulromano]

Description

This PR is an alternative complementary to #2725.

@bam241 This is what I was thinking -- for volume calculation mode (which is used by openmc-plotter), we still open the .h5 files but only load the atomic weight ratio since this is all that is needed for the sake of computing densities. This won't yet allow users to run the plotter without having cross sections specified, but it at least does speed up the loading time significantly since all the cross section data itself is not loaded.

Checklist

• I have performed a self-review of my own code
• I have run clang-format on any C++ source files (if applicable)
• I have followed the style guidelines for Python source files (if applicable)
• I have made corresponding changes to the documentation (if applicable)
• I have added tests that prove my fix is effective or that my feature works (if applicable)

[shimwell]

I see this PR is just 4 lines of code so that is a nice speed improvement for such a small modification

I guess #2725 is still preferable from the user point of view as it avoids the needs for having cross sections specified and be even faster as well.

[paulromano]

@shimwell Yes, from a user's point of view #2725 would be nice, but I think it will take a bit more work to get things working with the plotter since the plotter still assumes atomic masses and normalized densities are available. This PR is actually complementary in the sense that we could merge both (updated my original description to reflect that). This will simply make things faster when cross section data is available.

[shimwell]

This looks good to me, I just want to get some representative speed ups for peoples interest

[paulromano]

@shimwell The time spent "loading cross sections" for a volume calculation (or for loading openmc-plotter) goes from some value T to basically 0, so the speedup is T/0 = ∞ 😉

[shimwell]

A geometry with Eurofer material look around 5s to load for the volume calculation. Now it takes a fraction of a second.

Nominating this PR for biggest speed improvement per line of code change award

[shimwell]

Will merge tomorrow morning (UK time) unless there are objections