BayesCHIME

Many factors surrounding the transmission, severity of infections, and remaining susceptibility of local populations to COVID-19 remain highly uncertain. However, as new data on hospitalized cases becomes available, we wish to incorporate this data in order to update and refine our projections of future demand to better inform capacity planners. To that end we have extended CHIME to increase the epidemiological process realism and to coherently incorporate new data as it becomes available. This extension allows us to transition from a small number of scenarios to assess best and worst case projections based on parameter assumptions, to a probabilistic forecast representing a continuous distribution of likely scenarios.

Installation

We recommend using the anaconda python distribution which comes with most of the libraries you will need to run this application. Additional dependencies should be installed by running:

pip install -r requirements.txt

Usage

The best way to run sims is via a bash script. This allows for easy configurability of arguments. Here is an example script:

#!/bin/bash

# Set the arguments
chains=8
n_iters=5000
burn_in=2000
reopen_day=100
reopen_speed=.05
reopen_cap=.5 


for loc in 'CCH' 'LGH' 'PAH' 'Downtown' 'HUP' 'PMC' 'MCP'
do 
	echo $loc
	echo $chains
	python _01_GOF_sims.py -p data/"$loc"_parameters.csv -t data/"$loc"_ts.csv \
	-C $chains \
	-b \
	-f \
	-i $n_iters \
	-B $burn_in \
	-pp \
	--reopen_day $reopen_day \
	--reopen_speed $reopen_speed \
	--reopen_cap $reopen_cap \
	--prefix $loc
done

Arguments to the command-line interface:

usage: _01_GOF_sims.py [-h] [-c MY_CONFIG] [-P PREFIX] [-p PARAMETERS] [-t TS]
                       [-C N_CHAINS] [-i N_ITERS] [-f] [--penalty PENALTY]
                       [-s] [-o OUT] [-a AS_OF] [-b] [-v] [-B BURN_IN]
                       [-d N_DAYS] [-y Y_MAX] [-pp] [--reopen_day REOPEN_DAY]
                       [--reopen_speed REOPEN_SPEED] [--reopen_cap REOPEN_CAP]

Args that start with '--' (eg. -P) can also be set in a config file (specified
via -c). Config file syntax allows: key=value, flag=true, stuff=[a,b,c] (for
details, see syntax at https://goo.gl/R74nmi). If an arg is specified in more
than one place, then commandline values override config file values which
override defaults.

optional arguments:
  -h, --help            show this help message and exit
  -c MY_CONFIG, --my-config MY_CONFIG
                        config file path
  -P PREFIX, --prefix PREFIX
                        prefix for old-style inputs
  -p PARAMETERS, --parameters PARAMETERS
                        the path to the parameters csv
  -t TS, --ts TS        the path to the time-series csv
  -C N_CHAINS, --n_chains N_CHAINS
                        number of chains to run
  -i N_ITERS, --n_iters N_ITERS
                        number of iterations to run per chain
  -f, --fit_penalty     fit the penalty based on the last week of data
  --penalty PENALTY     penalty factor used for shrinkage (0.05 - 1)
  -s, --sample_obs      adds noise to the values in the time-series
  -o OUT, --out OUT     output directory
  -a AS_OF, --as_of AS_OF
                        number of days in the past to project from
  -b, --flexible_beta   flexible, vs simple, logistic represetation of beta
  -v, --verbose         verbose output
  -B BURN_IN, --burn_in BURN_IN
                        how much of the burn-in to discard
  -d N_DAYS, --n_days N_DAYS
                        make a census/admits plot out to n_days
  -y Y_MAX, --y_max Y_MAX
                        max y-scale for the census graph
  -pp, --plot_pairs     Plot posterior samples in a pair-plot grid
  --reopen_day REOPEN_DAY
                        day at which to commence evaluating the reopen
                        function
  --reopen_speed REOPEN_SPEED
                        how fast to reopen
  --reopen_cap REOPEN_CAP
                        how much reopening to allow 0: being fully open, 1: being no change in current social distancing

Docker

For anyone struggling with getting these python scripts running on their own machine, we've provided a Dockerfile to help you along with the process. This requires that you setup Docker on your machine. For Mac and Windows you'll need Docker Desktop. On linux you can use your package manager to install Docker.

The easy way

The easy way to get things going is through docker-compose

Build the container

docker-compose build

Run the container

docker-compose up

By default this will use the CCH files in the data directory. If you'd like to run this with a different set of input files you can override this using the $LOC environment variable. This can be done one of two ways

1. export LOC=NewLoc
2. Modify the .env file

The LOC env variable will be passed to full_pipe.sh. If you need to modify what is passed to the scripts you can modify full_pipe.sh accordingly

The more flexible way

The more flexible way to run the app in Docker is using the docker command

Build the container

docker build -t chime_sims .

Run the container

docker run -it --rm -v `pwd`:/chime_sims chime_sims /bin/bash

Note: You may have to replace `pwd` with the path to the chime_sims directory.

To run the container in Docker on Windows, you must enable sharing from the local drive where the chime_sims directory exists. The following steps are for Windows 10 and Docker Desktop version 2.2.0.5.

1. Right click on the Docker icon in the system tray and select settings.
2. Click Resources and navigate to File Sharing.
3. Select the checkbox for the appropriate local drive.
4. Click Apply & Restart (you may be prompted to enter your credentials). To run the container, insert the file path to the chime_sims directory on your local drive: docker run -it --rm -v InsertFilePath:/chime_sims chime_sims /bin/bash

Run the scripts

Once the container starts you should be put into a bash shell where you can then run scripts as described above.

Note: You may need to increase resources to run the scripts. To do so, right click on the Docker icon in the system tray and select settings then click Resources and adjust as needed.