Open
Conversation
Deleted old validation plot files and uploaded new validation do files.
Persistence.createEntityManagerFactory(...) was called on every invocation of all three methods. Building a factory is expensive — it validates the schema and sets up a connection pool. Two static fields replace this: - emfStartingPopulationPersist — shared by getProcessed() and persistProcessed() (both use getPersistDatabasePath()) - emfStartingPopulationRun — used by loadStartingPopulation() (uses RunDatabasePath) 2. loadStartingPopulation() — bug fix for LazyInitializationException With @onetomany collections already marked FetchType.LAZY, returning detached Household objects after em.close() without having initialised benefitUnits or members would throw LazyInitializationException the moment the caller iterated them. The force-initialisation loop triggers the SUBSELECT batch loads within the open session before closing it. 3. getProcessed() — dead variable removed processed_return was assigned the JOIN FETCH result but the method returned processed (the same L1-cached object). The two are identical references, so processed_return.resetDependents() and processed.resetDependents() were equivalent. The dead variable is removed and the call is chained directly. No behaviour change.
-DonorTaxImputation used Math.random() to decide whether a benefit unit receives Universal Credit based on the weighted mean UC receipt across donor candidates. Math.random() draws from Java's global, unseeded Random instance, producing different results on every JVM run regardless of the model's random seed setting. -Replaced with SimulationEngine.getRnd().nextDouble(), which draws from the model's seeded Random instance — the same one used throughout the rest of the model. With a fixed seed, UC/Legacy benefit assignment is now deterministic across runs.
AlignmentAdjustmentFactors fields (30 total): - Partnership: adj factor, sim share, target share (same computation) - Fertility: adj factor, sim rate, target rate (same computation) - In-school: adj factor, sim share, target share — new - Utility adj factors × 7 (same computation) - Employment shares: 7 × simulated + 7 × target — new
Optimized database hadling and replaced Statistics3 with AlignmentAdjustmentFactors
Remove all the code/references to IT
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Apologies for the confusion last time; it seems I'm still not as fluent with GitHub as I would like to be. Hopefully the new files are now in this branch...
I've copied the text from the previous branch below for reference:
In this branch, I have included the draft programs do files, in which all of the Excel files containing the process estimates (except wages) are produced using the programs file.
A few things worth noting:
The Excel file name is now dynamic and fed into the relevant program.
There are a few different possible processing programs. The one to use depends on the type of model you are running.
Probit and logit models are processed in the same way, even if the probit option is always used.
Review Request:
@andrewbaxter439 - It would be great if you could take a look at the files for the well-being, mental health and financial distress to see if they are working as expected.
Please let me know if you spot any bugs or have any issues.
The one obvious thing that is outstanding to me is that some of the labels are incorrect. Specifically, it is easier to create dummy variables from categorical variables with the name that will be picked up by SimPaths. Lags are dealt with in the code, but category names are not. I think the best place for these variables to be created is in the variable_update do-file. @dariaple is currently updating this file (along with the full set of estimation do files) with the new variable names, so perhaps this is best added after she has completed that work. However, this shouldn't impact the programs themselves it just alters what is fed into them.