Improve clipping of Mixture distributions

[paulromano]

Description

This PR adds an additional check in the Mixture.clip method for when the probability of one of the contained distributions is negligibly small. I've run into this when running shutdown dose rate calculations where an isotope at an extremely low concentration produces discrete lines in a photon spectrum that can safely be ignored.

Checklist

• I have performed a self-review of my own code
• I have run clang-format (version 15) 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]

This looks very handy for speeding up the shut down dose simulations. If I've understood this correctly then clip previously removed just gamma energies that were low probability. Now it removes two slightly different things, low probability gamma energies and low probability distributions. Perhaps the check for removing the distribution should not be on the probability of the distribution itself but if all the gamma probabilities (perhaps scaled by distribution probability) within the distribution are below the threshold? Then we have a single value that controls things consistently instead of a value that could apply to gamma probabilities or distribution probabilities. I might have got lost here with all the probabilities.

[paulromano]

@shimwell Correct, two slightly different things. The latter case (low probability distribution) is really targeted to handle the case where the distribution is not discrete and so it's not just a matter of checking the intensity. I ran into this where you can have a mixture of a Discrete and Tabular distributions. This was the best way I could think of to handle that but let me know if you have a better idea.

[shimwell]

Perhaps just adding two args, one for distribution clipping one for probability clipping. Not sure if it helps but perhaps worth considering