Instructions to get started with using gempyor and flepiMoP, using some provided example configs to help you.
Follow all the steps in the Before any run section to ensure you have access to the correct files needed to run your custom model or a sample model with flepiMoP.
Take note of the location of the directory on your local computer where you cloned the flepiMoP model code (which we'll call FLEPI_PATH), as well as where you cloned your project files (which we'll call PROJECT_PATH).
For example, if you cloned your Github repositories into a local folder called Github and are using flepimop_sample as a project repository, your directory names could be
On Mac:
/Users/YourName/Github/flepiMoP
/Users/YourName/Github/flepimop_sample On Windows: C:\Users\YourName\Github\flepiMoP
C:\Users\YourName\Github\flepimop_sample
Since you'll be navigating frequently between the folder that contains your project code and the folder that contains the core flepiMoP model code, it's helpful to define shortcuts for these file paths. You can do this by creating environmental variables that you can then quickly call instead of writing out the whole file path.
If you're on a Mac or Linux/Unix based operating system, define the FLEPI_PATH and PROJECT_PATH environmental variables to be your directory locations, for example
or, if you have already navigated to your parent directory
You can check that the variables have been set by either typing env to see all defined environmental variables, or typing echo $FLEPI_PATH to see the value of FLEPI_PATH.
If you're on a Windows machine
or, if you have already navigated to your parent directory
You can check that the variables have been set by either typing set to see all defined environmental variables, or typing echo $FLEPI_PATH$ to see the value of FLEPI_PATH.
If you don't want to bother defining the environmental variables, it's no problem, just remember to use the full or relative path names for navigating to the right files or folders in future steps
The code is written in a combination of R and Python. The Python part of the model is a package called gempyor, and includes all the code to simulate the epidemic model and the observational model and apply time-dependent interventions. The R component conducts the (optional) parameter inference, and all the (optional) provided pre and post processing scripts are also written in R. Most uses of the code require interacting with components written in both languages, and thus making sure that both are installed along with a set of required packages. However, Python alone can be used to do forward simulations of the model using gempyor.
First, ensure you have python and R installed. You need a working python3.7+ installation. We recommend using the latest stable python release (python 3.12) to benefit from huge speed-ups and future-proof your installation. We also recommend installing Rstudio to interact with the R code and for exploring your model outputs.
On Mac 🍏
Python 3 is installed by default on recent MacOS installation. If it is not, you might want to check homebrew and install the appropriate installation.
However, this may result in two versions of Python being installed on your computer. If there are multiple versions of Python (e.g., multiple versions of Python 3), you may need to specify which version to use in the installation. This can be done by following the instructions for using a conda environment, in which case the version of Python to use can be specified in the creation of the virtual environment, e.g., conda create -c conda-forge -n flepimop-env python=3.12 numba pandas numpy seaborn tqdm matplotlib click confuse pyarrow sympy dask pytest scipy graphviz emcee xarray boto3 slack_sdk. The conda environment will be activated in the same way and when installing gempyor, the version of pip used in the installation will reflect the Python version used in the conda environment (e.g., 3.12), so you can use pip install -e flepimop/gempyor_pkg/ in this case.
There is also the possibility that multiple versions of gempyor have been installed on your computer in the various iterations of Python. You will only want to have gempyor installed on the latest version of Python (e.g., Python 3.8+) that you have. You can remove a gempyor iteration installed for a given version of Python using pip[version] uninstall gempyor e.g., pip3.7 uninstall gempyor. Then, you will need to specify which version of Python to install gempyor on during that step (see above).
To install the python portions of the code (gempyor ) and all the necessary dependencies, go to the flepiMoP directory and run the installation, using the following commands:
A warning for Windows
Once gempyor is successfully installed locally, you will need to make sure the executable file gempyor-seir.exe is runnable via command line. To do this, you will need to add the directory where it was created to PATH. Follow the instructions here to add the directory where this .exe file is located to PATH. This can be done via GUI or CLI.
If you would like to install gempyor directly from GitHub, go to the flepiMoP directory and use the following command:
If you just want to run a forward simulation, installing python's gempyor is all you need.
