search

gwastro / pycbc

Core package to analyze gravitational-wave data, find signals, and study their parameters. This package was used in the first direct detection of gravitational waves (GW150914), and is used in the ongoing analysis of LIGO/Virgo data.
http://pycbc.org
GNU General Public License v3.0
307 stars 344 forks source link

issue when using PSD from a file #4780

Open aleynaakyuz opened 1 month ago

aleynaakyuz commented 1 month ago

I am working with the strain sensitivity curves for Cosmic Explorer, which are available here. When I read the data files using the pycbc.psd.read.from_txt function with df=0.1, low_freq_cutoff=5.1, and length=4000, I observed that the resulting curves appear flat, as shown in the attached plot.

After examining the issue, I found that this issue does not occur with different values for low_freq_cutoff. The problem arises due to floating point arithmetic. Specifically, the value of low_freq_cutoff is 5.1 and delta_f is 0.1, so kmin should ideally be 51. However, in Python, the calculation 5.1 / 0.1 yields 50.99999999999999, and since the code uses the int function, kmin is set to 50. This causes flow to be 5 and data_start to be -1 because the first value of freq_data is slightly above 50 for Cosmic Explorer.

I would suggest that you resolve this issue by replacing the int function with the round function in the relevant part of the code. This change will ensure that kmin is correctly calculated as 51 when low_freq_cutoff is 5.1 and delta_f is 0.1.

low_freq_cutoff5_1_40km

WuShichao commented 1 month ago

Thanks, I know this f_lower issue, I just use 5.05 Hz for CE. Any comments from @ahnitz?

ahnitz commented 2 weeks ago

@aleynaakyuz Can you please rebase this PR from master and check on the code climate report afterwards?

ahnitz commented 2 weeks ago

@WuShichao Can you verify your case as well, to see that you no longer have this restriction?

WuShichao commented 2 weeks ago

@ahnitz I just realized that my issue was not caused by this, just because the initial frequency of CE asd file is 5.000000000000000888e+00 not 5. I have checked Aleyna's case, this PR can fix her issue.

ahnitz commented 3 days ago

@aleynaakyuz You'll need to rebase this PR and make sure the unittests pass along with the code climate report (found as a link towards the bottom of the checks).