Develop specact depend

[rburghol]

@PaulDudaRESPEC
This push has 12 file changes in here, which look a bit voluminous, but the reality is that 2 Are clean-up, removing "test" functions, and old debug print statement (.gitignore and SPECL.py).

The rest are as follows:

• main.py - functionalized a block of code that sets up STATE variables for eligible domain variables (like HYDR and SEDTRN), and called new functions to take their place.
• om.py - added function to do dependency ordering for a given domain, i.e., model_domain_dependencies(), and tidied up some variable creation routines.
• om_model_linkage.py - cleaned up some test code and added a new linkage type (not used in this implementation)
• om_model_object.py - new function to create the end-point for things like RSED4, IVOL etc. Also added a class for ModelRegister tho that is unused at this point.
• om_special_action.py - small change to create a ModelConstant to insure that the dependency endpoint can be found by model_domain_dependencies() (it feels wierd to call somethign a "constant" that is bein changed, but rather than create a superfluous class like 'ModelVariable' I opted to re-use Constant. We can change if it seems wrong).
• HYDR.py - refactored function hydr_init_ix() to clean up arguments
• state.py - code clarity clean up and new function state_init_hsp2() called from main.py
• SEDTRN.py
  • SEDTRN now has the first implementation of the "domain dependency order", so that only those operational components affecting SEDTRN related variables are executed, and refactored function sedtrn_init_ix() to clean up arguments.
• test10.uci - This UCI (based on our email exchange) has 3 SPEC-ACTIONS all pointing at a single reach (RCHRES 5), and thus, only 3 operational model components are executed during the SEDTRN pass through RCHRES 5, and ignored during other
• check_depends_endpoint.py - A new test script added to explore this functionality is included at Single Component Dependency end-point function #133 so that you can see the ordering (this code is also located in tests/testcbp/HSP2Results/check_depends_endpoint.py.

[PaulDudaRESPEC]

Thanks @rburghol , this looks like good stuff. A few comments/requests if you don't mind:

• Instead of modifying test10.uci, let's create a new test10spact.uci with the special actions as you mocked them up. We run test10.uci as a baseline, and I'd like to keep it the same as it has always been. Sorry if my past suggestion wasn't clear.
• In HYDR.py, there's a minor change from njit to njit(cache=True)... was that intentional? If not, let's leave it as cache=True so we don't have an inadvertent change.

[rburghol]

Thanks for the review @PaulDudaRESPEC -- I'll make these uci changes asap. As for the @njit instead of @njit(cache=True), I made that change as a test, and should have communicated about it. My sense is that it was important to do during debugging, as the cache=True argument appeared to be requiring me to reload my environment occasionally. That may have been a misunderstanding on my part -- regardless, I did not intend to leave it in there, though it might be worth seeing if it actually does anything positive and/or if it indeed can cause unwanted caching (I can verify that it does not eliminate pre-compiled speed ups).

[rburghol]

Changes done. Thanks again!

[PaulDudaRESPEC]

Cool, thanks @rburghol . Would you also please revert your changes to Test10.uci?

[rburghol]

Ah - thanks @PaulDudaRESPEC -- in all the switching back and forth I missed that. Done!

[PaulDudaRESPEC]

Perfect, thank you!