freyberg_prior_monte_carlo in part 2

[laat0003]

Hi,

I am unsure if this is related to the other bug report but the Prior Monte Carlo notebook currently fails when you first try to use pyemu.plot_utils.ensemble_res_1to1(). From what I can tell the function _get_plotlims() is the problem. Specifically, the following lines, which seem to work fine when passing in a base_ensemble but fail when base_ensemble = None.

line1413 bemin = np.min([bemin, beni.min().min()]) line1414 bemax = np.max([bemax, beni.max().max()])

My band-aid solution to get the notebook working was to do the following

if base_ensemble is not None: #I added this bemin = np.min([bemin, beni.min().min()]) # original code bemax = np.max([bemax, beni.max().max()]) #original code else: #I added this bemin = np.min([bemin, beni.min()]) #I added this bemax = np.max([bemax, beni.max()]) #I added this

It all worked fine after that.

Cheers, Tariq

[jtwhite79]

Thanks @laat0003 - that function is a problem in general; it was causing issues with the latest jump to pandas 2.X in pyemu also and I lost interest in trying to patch it. Are you using the conda env and the the pyemu src that comes with the notebooks?

[laat0003]

My conda env may not be the same but I am using the pyemu src that comes with the notebooks. I've attached my conda env if it is of any use.

Cheers for the quick reply.

my_env.txt

[rhugman]

Hey @laat0003,

I've updated the env yml file to require pandas <2.0 and plotting is still working with a fresh env install. Sounds like an incompatibility with your env. This is why we try to keep the repo self-contained.

btw @jtwhite79, we should at some point update all of this with the post-pandas 2.0 version of pyemu...adding to the to-do list.

Cheers, R

[laat0003]

Curses! You are correct. My apologies @rhugman and @jtwhite79. I'm too big of a fanboi for these notebooks so just downloaded and binged them without reading the fine print. I'll make sure to double-check env compatibility in the future.

All the best, Tariq