Adding functionalities to create calibration config files

[sum33it]

To incorporate calibration uncertainties in PE runs, one need to provide a calibration configuration files to pycbc_inference code. Although, the support for the calibration config file exist, we still do not have a way to generate these config files, within PyCBC infrastructure. In this PR, I include the full infrastructure to generate the config files for PE runs for any GPS time from the publicly available calibration envelopes provided by LIGO-Virgo.

[ahnitz]

@sum33it I don't think there are any strong objections to this. Maybe you can add a short example of using this? If that works, then I think we can go ahead and merge.

[ahnitz]

@sum33it What's the status of this? Can we get this rebased and then merged?

[sum33it]

@ahnitz I think this PR is ready for the review.

[ahnitz]

@sum33it Can you add an example to the documentation? I think in this case, that would be very useful and help verify the code continues to work.

[sum33it]

I will open a follow-up PR for detailed documentation and examples. But here, I have added an example script of how to generate a calibration config file for GW150914. The extension to other events is trivial: one needs to provide appropriate arguments like ifo-list and trigger-time.

[mj-will]

I've had a look over the new function for reading the calibration envelopes and added a few minor comments.

[mj-will]

I don't think this is necessary.

[sum33it]

In general, it might not be needed, but I still want to keep it in case we go beyond the frequency range of the calibration provided by the calibration envelopes. For example, if someone wants to do PE with a lower cutoff frequency than the standard, they might be tempted to give the same f_lower for calibration config file generation. This option should safeguard against that.

[mj-will]

Ah, just to clarify, I mean then else part of the if/else statement. I think it's a good idea to keep the checks you implemented.

[sum33it]

Got it. Thanks for the clarification. I agree with that. Will do in the following commit.

[sum33it]

@mj-will I have made the necessary changes and pushed the new commit.

[mj-will]

Same as above

[sum33it]

Hi @mj-will
Thank you for the review. I have made appropriate changes. Could you please check if you are happy with this?

[mj-will]

Thanks @sum33it for addressing my comments. I'm happy to approve this (assuming the tests pass and someone which the relevant expertise reviews the executable.)

[sum33it]

Thanks a lot @mj-will .
@ahnitz if you don't have objection, shall we merge this PR?

[ahnitz]

@sum33it The unit tests need to pass. That's why I said to rebase, then see if anything needs to be addressed there. Otherwise, I don't have other objections.

[sum33it]

@ahnitz I have rebased it and there is one failed test and it seems that it is not originating from my commits. What do you suggest?

[GarethCabournDavies]

@sum33it Try rebasing to gwastro:master and starting the checks again - they've expired now so I can't check the run failures, but I will take a look at them if it fails again

[sum33it]

@GarethCabournDavies The two failed tests are related to i) py-docs, and in ii) retrieving frame data with bash -e examples/search/get.sh. I don't see any of them related to my commits.

[GarethCabournDavies]

I think the sphinx documentation failure is real - when I try and run the examples/inference/generate_calibration_config.sh script, I am getting an error:

Traceback (most recent call last):
  File ".../bin/pycbc_inference_create_calibration_config", line 127, in <module>
    calibration_files_dict = get_calibration_files_O1_O2_O3(opts.ifos, gps_time, calibration_files_path)
  File ".../pycbc/pycbc/strain/recalibrate.py", line 610, in get_calibration_files_O1_O2_O3
    ifo_calibration_file = '%s/%s'%(os.getcwd(), all_calibration_files[dt_list.argmin()])
ValueError: attempt to get argmin of an empty sequence

The documentation will run the examples in the examples/inference folder, but it currently doesn't give the failure / output nicely

[GarethCabournDavies]

(which makes sense as it hardcodes files in your own path)

[GarethCabournDavies]

These changes should work to make the example run locally.

As an aside, it would neaten up the code if you could change print statements to logging, and use the pycbc.add_common_pycbc_options and pycbc.init_logging functions, but no worries if you can't.

[sum33it]

Thanks @GarethCabournDavies I implemented your changes in the example folder and now we have only sphinx documentation related failure. Shall I just merge this at this point?

[sum33it]

@ahnitz Do you have any objection on merging it? Failed test seems to be coming from sphinx documentation failure which is not related to this PR.

[spxiwh]

The documentation failure is related to this patch. It must be fixed before this can be merged:

/home/runner/work/pycbc/pycbc/.tox/py-docs/lib/python3.11/site-packages/pycbc/strain/recalibrate.py:docstring of pycbc.strain.recalibrate.get_calibration_files_O1_O2_O3:6: WARNING: Title underline too short.

[sum33it]

Thanks a lot @spxiwh for pointing this out. I have fixed that and now there is one skipped test related to deploy_documentation. I can't see it being related to my PR. I hope it is OK to merge this PR now.

[sum33it]

@spxiwh, @GarethCabournDavies , @ahnitz: If you have no further objections, shall we merge it now?

[GarethCabournDavies]

You can got for it merging for me - Ive made a couple of suggestions for using logging rather than printing.

It looks like one of the print statements should be an error. You can change this to logging.warning if preferred though

[sum33it]

Thanks @GarethCabournDavies for suggestions! I have accepted your suggested commits. I will merge it sometime today.

[GarethCabournDavies]

Weirdly, this passed without #5156 ??