Should we call the infection seeding process "seeding"?

[dylanhmorris]

Minor: Do you think the risk is low that someone misconstrues seed here for reproducibility / something involved in numpyro.sample() vs. interpreting post_seed_latent_infections as the infections after seeded infections $I_{v:0}$ have been incorporated?

I think the risk is low, given that those who might be interested in using this enough to peruse the source code would understand this.

Originally posted by @AFg6K7h4fhy2 in #204 (comment)

[AFg6K7h4fhy2]

Bumping this. Any thoughts here @damonbayer @dylanhmorris?

[damonbayer]

I like it as is, but, if I had to suggest something, it would be to replace "seeding" with "initialization." The main problem with that is ambiguity around "initial infections" which (to me) suggests a single time point, while "seed infections" (to me) could suggest multiple time points.

[dylanhmorris]

post_initialization_latent_infections seems pretty clear to me.

If we ever have ambiguity, initial_infections_vector or initial_infections_timeseries can clarify that it's not necessarily a single timepoint.

[seabbs]

In hhh4 this is called the endemic-epidemic framework (so seeding is endemic). In epiestim it would be called importation I think

[AFg6K7h4fhy2]

In conferring with @damonbayer and @dylanhmorris, we decided that a seemingly good naming convention that MSR can adopt for "infections seeding process" is post_initialization_latent_infections. As such, I will ensure this convention is correctly adopted in the code.