I am trying to run ILAMB using the ACCESS-NRI version. I run it on gadi, calling ilamb-run from a bash shell script. In that script, I first do ‘module use /g/data/xp65/public/modules’ and ‘module load conda/access-med’. Then I run the ilamb-run command. This generates the error/warning messages printed below. Any guidance on how to tackle this would be appreciated. Cheers, Juergen.
/g/data/xp65/public/apps/med_conda/envs/access-med-0.8/lib/python3.10/site-packages/ILAMB/ModelResult.py:138: SyntaxWarning: “is not” with a literal. Did you mean “!=”?
if self.regex is not “”:
Traceback (most recent call last):
File “/g/data/xp65/public/apps/med_conda/envs/access-med-0.8/bin/ilamb-run”, line 44, in
RegisterCustomColormaps()
File “/g/data/xp65/public/apps/med_conda/envs/access-med-0.8/lib/python3.10/site-packages/ILAMB/Post.py”, line 1318, in RegisterCustomColormaps
plt.register_cmap(“score”, cm)
AttributeError: module ‘pylab’ has no attribute ‘register_cmap’
@Juergen, I’ve pinged the MED team. It seems there is an issue with the current version (0.8), they are going to roll back to 0.6 while they investigate.
I’ll look after this ticket for the time being, please let us know how you go once the rollback occurs.
Thank you for bringing this to my attention. It appears there’s an incompatibility between Matplotlib and ILAMB 2.7, which has been addressed by the developers in version 2.7.1. However, this version is not yet available on PyPI or conda-forge. I’ve raised the issue on GitHub here, and we’re working to resolve it as soon as possible.
In the meantime, I’ve reverted the default conda/access-med environment to point to a stable version of the ILAMB-workflow analysis environment. Please let me know if you encounter any further issues. I’ll update the ticket once version 2.7.1 is available.
ILAMB is running now, thanks again for your support. I was trying to run it with GPP, but got a ‘VarsNotComparable’ error. The ILAMB01.log file tells me that ‘Datasets have differing numbers of time intervals: reference = 408, comparison = 409’.
For some reason, ILAMB does not seem to cut the GPP file correctly even though the time axis is correct (at least that’s what I think).
It would be great if you could have a look and check if I am missing something. You can find all the data here:
Hi, @Juergen, I will look into this. but I dont have permissiong to group rp23, is it posiable to move your script to somewhare else so I could access? like in /scratch/public
Thanks @RhaegarZeng . Feel free to use these data as you see fit. You should have permissions for all the files, but let me know if not. The ILAMB log file is under RESULTS and I have also copied the output file of the bash script that run ILAMB to this directory (the *.o file), as well as the script itself (postprocess_TRENDY.sh). Let me know if you need anything else.
Hi, @Juergen, I looked into this issue and found that this is caused by the ClipTime function in ILAMB which didn’t clip the right time range.
In your data, the time in each time stamp was in 15th of each month, but in reference data their are in 16th of each month, I suggest that you change the time in your data to 16th as well, that will make the run successfully. I will also post this issue to ILAMB developer group. the data in /scratch/public/jk8585/ILAMB/MODELS/S3_new5/CABLE-POP_S3_new5_gpp.nc has already been modified, you can use it directly.
Hi @RhaegarZeng . Many thanks for looking into this. I can confirm that it works for GPP when I set the date one day later to the 16th of each month. I will now try the other variables.
Good idea to share this issue with the ILAMB developer group. I had this issue for quite a while already.
Hi, @Juergen:
Sorry for the late reply, really glad to here that your problem has been solved, if there is no other issue, I will close this topic, feel free to post it on the forum if you have further more questions.