Skip to content

Bug fixes and other adjustments from comparison to Epicast #142

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Open
stevenhofmeyr wants to merge 37 commits into
base: development
Choose a base branch
from
Open

Bug fixes and other adjustments from comparison to Epicast #142

stevenhofmeyr wants to merge 37 commits into from

Conversation

@stevenhofmeyr
Copy link
Collaborator

@stevenhofmeyr stevenhofmeyr commented Dec 3, 2025

This branch includes a number of changes to produce a closer match to Epicast, including a few bug fixes.

  • Changed compliance to be over three days, with different probabilities per age group
  • Changed hospitalization days to use the normal six categories of ages, instead of the three that were used only for this. This clarifies the parameters and reduces the chance of misconfig
    uation.
  • Implemented a set of macros to make implementation of reduction operations with many fields cleaner. Previously, there were four different places where a set of values/variables had to be
    explicitly listed, equal in size to the number of operations. These now can be specified with a single parameter, which is much cleaner and less error prone.
  • Fixed a bug in the way household clusters were allocated and used. This now allocates the HC in the initialization, and uses an additional agent field (hh_cluster) to access the number. I
    t also accesses this value correctly during interactions.
  • Added additional tracking of newly changed counts of infcetions, asymptomatic and presymptomatic agents.
  • Updated use of transitions periods such as incubation, hospital_delay, etc.
  • Ensured hospital_delay is at least 2.
  • Simplified functions for reading parameter lists.
  • Modified updateAgents to implement compliance over three days with hospitalization possible on the second day, as in Epicast.
  • Fixed a bug in reading and writing checkpoints for UrbanPop.
  • Changed allocation of initial infections to only do one at a time, for a better spread over all communities.
  • Changed the calculation of neighborhoods in UrbanPop to give a closer match of the average to the desired neighborhood size parameter.
  • Added helper scripts in utilities to facilitate comparisons with Epicast.
  • Enabled transmission of hospitalized at home, as is done in Epicast.

@stevenhofmeyr
Added various hacks and changes to compare to Epicast:
ab206b6 
- track newly infected
- disable hospital delays
- hacked compliance to match the three day probabilities in Epicast
- hacked hospitalization to only be checked on second day, like in Epicast
- allow household transmission when individuals are hospitalized
@stevenhofmeyr
added python script for comparing epicast and exaepi
5a7000e
@stevenhofmeyr
fixed to properly use the transition probs of epicast
e83e016
@stevenhofmeyr
dump agents in debug mode.
20d1300 
fixed some debug warnings
@stevenhofmeyr
added household cluster and do it per community, not neighborhood
b72f4ad
@stevenhofmeyr
initialize hh cluster for census
fd1b4a6
@stevenhofmeyr
hospitalize on day 2 (as per epicast), using hospital_delay to set
cb0517f
@stevenhofmeyr
used some macros to simplify parallel reductions
3a12c7e
@stevenhofmeyr
fixed issue with dead become immune
c408aa4
@stevenhofmeyr
Create a #define for enabling Epicast comparisons
6f440ad
@stevenhofmeyr
converted quaryArray into a templated function
3527ad9
@stevenhofmeyr
Made withdrawal compliance into arrays for days 0, 1, 2 with differen…
4c3f7a2 
…t probabilities for different age groups.
@stevenhofmeyr
Removed the AgeGroups_Hosp. Confusing and potential for error
56a7318
@stevenhofmeyr
Don't allow hospital treatment on first day
09d11c0
@stevenhofmeyr
Very minor tweaks
236c2f2
@stevenhofmeyr
Fixed incorrect use of fields for Epicast data. Needed the ones with …
935b503 
…ctx_ prepended.
@stevenhofmeyr
include ventilated and icu in hosp counts for comparison to epicast
001271a
@stevenhofmeyr
Don't use epicast hack transition probabilities.
e4889bf 
Added script for estimating gamma parameters
@stevenhofmeyr
transit actions should be commented out
cde880f
@stevenhofmeyr
cleaned up code related to neighborhoods
ddfe642
@stevenhofmeyr
Assign initial infections one at a time to spread more across communi…
81101ab 
…ties
@stevenhofmeyr
A few changes to bring the code more in line with development
54c1eae
@stevenhofmeyr
minor tweak
c0aaaf4
@stevenhofmeyr
set defaults for several parameters to match the Epicast defaults
6dde589
@stevenhofmeyr
fixed GPU build issue
2989b2e
@atmyers
Copy link
Contributor

atmyers commented Dec 16, 2025

There are non-trivial conflicts because of PRs #139 and #140 - let me try to resolve them.

@atmyers
Copy link
Contributor

atmyers commented Dec 16, 2025

I think I've resolved everything - could you make sure the results are still good in your comparison?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@atmyers atmyers Awaiting requested review from atmyers

@debog debog Awaiting requested review from debog

@tannguyen153 tannguyen153 Awaiting requested review from tannguyen153

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

3 participants

@stevenhofmeyr @atmyers