These numbers have been compared to the reviewed results for GW190814 `combinedPHM` at the bottom of this page [1]
These numbers have been compared to the reviewed results for GW190814 `combinedPHM` at the bottom of this page [1], which correspond to numbers cited in the discovery paper [2], which we report her:
* upper value `mass ratio`
* lower value `total_mass_source`
* lower value `mass_1_source`
* lower limit `theta_jn `
* both limits for `L1_matched_filter_snr_abs ` (0.05 difference for the lower limit)
* Same for `H1_matched_filter_snr_abs `
* Same for `V1_matched_filter_snr_abs` , with a 0.13 difference.
My guess is that this is due to different versions of `np.percentile` (pesummary uses ). To cross-check this, I have personally run the script at the page linked [1]:
```python
importnumpyasnp
frompesummary.gw.file.readimportread
parameter_dict={
"chirp_mass_source":"Chirp mass $\mathcal{M}/M_{\odot}$",
"total_mass_source":"Total mass $M/M_{\odot}$",
"mass_ratio":"Mass ratio $q$",
"mass_1_source":"Primary mass $m_{1}/M_{\odot}$",
"mass_2_source":"Secondary mass ${m_{2}/M_{\odot}}$",
I have adapted the script to run only on `combinedPHM` data. I had also to change `matched_filter_snr` tag in `matched_filter_snr_abs`. Plus I have added two lines that print the intervals in the form [min, max]
[1]:[https://git.ligo.org/publications/gw190814/gw190814-discovery/-/wikis/PE-result-table](https://git.ligo.org/publications/gw190814/gw190814-discovery/-/wikis/PE-result-table) . There are sum difference in the following parameters
[1]:[https://git.ligo.org/publications/gw190814/gw190814-discovery/-/wikis/PE-result-table](https://git.ligo.org/publications/gw190814/gw190814-discovery/-/wikis/PE-result-table) . There are sum difference in the following parameters