Problem determining nnmax for slabs about tensoap

[dilkins]

This might be the best option; when writing the code originally, there was quite a bit of flipping between setting a hard nnmax, which is "safe" but often slow, and setting one by calculating the density, which is faster but only really works for systems with uniform density; having a flag that sets nnmax to a hard limit is inelegant (especially as sagpr_get_PS has so many command-line flags already), but it's probably the best option. Feel free @alanmlewis to implement and open a PR -- otherwise I will be able to get to it in a couple of days' time.

[alanmlewis]

Another possibility would be to offer a flag where the user could set the occupied fraction of the cell at runtime, and then simply multiply the volume by this fraction? That may require a little trial and error on the user's part, but is more flexible than simply setting a hard limit.

[dilkins]

Agreed, that's probably a better idea; since we're admitting trial and error, probably some extra exception handling is needed so that the code doesn't crash so messily.

[alanmlewis]

Yes. for sure; I'll have a go at implementing that tomorrow.

[alanmlewis]

Resolved by commit f448a04