Ian Harry
Ian Harry
I have noticed that `pycbc_live` makes many calls to the expensive `logsignalrate` function with no triggers (or at least it does in the first minutes). These calls at least can...
Responding to #4104 this removes the `test/data` directory. Data files are now moved to the `pycbc-config` repository and pulled from there using astropy. This limits the filesize of PyCBC distributions,...
One more patch to try to help with `pycbc_live` performance: This one moves the FFTs in PSD estimation (done during inverse spectrum truncation), to the new `cached_fft` functions. I also...
We are having difficulties in getting `pycbc_live` to keep up with data in real-time in its current configuration. There may be a few patches from me aimed at this (even...
At the moment it's not easy to get the CVMFS builds to run on LVK machines. This is because the python 3.9 (and 3.6) libraries are available but not 3.8...
Similar to #4015. Recomputing FFTW plans (for data reading/high-pass filtering etc.) can be expensive. I know that the class-based API would be more efficient (expect that PR later), but it...
The conditional here: https://github.com/gwastro/pycbc/blob/master/pycbc/results/scatter_histograms.py#L598 is buggy. If there are 2 parameters and `--plot-marginal` is not given then only one image is created in the corner plot and the length of...
``` echo "usage: pycbc_copy_output_map [-h] input_map output_map [optional arguments]" ``` Although I don't think this code can or does take an input or output map as arguments. I think it...
Currently H and L are required for one of the main statistics: https://github.com/gwastro/pycbc/blob/master/pycbc/events/stat.py#L1313
`pycbc_plot_trigger_timeseries ` needs an option to show the SG veto new SNR if that's what we're using now for single detector ranking. Without this it is confusing why apparently louder...