To run an inference run and to explore your model outputs using provided post-processing functionality, there are some packages you'll need to install in R. Open your R terminal (at the bottom of RStudio, or in the R IDE), and run the following command to install the necessary R packages:
On Linux
The R packages "sf" and "ggraph" require you to have libgdal-dev and libopenblas-dev installed on your local linux machine.
This step does not need to be repeated unless you use a new computer or delete and reinstall R.
Now return to your system terminal. To install flepiMop-internal R packages, run the following from command line:
After installing the flepiMoP R packages, we need to do one more step to install the command line tools for the inference package. If you are not running in a conda environment, you need to point this installation step to a location that is on your executable search path (i.e., whenever you call a command from the terminal, the places that are searched to find that executable). To find a consistent location, type
The location that is returned will be of the form EXECUTABLE_SEARCH_PATH/gempyor-simulate. Then run the following in an R terminal:
To install the inference package's CLI tools.
Each installation step may take a few minutes to run.
Note: These installations take place on your local operating system. You will need an active internet connection for installation, but not for other steps of running the model. If you are concerned about disrupting your base environment that you use for other projects, try using Docker or conda instead.
Everything is now ready 🎉 .
The next step depends on what sort of simulation you want to run: One that includes inference (fitting model to data) or only a forward simulation (non-inference). Inference is run from R, while forward-only simulations are run directly from the Python package gempyor.
First, navigate to the project folder and make sure to delete any old model output files that are there.
For the following examples we use an example config from flepimop_sample, but you can provide the name of any configuration file you want.
To get started, let's start with just running a forward simulation (non-inference).
Stay in the PROJECT_PATH folder, and run a simulation directly from forward-simulation Python package gempyor. Call gempyor-simulate providing the name of the configuration file you want to run. For example here, we use config_sample_2pop.yml from flepimop_sample.
This will produce a model_output folder, which you can look using provided post-processing functions and scripts.
We recommend using model_output_notebook.Rmd from flepimop_sample as a starting point to interact with your model outputs. First, modify the yaml preamble in the notebook, then knit this markdown. This will produce some nice plots of the prevalence of infection states over time. You can edit this markdown to produce any figures you'd like to explore your model output.
The first time you run all this, it's , it's better to run each command individually as described above to be sure each exits successfully. However, eventually you can put all these steps together in a script, like below
Note that you only have to re-run the installation steps once each time you update any of the files in the flepimop repository (either by pulling changes made by the developers and stored on Github, or by changing them yourself). If you're just running the same or different configuration file, just repeat the final steps
An inference run requires a configuration file that has the inference section. Stay in the $PROJECT_PATH folder, and run the inference script, providing the name of the configuration file you want to run. For example here, we use config_sample_2pop_inference.yml from flepimop_sample.
This will run the model and create a lot of output files in $PROJECT_PATH/model_output/.
The last few lines visible on the command prompt should be:
[[1]]
[[1]][[1]]
[[1]][[1]][[1]]
NULL
If you want to quickly do runs with options different from those encoded in the configuration file, you can do that from the command line, for example
where:
n is the number of parallel inference slots,
j is the number of CPU cores to use on your machine (if j > n, only n cores will actually be used. If j <n, some cores will run multiple slots in sequence)
k is the number of iterations per slots.
Again, it is helpful to run the model output notebook (model_output_notebook.Rmd from flepimop_sample) to explore your model outputs. Knitting this file for an inference run will also provide an analysis of your fits: the acceptance probabilities, likelihoods overtime, and the fits against the provided ground truth.
The first time you run all this, it's , it's better to run each command individually as described above to be sure each exits successfully. However, eventually you can put all these steps together in a script, like below
Note that you only have to re-run the installation steps once each time you update any of the files in the flepimop repository (either by pulling changes made by the developers and stored on Github, or by changing them yourself). If you're just running the same or different configuration file, just repeat the final steps
If your run is successful, you should see your output files in the model_output folder. The structure of the files in this folder is described in the Model Output section. By default, all the output files are .parquet format (a compressed format which can be imported as dataframes using R's arrow package arrow::read_parquet or using the free desktop application Tad for quick viewing). However, you can add the option --write-csv to the end of the commands to run the code (e.g., > gempyor-simulate -c config.yml --write-csv) to have everything saved as .csv files instead ;
These configs and notebooks should be a good starting point for getting started with flepiMoP. To explore other running options, see How to run: Advanced.