Merge penn fix conflicts

[sam-writer]

This is just #49 but with 2 commits of mine fixing the merge conflict.

Some notes I am copying from that PR:

This should be merged before #50 or #51 or I worry they will get overwritten.

A lot of these changes are things like adding definitions, and using Length of Stay instead of LOS, as requested by a non-native English speaker working on translating this. However, there is also the change to the denominator of beta, which now includes * (1-relative_contact_rate), which propagates through numbers that feed the charts. So this also affects the analysis.

[quinn-dougherty]

looks good and i tested it locally in my browser

[mdbecker]

hey @sam-qordoba I'm seeing several regressions against the current prod app. I'll start a list of them here:
Prod:

Branch:

Prod:

Branch:

Prod:

Branch:

I'll add more as I find them

[cjbayesian]

A key difference between this and current prod is the user input of prob of detection, rather than inferring it directly from number of hospitalizations, hospitalization %, and market share.

[quinn-dougherty]

I think the discrepency w the data/results are because of The estimated number of currently infected individuals is 1260. (this branch) vs. The estimated number of currently infected individuals is 533. (prod). other than that the calculation might not be off?

i'm currently fixing the input order in the sidebar.

@cjbayesian probability of detection isn't in prod, so where should i put that on the sidebar (between what 2 parameters)? or should i bake in a default and remove it from UI? if there's an earlier version where that parameter was inferred, could you point me to that strip of code so i can restore it?

[cjbayesian]

@quinn-dougherty Thanks. It's https://github.com/pennsignals/chime/blob/f4eb1a5e3e25af105dc76e32e099eedad887f307/app.py#L62 and https://github.com/pennsignals/chime/blob/f4eb1a5e3e25af105dc76e32e099eedad887f307/app.py#L63

There the detection_prob and total_infections are being inferred based on current_hosp, Penn_market_share, hosp_rate, and initial_infections.

[cjbayesian]

So we should remove the probability of detection input from the sidebar. All of the description text is in line with the method of calculating the total number of currently affected individuals that is in current prod.

We can consider changing the way this is being estimated in a future version, but would need to ensure that all description text, documentation, explanations are consistent with the approach.

[cjbayesian]

Side-by-side now looks good to me. @mdbecker to confirm.

[mdbecker]

Checking now. Thanks!

[mdbecker]

Looks great!!!