Feat/echam streams #1240
Feat/echam streams #1240
Conversation
src/esm_runscripts/echam.py
Outdated
| namelist = f90nml.read(econfig["namelist_dir"] + "/namelist.echam") | ||
| mvstream_tags = _get_mvstream_tags_from_namelist(namelist) | ||
| jsbach_streams = config["jsbach"]["streams"] | ||
| ignore_these_tags = econfig.get("ignore_tags", jsbach_streams) |
There was a problem hiding this comment.
Why is the default ignore ignore jsbach_streams? Is it so that the jsbach streams are moved to the jsbach directory instead of echam?
I find this behavior a bit confusing because if I declare an echam.ignore_tags then the jsbach streams won't be ignored anymore, so there is a big change of behavior behind the scenes. What do you think @pgierz? Maybe I am not understanding this part...
There was a problem hiding this comment.
Good point, see new commit (coming later today)
| return mvstream_tags | ||
|
|
||
|
|
||
| def append_namelist_dependent_sources(config): |
There was a problem hiding this comment.
Shouldn't this happen for echam but also jsbach? Doesn't jsbach also have its own streams?
There was a problem hiding this comment.
This is a little bit convoluted: jsbach and echam are controlled by the same namelist, so I would need some way to filter out which tags go to which model. In principle this would just be defined by which model config they come from.
There was a problem hiding this comment.
jsbach and echam are controlled by the same namelist
Do you mean that they can be controlled from the same namelist? If that is the case, then I understand the rest of your reply and the new jsbach.py
There was a problem hiding this comment.
Yes, the output generated from jsbach is defined in mvstreamctl chapters of namelist.echam (since JSBACH technically speaking is just a submodule inside of ECHAM), and does not have any output control in it's own namelist.
Add suggestions by @mandresm Co-authored-by: Miguel <[email protected]>
|
The latest few commits use the extremely useful feature of being able to write tests directly in your docstrings under the examples section. The included files are referenced directly from the root of the repo in the CI actions. Check it out 😄
|
|
Some of the dryruns in esm_tests are failing |
| return mvstream_tags | ||
|
|
||
|
|
||
| def append_namelist_dependent_sources(config): |
There was a problem hiding this comment.
jsbach and echam are controlled by the same namelist
Do you mean that they can be controlled from the same namelist? If that is the case, then I understand the rest of your reply and the new jsbach.py
|
Once the CI tests work, I would recommend running |
|
Should be good to go. The failing levante tests mainly fail due to changes in how the streams get picked up, as well as new main keys in the echam config, e.g.: Differences in run/awicm/awicm1-CMIP6-initial-monthly//run_18500101-18500131/config//awicm1-CMIP6-initial-monthly_finished_config.yaml:
---
+++
@@ -118,6 +118,7 @@
forcing_dir: /pool/data/ECHAM6//input/r0007/T127
greenhouse_dir: /pool/data/ECHAM6/
namelist_dir: <HOME_DIR>/esm_tools/namelists//echam/6.3.04p1/PI-CTRL
+ has_namelist_streams: true |
Automatically create ECHAM streams from mvstreamctl filetags.
Should be useful for #1239