# Environment Variables

***

### description: >- A library of environment variables in the flepiMoP codebase. These variables may be updated or deprecated as the project evolves.

## Environment Variables

Below you will find a list of environment variables (envvars) defined throughout the flepiMoP codebase. Often, these variables are set in response to command-line argument input. Though, some are set by `flepiMoP` without direct user input (these are denoted by a 'Not a CLI option' note in the **Argument** column.)

<table><thead><tr><th width="135">Envvar.</th><th width="100">Argument</th><th width="300">Description</th><th width="115">Default</th><th width="150">Valid values</th><th width="150">Key file locations (inexhaustive)</th></tr></thead><tbody><tr><td><code>BATCH_SYSTEM</code></td><td>Not a CLI option.</td><td>System you are running on (e.g., aws, SLURM, local).</td><td>N/A</td><td>e.g., <code>aws</code>, <code>slurm</code></td><td><code>inference_job_launcher.py</code></td></tr><tr><td><code>CENSUS_API_KEY</code></td><td>Not a CLI option.</td><td>A unique key to the API for census data.</td><td>N/A</td><td><a href="https://api.census.gov/data/key_signup.html">Get your own API key</a></td><td><code>slurm_init.sh</code>, <code>build_US_setup.R</code></td></tr><tr><td><code>CONFIG_PATH</code></td><td><code>-c</code>, <code>--config</code></td><td>Path to a configuration file.</td><td>--</td><td><code>your/path/to/config_file</code></td><td><code>build_covid_data.R</code>, <code>build_US_setup.R</code>, <code>build_initial_seeding.R</code>, <code>build_flu_data.R</code>, <code>config.R</code>, <code>preprocessing/</code> files</td></tr><tr><td><code>DELPHI_API_KEY</code></td><td><code>-d</code>, <code>--delhpi_api_key</code></td><td>Your personalized key for the Delphi Epidata API. Alternatively, this key can go in the config inference section as <code>gt_api_key</code>.</td><td>--</td><td><a href="https://cmu-delphi.github.io/delphi-epidata/api/api_keys.html">Get your own API key</a></td><td><code>build_covid_data.R</code></td></tr><tr><td><code>DIAGNOSTICS</code></td><td><code>-n</code>, <code>--run-diagnostics</code></td><td>Flag for whether or not diagnostic tests should be run during execution.</td><td><code>TRUE</code></td><td><code>--run-diagnostics FALSE</code> for <code>FALSE</code>, <code>--run-diagnostics</code> or no mention for <code>TRUE</code></td><td><code>run_sim_processing_SLURM.R</code></td></tr><tr><td><code>DISEASE</code></td><td><code>-i</code>, <code>--disease</code></td><td>Which disease is being simulated in the prsent run.</td><td><code>flu</code></td><td>e.g., <code>rsv</code>, <code>covid</code></td><td><code>run_sim_processing_SLURM.R</code>/td></td></tr><tr><td><code>DVC_OUTPUTS</code></td><td>Not a CLI option, but defined using <code>--output</code></td><td>The names of the directories with outputs to save in S3 (separated by a space).</td><td><code>model_output model_parameters importation hospitalization</code></td><td>e.g., <code>model_output model_parameters importation hospitalization</code></td><td><code>scenario_job.py</code>, <code>AWS_scenario_runner.sh</code></td></tr><tr><td><code>FILENAME</code></td><td>Not a CLI option.</td><td>Filenames for output files, determined dynamically during inference.</td><td>N/A</td><td><code>file.parquet</code>, <code>plot.pdf</code></td><td><code>AWS_postprocess_runner.sh</code>, <code>SLURM_inference_job.run</code>, <code>AWS_inference_runner.sh</code></td></tr><tr><td><code>FIRST_SIM_INDEX</code></td><td><code>-i</code>, <code>--first_sim_index</code></td><td>The index of the first simulation.