Error traces back to line 468 in compute_locations.py
Reproduced and identified error using master and tiny-data of celavi-data, fresh install following instructions in this repo's README
groupby.mean() is trying to average columns containing string values, resulting in the error
Fix: Specify that only the lat and long columns should be averaged during this groupby
Related fix: Pull out only the locations of power plants to calculate average lat/longs (by state), instead of all facility type locations
Related fix: At least one to_csv in scenario.py (line 518) has an unexpected keyword argument "line_terminator" - looks like this keyword changed to "lineterminator" at some point, per pandas.to_csv documentation.
After fixing lineterminator, now postprocessing hangs indefinitely after generating count_cumulative_histories.csv ...
Tracked down the hanging to line 164 of diagnostic_viz.py where the figure is written to file. Appears to be the same issues as here: https://community.plotly.com/t/static-image-export-hangs-using-kaleido/61519/4
The fix as per the webpage is to install Kaleido 0.1.0 (via conda, that's conda install python-kaleido==0.1.0) in the environment
(in the environment installed using the repo instructions, Kaleido 0.2.0 gets installed)
Had to use the conda-forge channel and it took an unnerving amount of time to solve the environment, but installation was successful
Almost, figure creation no longer hangs but there's one more line_terminator keyword in costgraph.py
Before marking this solved and merging, I'll create a brand new environment with a fresh celavi install to make sure the kaleido issue was cleared up in 61d475c - and test again on tiny-data
Revised the exact version (0.1.0.post1 rather than 0.1.0) and successfully completed a full tiny-data run
rjhanes
commented
7 months ago
Error traces back to line 468 in compute_locations.py
Reproduced and identified error using
masterandtiny-dataof celavi-data, fresh install following instructions in this repo's READMEgroupby.mean()is trying to average columns containing string values, resulting in the errorFix: Specify that only the
latandlongcolumns should be averaged during thisgroupbyRelated fix: Pull out only the locations of power plants to calculate average lat/longs (by state), instead of all facility type locations
Related fix: At least one
to_csvin scenario.py (line 518) has an unexpected keyword argument "line_terminator" - looks like this keyword changed to "lineterminator" at some point, perpandas.to_csvdocumentation.After fixing lineterminator, now postprocessing hangs indefinitely after generating count_cumulative_histories.csv ...
Tracked down the hanging to line 164 of diagnostic_viz.py where the figure is written to file. Appears to be the same issues as here: https://community.plotly.com/t/static-image-export-hangs-using-kaleido/61519/4 The fix as per the webpage is to install Kaleido 0.1.0 (via conda, that's
conda install python-kaleido==0.1.0) in the environment(in the environment installed using the repo instructions, Kaleido 0.2.0 gets installed)
Had to use the conda-forge channel and it took an unnerving amount of time to solve the environment, but installation was successful
Almost, figure creation no longer hangs but there's one more
line_terminatorkeyword in costgraph.pySuccessfully completed simulation, diagnostic graphs created:
Before marking this solved and merging, I'll create a brand new environment with a fresh celavi install to make sure the kaleido issue was cleared up in 61d475c - and test again on tiny-data
Revised the exact version (0.1.0.post1 rather than 0.1.0) and successfully completed a full tiny-data run