monocongo/climate_indices

all tests working except palmers

Closed this issue · 4 comments

all tests have run successfully except palmers. Attached are screenshots of the errors when trying to run. Below is the command and initial error... there is also a long list of other exceptions occurred when handling this one.

(palmers) Caryns-MacBook-Pro:climate_indices-master caryn.ross$ process_climate_indices --index palmers --periodicity monthly --netcdf_precip /Users/caryn.ross/Desktop/climate_indices-master/example_data/input/nclimgrid_lowres_prcp.nc --var_name_precip prcp --netcdf_pet /Users/caryn.ross/Desktop/climate_indices-master/example_data/input/nclimgrid_lowres_pet.nc --var_name_pet pet --netcdf_awc /Users/caryn.ross/Desktop/climate_indices-master/example_data/input/nclimgrid_lowres_soil.nc --var_name_awc awc --output_file_base /Users/caryn.ross/Desktop/climate_indices-master/example_data/results/palmers/nclimgrid_lowres --calibration_start_year 1951 --calibration_end_year 2010 --multiprocessing all

2020-07-23 09:13:58 INFO Start time: 2020-07-23 09:13:58.668712
2020-07-23 09:13:58 INFO Computing PALMERS
2020-07-23 09:14:18 ERROR Failed to complete

multiprocessing.pool.RemoteTraceback:
"""
Traceback (most recent call last):
File "/opt/miniconda3/envs/palmers/lib/python3.7/site-packages/numba/core/errors.py", line 745, in new_error_context
yield
File "/opt/miniconda3/envs/palmers/lib/python3.7/site-packages/numba/core/lowering.py", line 273, in lower_block
self.lower_inst(inst)
File "/opt/miniconda3/envs/palmers/lib/python3.7/site-packages/numba/core/pylowering.py", line 171, in lower_inst
raise NotImplementedError(type(inst), inst)
NotImplementedError: (<class 'numba.core.ir.EnterWith'>, <numba.core.ir.EnterWith object at 0x7fb4ef75e4e0>)

This appears to be related to numba. Can you please set the environment variable to ignore numba's just-in-time compilation and try again? I.e. do this:

$ export NUMBA_DISABLE_JIT=1

@monocongo it works! thank you!!

OK, great news! A word of caution, the Palmers are the least scientifically reliable (the jury is still out as far as a comparison evaluation vs. the original NOAA PDSI Fortran code on which our implementation is based) and take the longest to compute. In any event if you have other questions let us know here. Happy indexing!