AGNI no-solve configuration and minor changes #334

nichollsh · 2025-02-11T10:43:30Z

Add option for running AGNI without solver, just prescribing a TP profile. This is very fast and could easily be put into a test or an example for the JOSS paper. Closes AGNI no-solve configuration #332.
As part of the above, removes the global atm variable into the Proteus class.
Store runtime info in helpfile. Closes Store model wall-clock runtime in helpfile #331.
Minor fixes to plots and docs.
Option for disabling the 'strict' check on the convergence criteria, which requires them to be met for two consecutive iterations.
Time stepping optimisations. The model checks if it is close to approaching solidification / radeqm in order to make sure that the time-step does not exceed the expected point of model convergence. However, this only matters when these criteria are enabled. The model now does not check for this if the criteria are disabled.

…strict convergence check not passed. Added wall-clock runtime to helpfile

… test the model easily even if new columns have been added.

timlichtenberg · 2025-02-11T12:49:55Z

input/demos/spider.toml

Why no demo SPIDER file anymore?

I created it for demo-ing the tidal heating stuff, but that works well now and can be tested through the other configs if necessary.

timlichtenberg · 2025-02-11T12:52:39Z

input/planets/k218b.toml

I feel we have to pick up the discussion again how to handle the config files. Since the number of parameters keeps rising, it seems to be really annoying to edit all files everytime... :> So the solution may be to instead have one "default" file and store only the parameters that are changed from the default in the specific demo cases. This will require some tweaking of how config files are read. This may actually make a good issue for @HannoSpreeuw to consider sometime in early March. What do you think @lsoucasse?

Good point @timlichtenberg !

I agree. We decided before to not have a default one, because it could be confusing, but it feels like we have to go through and edit a bunch every time someone changes the code.

I think it should be quite easy to do though

So the idea would be to call a specific .toml file, which only contains the parameters that are different from default.toml, correct? In this case, both the "case.toml" and "default.toml" file should be copy-pasted into the simulation directory, so that afterwards it is crystal clear which parameters were used.

Yeah I think that would be a good solution. As long as it's all compiled into a file that the user can check, it should hopefully avoid issues.

EmmaPostolec

Looks good to me, nice job with the double check for solidification/termination @nichollsh !

timlichtenberg · 2025-02-11T12:59:23Z

src/proteus/proteus.py

@@ -107,11 +110,19 @@ def start(self, *, resume: bool = False, offline: bool = False):
            Run in offline mode; do not try to connect to the internet.
        """

-        # Import
-        from proteus.atmos_clim import RunAtmosphere


Define some common nomenclature in Docs regarding function names etc?

They keep changing from CapitalLetter to under_score_naming and back.

The code was originally written with CapitalLetter format, but we later decided that the standard should be lower_case_with_underscores. I think the latter is what Python is written in.

But yes, I agree we should add this in the "contributing" section of the docs.

timlichtenberg · 2025-02-11T13:01:08Z

src/proteus/utils/terminate.py

@@ -149,6 +149,14 @@ def check_termination(handler: Proteus) -> bool:

    # Quantify model state at this iteration
    finished = False
+    handler.finished2 = False


"finished2" and "finished1" are difficult to parse and understand variable names. Something more descriptive?

I'll improve this.

timlichtenberg · 2025-02-11T13:01:42Z

tests/helpers/helpers.py

@@ -5,6 +5,7 @@
 from PIL import Image

 PROTEUS_ROOT = Path(__file__).parents[2]
+NEGLECT = ["CH4_mol_atm", "CH4_mol_total", "CH4_kg_atm", "CH4_kg_total", "CH4_bar", "runtime"]


Why only these?

In this PR I have just moved this functionality to a new file. We should try to keep the list of neglected keys as short as possible. CH4 is included because it's very sensitive to the numerics. Runtime is included because it depends on the performance of the computer used.

timlichtenberg

A few comments that may lead to new issues.

nichollsh · 2025-02-11T13:10:50Z

Thanks both! I'll make some minor changes and re-request a review.

timlichtenberg

Great, improves the readability and contribution instructions a lot!

nichollsh · 2025-02-11T13:50:15Z

Thank you!

nichollsh and others added 10 commits February 4, 2025 09:53

Add option for toggling strict convergence check

87d77db

Improved timestepping near tidal regime. Update status to Running on …

37940cc

…strict convergence check not passed. Added wall-clock runtime to helpfile

Timestep optimsations based on termination criteria

99f4997

Update plot label

deb7e79

AGNI no-solve configuration

a509aeb

AGNI no-solve config with condensation

1211f15

Merge branch 'main' into hn/dev

e0b658e

Only compare overlapping columns in test dataframe. This allows us to…

8fdd6d3

… test the model easily even if new columns have been added.

Merge branch 'hn/dev' of github.com:FormingWorlds/PROTEUS into hn/dev

cb3bbac

Update test following new time-stepping

9f04296

nichollsh marked this pull request as ready for review February 11, 2025 11:14

nichollsh added 2 commits February 11, 2025 12:15

Style

f3d68a3

Style2

47706e8

nichollsh requested review from timlichtenberg and EmmaPostolec February 11, 2025 12:01

timlichtenberg reviewed Feb 11, 2025

View reviewed changes

EmmaPostolec approved these changes Feb 11, 2025

View reviewed changes

timlichtenberg reviewed Feb 11, 2025

View reviewed changes

timlichtenberg approved these changes Feb 11, 2025

View reviewed changes

Renamed variables for clarity. Info on code style in docs

24315ff

nichollsh requested review from timlichtenberg and EmmaPostolec February 11, 2025 13:42

timlichtenberg approved these changes Feb 11, 2025

View reviewed changes

nichollsh merged commit ed649c4 into main Feb 11, 2025
5 checks passed

nichollsh deleted the hn/dev branch February 11, 2025 13:50

EmmaPostolec approved these changes Feb 11, 2025

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

AGNI no-solve configuration and minor changes #334

AGNI no-solve configuration and minor changes #334

nichollsh commented Feb 11, 2025 •

edited

Loading

timlichtenberg Feb 11, 2025

nichollsh Feb 11, 2025

timlichtenberg Feb 11, 2025

EmmaPostolec Feb 11, 2025

nichollsh Feb 11, 2025

nichollsh Feb 11, 2025

timlichtenberg Feb 11, 2025

nichollsh Feb 11, 2025

EmmaPostolec left a comment •

edited

Loading

timlichtenberg Feb 11, 2025

timlichtenberg Feb 11, 2025 •

edited

Loading

nichollsh Feb 11, 2025

nichollsh Feb 11, 2025

timlichtenberg Feb 11, 2025

nichollsh Feb 11, 2025

timlichtenberg Feb 11, 2025

nichollsh Feb 11, 2025

timlichtenberg left a comment

nichollsh commented Feb 11, 2025

timlichtenberg left a comment

nichollsh commented Feb 11, 2025

AGNI no-solve configuration and minor changes #334

AGNI no-solve configuration and minor changes #334

Conversation

nichollsh commented Feb 11, 2025 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

EmmaPostolec left a comment • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

timlichtenberg Feb 11, 2025 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

timlichtenberg left a comment

Choose a reason for hiding this comment

nichollsh commented Feb 11, 2025

timlichtenberg left a comment

Choose a reason for hiding this comment

nichollsh commented Feb 11, 2025

nichollsh commented Feb 11, 2025 •

edited

Loading

EmmaPostolec left a comment •

edited

Loading

timlichtenberg Feb 11, 2025 •

edited

Loading