</td><td><code>1</code></td><td><code>int</code></td><td><code>shared_cli.py</code></td></tr><tr><td><code>FLEPI_BLOCK_INDEX</code></td><td><code>-b</code>, <code>--this_block</code></td><td>Index of current block.</td><td><code>1</code></td><td><code>int</code></td><td><code>flepimop-inference-main.R</code>, <code>utils.py</code>, <code>AWS_postprocess_runner.sh</code>, <code>AWS_inference_runner.sh</code>, <code>SLURM_inference_job.run</code>, <code>inference_job_launcher.py</code></td></tr><tr><td><code>FLEPI_CONTINUATION</code></td><td><code>--continuation</code>/<code>--no-continuation</code></td><td>Flag for whether or not to use the resumed run seir files (or provided initial files bucket) as initial conditions for the next run.</td><td><code>FALSE</code></td><td><code>--continuation TRUE</code> for <code>TRUE</code>, <code>--continuation</code> or no mention for <code>FALSE</code></td><td><code>SLURM_inference_job.run</code>, <code>inference_job_launcher.py</code></td></tr><tr><td><code>FLEPI_CONTINUATION_FTYPE</code></td><td>Not a CLI option.</td><td>If running a continuation, the file type of the initial condition files.</td><td><code>config['initial_conditions']['initial_file_type']</code></td><td>e.g., <code>.csv</code></td><td><code>SLURM_inference_job.run</code>, <code>inference_job_launcher.py</code></td></tr><tr><td><code>FLEPI_CONTINUATION_LOCATION</code></td><td><code>--continuation-location</code></td><td>The location (folder or an S3 bucket) from which to pull the /init/ files (if not set, uses the resume location seir files).</td><td>--</td><td><code>path/to/your/location</code></td><td><code>SLURM_inference_job.run</code>, <code>inference_job_launcher.py</code></td></tr><tr><td><code>FLEPI_CONTINUATION_RUN_ID</code></td><td><code>--continuation-run-id</code></td><td>The ID of run to continue at, if doing a continuation.</td><td>--</td><td><code>int</code></td><td><code>SLURM_inference_job.run</code>, <code>inference_job_launcher.py</code></td></tr><tr><td><code>FLEPI_INFO_PATH</code></td><td>Not a CLI option.</td><td><em>pending</em></td><td><em>pending</em></td><td><em>pending</em></td><td><code>info.py</code></td></tr><tr><td><code>FLEPI_ITERATIONS_PER_SLOT</code></td><td><code>-k</code>, <code>--iterations_per_slot</code></td><td>Number of iterations to run per slot.</td><td>--</td><td><code>int</code></td><td><code>flepimop-inference-slot.R</code>, <code>flepimop-inference-main.R</code>, <code>SLURM_inference_job.run</code>, <code>inference_job_launcher.py</code></td></tr><tr><td><code>FLEPI_MAX_STACK_SIZE</code></td><td><code>--stacked-max</code></td><td>Maximum number of iterventions to allow in a stacked intervention.</td><td><code>5000</code></td><td><code>int >=350</code></td><td><code>StackedModifier.py</code>, <code>inference_job_launcher.py</code></td></tr><tr><td><code>FLEPI_MEM_PROFILE</code></td><td><code>-M</code>, <code>--memory_profiling</code></td><td>Flag for whether or not memory profile should be run during iterations.</td><td><code>FALSE</code></td><td><code>--memory_profiling TRUE</code> for <code>TRUE</code>, <code>--memory_profiling</code> or no mention for <code>FALSE</code></td><td><code>flepimop-inference-slot.R</code>, <code>flepimop-inference-main.R</code>, <code>inference_job_launcher.py</code></td></tr><tr><td><code>FLEPI_MEM_PROF_ITERS</code></td><td><code>-P</code>, <code>--memory_profiling_iters</code></td><td>If doing memory profiling, after every X iterations, run the profiling.</td><td><code>100</code></td><td><code>int</code></td><td><code>flepimop-inference-slot.R</code>, <code>flepimop-inference-main.R</code>, <code>inference_job_launcher.py</code></td></tr><tr><td><code>FLEPI_NJOBS</code></td><td><code>-j</code>, <code>--jobs</code></td><td>Number of parallel processors used to run the simulation. If there are more slots than jobs, slots will be divided up between processors and run in series on each.