From 12eaf3b924f22d713777be70254b76279737c73b Mon Sep 17 00:00:00 2001 From: Pablo Lichtig Date: Mon, 26 Aug 2024 15:10:18 -0600 Subject: [PATCH] Update docs/appendix/troubleshooting.rst Co-authored-by: Zachary Moon --- docs/appendix/troubleshooting.rst | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/appendix/troubleshooting.rst b/docs/appendix/troubleshooting.rst index 5a01e02e..7d18c80e 100644 --- a/docs/appendix/troubleshooting.rst +++ b/docs/appendix/troubleshooting.rst @@ -13,6 +13,6 @@ Runtime issues * The plots are not produced. The error message references ``leg.legendHandles``: * You are probably using Matplotlib 3.9+ with pandas 1.x. Downgrade Matplotlib to 3.8 (upgrading pandas should also work, but you might run into some incompatibilities for some specific functionalities, especially those related to downloading observational data with MONETIO. Check :doc:`/getting_started/installation`). * NaN does not exist, or pandas not detecting NaN values, etc: - * numpy v2.0.0 removed np.NaN, which has been completely replaced by np.nan. np.NaN is required by some versions of MONETIO. There are two solutions: (i) downgrading numpy to previous versions or (ii) upgrading MONETIO. If you go with option (ii), you might need to use the _develop_ branch of MONETIO (this should make it into _stable_ soon). + * In NumPy 2.x, ``np.NaN`` has been completely replaced by ``np.nan``. ``np.NaN`` is required by some versions of MONETIO. There are two solutions: (i) downgrading NumPy to previous versions (``'numpy<2'``) or (ii) upgrading MONETIO (``'monetio>=0.2.7'``). If you go with option (ii), you might need to use the _develop_ branch of MONETIO (this should make it into _stable_ soon). * Failure downoloading maps: * MELODIES-MONET uses Cartopy for mapping. Cartopy downloads shapefiles automatically to create the maps, and if you are offline or working on a machine with download restrictions, this might fail. Check :doc:`/appendix/machine-specific-install`. This is the situation for *NOAA HPC Hera*, and the solution discussed there should work.