Interpolate x value for plotting #64
Conversation
Codecov Report
❗ Your organization needs to install the Codecov GitHub app to enable full functionality. @@ Coverage Diff @@
## main #64 +/- ##
==========================================
- Coverage 96.09% 96.08% -0.01%
==========================================
Files 13 13
Lines 972 971 -1
==========================================
- Hits 934 933 -1
Misses 38 38
📣 Codecov offers a browser extension for seamless coverage viewing on GitHub. Try it in Chrome or Firefox today! |
|
Nice idea. You could check how it looks like in a scatter plot of the numerical data overlayed on the reference data. If you don't see an improvement of the phase shift there, there is no chance to get better results using intrpolations. |
|
A scatter plot does not really change much, but there is also a slight difference in the frequency, such that the phase shift becomes smaller over time. Looks like the results are closer to the solution from Svärd and Kalisch (at their finest grid) than to the experimental data. |
|
I would like to merge this anyway since linear interpolation should be more appropriate than constant. Technically, I would call this a breaking change since it changes the output of a plot. Since #69 was also breaking already, I will bump the minor version after merging this PR. |
I was hoping to improve the plots by not using the constant interpolation at a position
x, but using higher order interpolation, in particular I was hoping to have a smaller phase shift in some plots, where we plot the solution at certainxvalues that are not on the grid. Unfortunately, the plots did not really improve, also not for cubic interpolation.