Add support for QBOi variables - #179
Conversation
|
@jinboxie here is a development branch for supporting QBOi variables. And i started a template for one variable. I will check if you have permission for this repo and push codes. |
Codecov ReportBase: 28.98% // Head: 28.93% // Decreases project coverage by
Additional details and impacted files@@ Coverage Diff @@
## master #179 +/- ##
==========================================
- Coverage 28.98% 28.93% -0.06%
==========================================
Files 93 93
Lines 3981 3988 +7
==========================================
Hits 1154 1154
- Misses 2827 2834 +7
Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here. ☔ View full report at Codecov. |
|
@jinboxie I updated the utendogw handler. But ran into the issue in cmor: I ran the code on Cori, with |
|
Hi Jill. Yes, the data wanted zonal mean instead of 3D. I reprocessed the data now, please try again. |
|
I made some updates, now it complains about missing lat_bnds. I think somehow latitude bounds are missing from zonal averaging in your processing. |
|
@chengzhuzhang Hi Jill. I created a new file during this postprocess zonal mean calculation. It is now added the new bnds and every variables from the previous files. It should be good now. |
|
Hi @jinboxie it looks like the longitude related data are still present in the data files. I think we picked a hard variable to start, e3sm_to_cmip currently doesn't support zonal mean variables, which is not standard CMIP requests. And implementing it elegantly will take sometime. Another difficulty I have is to actually understand how cmor interact with those QBOi tables, it seems to me cmor is always try to pick up CMIP6 tables, I haven't figured out how to force cmor to pick up other tables. I have to table this after AMS. At the mean time, I think you can generate a mapping of the conversion, In any case that would be useful either we use e3sm_to_cmip or directly use cmor. |
|
@chengzhuzhang ok! |
|
Hi @jinboxie you could look here https://github.com/E3SM-Project/e3sm_to_cmip/pull/179/files for code change I made to temporarily make You could follow below commands to install get this and following is my command to run on the utendogw variable. e3sm_to_cmip -i /global/cscratch1/sd/xie7/forJill/data -o ~/tests -v qboi_utendogw -t /global/homes/c/chengzhu/qboi-cmor-tables/Tables/ -u /global/homes/c/chengzhu/e3sm_to_cmip/e3sm_to_cmip/resources/qboi_template.json You could try patch the lib.py script and fit the qboi variables.. |
Thank you Jill. |
|
@jinboxie no problem. I think the lib.py script is not easy to read and perhaps needs some refactoring. I'm also tagging @tomvothecoder here and we can probably think about a re-write in the long run. |
Hi Jill @chengzhuzhang . I'm having a problem with the command. this was even the same with ua for some reason 0%| | 0/1 [00:00<?, ?it/s] |
|
@jinboxie , based on our discussion. It is best to start from standard e3sm_to_cmip use case, so please download the CMIP tables from here: https://github.com/PCMDI/cmip6-cmor-tables and point to downloaded tables by using |
|
Hi Jill @chengzhuzhang , please refer to https://github.com/E3SM-Project/e3sm_to_cmip/tree/jinboxie_qboi |
|
Thank you, @jinboxie ! |
This PR is to support variables requested by QBOi project, using the customized qboi cmor tables. https://gitlab.com/JamesAnstey/qboi-cmor-tables