Setup Instructions: Load Forecasts from Local Zoltar sqlite File

Serena Wang, Matthew Cornell, Yuxin David Huang

31 January 2024

Instructions

We now support loading forecast from a local zoltar sqlite file in covidHubUtils. To use this functionality, you need to set source = "local_zoltar" and provide the local_zoltpy_path and zoltar_sqlite_file parameters in load_forecasts().

Please follow these steps to set up the required environment prior to using this functionality for the first time:

Note: This feature is currently beta, so please contact us to create an sqlite file for you that contains the forecasts you want to work with.

1. Install Python 3.6+.
2. Install pipenv globally (i.e., without --user) via:

pip3 install pipenv

3. Set up your R environment so that the PATH environment variable includes both the python3 and pipenv commands. You can read more about R and environment variables here.

4. Clone the zoltpy package from here.

5. In the cloned zoltpy directory, run:

pipenv --three

6. Start your virtual environment via

pipenv shell

7. In your virtual environment, install all required packages with

pipenv install

8. Install R libraries: DBI and RSQLite in Rstudio
9. Test installation
   • Download a zoltar sqlite file form here.
   • Run a test call of covidHubUtils::load_forecasts() with your actual absolute or relative paths. For example:

covidHubUtils::load_forecasts(
  models = "COVIDhub-ensemble", 
  locations = "US", 
  source = "local_zoltar", 
  local_zoltpy_path = "<your_zoltpy_dir>", 
  zoltar_sqlite_file = "<your_db.sqlite3_dir>")