This PR updates and adds some functionality of result.plot_corner().
It addresses the following problems:
Previously, the font size of the legend was fixed to 50 which was sufficient for a full 14 dimensional posterior plot. However, if only a few parameters are included via parameters = ["chirp_mass", "mass_ratio"], the legend was so large that it overlapped completely with the corner plot.
It was not possible to pass the truth values to the corner plot which is helpful for injections.
So far, the DINGO-IS samples were not plotted when weights was not included in the parameters list. If the user wanted to only use e.g. parameters = ["chirp_mass", "mass_ratio"], it would only plot the DINGO samples.
After this PR, the user can:
specify the font size of the legend via legend_font_size=... and adjust it for a lower dimensional corner plot.
visualize truth values via truths=... as well as the color truth_color=....
plot the DINGO-IS samples although weights are not specified as parameters.
Illustration of an example before the PR where the importance weighted samples are not shown due to the bug:
How the plot looks after the PR (when specifying an appropriate font size for the legend):
This PR updates and adds some functionality of
result.plot_corner().It addresses the following problems:
parameters = ["chirp_mass", "mass_ratio"], the legend was so large that it overlapped completely with the corner plot.DINGO-ISsamples were not plotted whenweightswas not included in the parameters list. If the user wanted to only use e.g.parameters = ["chirp_mass", "mass_ratio"], it would only plot theDINGOsamples.After this PR, the user can:
legend_font_size=...and adjust it for a lower dimensional corner plot.truths=...as well as the colortruth_color=....DINGO-ISsamples although weights are not specified as parameters.Illustration of an example before the PR where the importance weighted samples are not shown due to the bug:
How the plot looks after the PR (when specifying an appropriate font size for the legend):