We need to profile the ecc measurement code

[vijayvarma392]

Would be good to understand which part of the code is slowest, etc.

[vijayvarma392]

Here's something that would be nice for the paper:
For waveforms with approx params of GW150914 and GW170817, run the different methods and document the costs in a table. The reason this would be nice is we can say in the paper how much extra cost is involved in our postprocessing.

For this, I would write a script and add it to gw_eccentricity/EccTests/TimingTests so that the test can be run repeatedly as we make improvements.

[vijayvarma392]

This also sounds like higher priority, but as time permits. More important than the table for the paper is to run a python profiler and see if there are bottlenecks that can be avoided.

[vijayvarma392]

@md-arif-shaikh: I think this is probably a high priority item now, as we get ready to review gw_ecc. We need to understand:

1. How fast is the gw_ecc evaluation time for a single waveform for a 20 orbit signal and a 100 orbit signal.
2. Which part of the code is dominating the cost? Can that be improved with some optimization?
3. Should we only compute ecc using a small segment of the signal? What is the error introduced by this?

[md-arif-shaikh]

@vijayvarma392 sure I will check this out. I will work on it after #215 is merged.