New header **kwargs patch for plot_data_greens3#333
Conversation
This should solve issue #332.
|
I am fine waiting a bit for you to tighten things up. My workaround currently works fine for this particular project |
…r in waveform figures. This triggers per-line (only the line that would overflow is rescaled). I tested it on the mtuq examples and it looks okay to me. Also tested with plot_data_greens3 and everything looks fine as well. I turned it on by default, but we could change that if needed.
|
Hi @ammcpherson, Could you please test this one out, I think it should do the trick! (It should shrink down the overflowing line). |
|
Not sure why the build isn't working with this one, I have not added any dependencies. |
|
@thurinj So this branch would not have the new misfit function? That is more vital to this particular project of mine than the header/waveform plotting issue |
|
Hi @ammcpherson, the branch doesn't have it, but the PR only brings the changes in the header.py file (see the list of modified files in the PR GitHub recap). |
Python 3.14 (released oct of this year) seems to cause issues with the continuous integration tests, so I fixed it to 3.13. So far the tests are working on my machine.
thurinj
force-pushed
the
waveform_plot_update
branch
from
0244c8c to
c8cc4a7
Compare
Changed default behavior for collect_synthetics (but also collect_attributes, even if not really used much in the code -- for consistency). I also noticed that if the normalized misfit per trace was not stored as attributes in the level0, then it would divide the raw misfit by the total (normalized) misfit. So in the 2009 eq event, this translates to doing ~1e-9 / ~1e0 'hybrid' had not been properly implemented, so I did a light rework of the calculate_norm_data function.
|
Since it was loosely related with the waveform plot, I also made changes to address issue #335. |
2 participants
This should solve issue #332 - fairly minimal fix.
I am not a huge fan of the resulting header when using 3 different filtering parameters:
(the time window duration is clipped out of the figure).
I can merge the PR as is in the next day or two if you want (so @ammcpherson can move forward with her work), or take a bit more time to rework the overall text arrangement so everything fits tightly in frame of the figure. (We can probably have the layout a bit smarter depending on what we have to insert in the figure's header.)
Let me know what's the best for you.
tag: @rmodrak, @ammcpherson