Add cache to approximant finding

The call to the lalsim approximant function happens during every waveform generation with always the same input.

This MR will drastically reduce the number of those calls as long as fewer than 10 different waveform approximants are used during an analysis (that should be safe for most users.)

While the function is fast, the repeated calls can add up over an analysis.

added Efficiency label

approved this merge request

@colm.talbot what do you mean by "fewer than 10 different waveforms". Looking at the code, it looks like, if a user calls more than 10, they simply won't get the savings from the cache.

changed the description

changed milestone to %1.2.0

approved this merge request

resolved all threads

merged

mentioned in commit cbcf32a0