make sure FD white_noise has the same length as asd

[WuShichao]

Standard information about the request

This is a: bug fix

This change affects: noise generation, inference

This change will: avoid noise generation bug in some corner LISA PE examples

Motivation

I encountered the len not match issue in some LISA tests, not all of them:

2024-09-12T22:22:28.777+00:00 INFO : Making colored noise
2024-09-12T22:22:30.494+00:00 INFO : len(psd): 2211906
2024-09-12T22:22:30.649+00:00 INFO : len(asd): 2211906
2024-09-12T22:22:31.943+00:00 INFO : int(wn_dur * sample_rate) // 2 + 1: 2211906
2024-09-12T22:22:31.944+00:00 INFO : len(white_noise): 2211907

Contents

Force resize the FD white_noise.

Links to any issues or associated PRs

Testing performed

Additional notes

• The author of this pull request confirms they will adhere to the code of conduct

[ahnitz]

I'm not sure this is exactly what we want to do here. Maybe the fix should be in the function just above. We definitely can't just resize the freuqency-domain object, but that distorts the sample rate (perhaps only slightly if the vector is long, but still not quite right).

[spxiwh]

I don't understand the problem this is trying to solve. This function takes start_time, end_time and sample_rate, which completely specifies the noise (and so we shouldn't do this).

But if this is a problem, it implies an issue above, as we do do a psd.resize previously to make things match.

Can you give an example of a failure?