SNR optimizer option types

[tdent]

Fix issue in non-default settings for some options that Max complained about

[ArthurTolley]

Just tested this change and it fixes the previous TypeError: can't multiply sequence by non-int of type 'float' error. Happy to approve

[tdent]

Was that for the time window option or f-lower ?

[ArthurTolley]

@tdent I can re-run the test again to find the exact error but I simply gave --snr-opt-extra-opts \ "--chirp-time-f-lower 20.0 --chirp-time-window 2.0 " \ and the error was there before the fix and not there after the fix :), like I said, I will re-run this in a minute.

[ArthurTolley]

This is the full error:

Traceback (most recent call last):
  File "/home/arthur.tolley/.conda/envs/live_rank_stat/bin/pycbc_optimize_snr", line 141, in <module>
    tau0 = cv.tau0_from_mass1_mass2(fp['mass1'][()], fp['mass2'][()], tau0flow)
  File "/home/arthur.tolley/.conda/envs/live_rank_stat/lib/python3.9/site-packages/pycbc/conversions.py", line 364, in tau0_from_mass1_mass2
    return tau0_from_mtotal_eta(mtotal, eta, f_lower)
  File "/home/arthur.tolley/.conda/envs/live_rank_stat/lib/python3.9/site-packages/pycbc/conversions.py", line 337, in tau0_from_mtotal_eta
    return _a0(f_lower) / (mtotal**(5./3.) * eta)
  File "/home/arthur.tolley/.conda/envs/live_rank_stat/lib/python3.9/site-packages/pycbc/conversions.py", line 322, in _a0
    return 5. / (256. * (numpy.pi * f_lower)**(8./3.))
TypeError: can't multiply sequence by non-int of type 'float'

So f_lower.

[tdent]

Yup, f_lower gets multiplied whereas the window gets subtracted or added ;)
Now I just don't know what is causing the 'run small workflow/search' things to fail ...

[ArthurTolley]

For the CI, maybe related to the pegasus upgrade?

Just for completeness, when applying the fix to f_lower and not chirp_time_window we get another error which is also fixed by this explicit type :)

Traceback (most recent call last):
  File "/home/arthur.tolley/.conda/envs/live_rank_stat/bin/pycbc_optimize_snr", line 143, in <module>
    mintau0 = max(tau0 - args.chirp_time_window,
numpy.core._exceptions._UFuncNoLoopError: ufunc 'subtract' did not contain a loop with signature matching types (dtype('float64'), dtype('<U3')) -> None

[maxtrevor]

While we are improving the command line options, can we please change the verbosity option to use https://github.com/gwastro/pycbc/blob/master/pycbc/__init__.py#L69 ?

[tdent]

Max if you make a code suggestion (via the comment interface) we can accept it, although the verbose option is not related to this specific fix

[maxtrevor]

made the requested change myself, and somehow it seems to have gone through without needing to be accepted?

[tdent]

Oh, yeah, github does let you edit someone else's PR just like that ..

[titodalcanton]

Remember to apply the right label to the PR, otherwise this risks not being deployed in the 2.1 release line.