Closed rburghol closed 2 months ago
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).
Changes done. Thanks again!
Ah - thanks @PaulDudaRESPEC -- in all the switching back and forth I missed that. Done!
Perfect, thank you!
@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
andSPECL.py
).The rest are as follows:
main.py
- functionalized a block of code that sets up STATE variables for eligible domain variables (likeHYDR
andSEDTRN
), 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 likeRSED4
,IVOL
etc. Also added a class forModelRegister
tho that is unused at this point.om_special_action.py
- small change to create aModelConstant
to insure that the dependency endpoint can be found bymodel_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 functionhydr_init_ix()
to clean up argumentsstate.py
- code clarity clean up and new functionstate_init_hsp2()
called frommain.py
SEDTRN.py
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 othercheck_depends_endpoint.py
- A new test script added to explore this functionality is included at #133 so that you can see the ordering (this code is also located intests/testcbp/HSP2Results/check_depends_endpoint.py
.