</td><td>Number of cores detected as available at computing cluster.</td><td><code>int</code></td><td><code>flepimop-inference-slot.R</code>, <code>flepimop-inference-main.R</code>, <code>calibrate.py</code></td></tr><tr><td><code>FLEPI_NUM_SLOTS</code></td><td><code>-n</code>, <code>--slots</code></td><td>Number of independent simulations of the model to be run.</td><td>--</td><td><code>int >=1</code></td><td><code>flepimop-inference-slot.R</code>, <code>flepimop-inference-main.R</code>, <code>calibrate.py</code>, <code>inference_job_launcher.py</code></td></tr><tr><td><code>FLEPI_OUTCOME_SCENARIOS</code></td><td><code>-d</code>, <code>--outcome_modifiers_scenarios</code></td><td>Name of the outcome scenario to run.</td><td><code>'all'</code></td><td><em>pending</em></td><td><code>flepimop-inference-slot.R</code>, <code>flepimop-inference-main.R</code>, <code>SLURM_inference_job.run</code>, <code>inference_job_launcher.py</code></td></tr><tr><td><code>FLEPI_PATH</code></td><td><code>-p</code>, <code>--flepi_path</code></td><td>Path to the flepiMoP directory.</td><td><code>'flepiMoP'</code></td><td><code>path/to/flepiMoP</code></td><td>several <code>postprocessing/</code> files, several <code>batch/</code> files, several <code>preprocessing/</code> files, <code>info.py</code>, <code>utils.py</code>, <code>_cli.py</code></td></tr><tr><td><code>FLEPI_PREFIX</code></td><td><code>--in-prefix</code></td><td>Unique name for the run.</td><td>--</td><td>e.g., <code>project_scenario1_outcomeA</code>, etc.</td><td><code>SLURM_inference_job.run</code>, <code>inference_job_launcher.py</code>, <code>AWS_postprocess_runner.sh</code>, <code>calibrate.py</code>, several <code>preprocessing/</code> files, several <code>postprocessing/</code> files, several <code>batch/</code> files</td></tr><tr><td><code>FLEPI_RESET_CHIMERICS</code></td><td><code>-L</code>, <code>--reset_chimeric_on_accept</code></td><td>Flag for whether or not chimeric parameters should be reset to global parameters whena global acceptance occurs.</td><td><code>TRUE</code></td><td><code>--reset_chimeric_on_accept FALSE</code> for <code>FALSE</code>, <code>--reset_chimeric_on_accept</code> or no mention for <code>TRUE</code></td><td><code>flepimop-inference-slot.R</code>, <code>flepimop-inference-main.R</code>, <code>slurm_init.sh</code>, <code>hpc_init</code>, <code>inference_job_launcher.py</code></td></tr><tr><td><code>FLEPI_RESUME</code></td><td><code>--resume</code>/<code>--no-resume</code></td><td>Flag for whether or not to resume the current calibration.</td><td><code>FALSE</code></td><td><code>--resume TRUE</code> for <code>TRUE</code>, <code>--resume</code> or no mention for <code>FALSE</code></td><td><code>flepimop-inference-slot.R</code>, <code>flepimop-inference-main.R</code>, <code>slurm_init.sh</code>, <code>hpc_init</code>, <code>inference_job_launcher.py</code></td></tr><tr><td><code>FLEPI_RUN_INDEX</code></td><td><code>-u</code>, <code>--run_id</code></td><td>Unique ID given to the model run. If the same config is run multiple times, you can avoid the output being overwritten by using unique model run IDs.</td><td>Auto-assigned run ID</td><td><code>int</code></td><td><code>copy_for_continuation.py</code>, <code>flepimop-inference-slot.R</code>, <code>flepimop-inference-main.R</code>, <code>shared_cli.py</code>, <code>base.py</code>, <code>calibrate.py</code>, several <code>batch/</code> files, several <code>postprocessing/</code> files</td></tr><tr><td><code>FLEPI_SEIR_SCENARIOS</code></td><td><code>-s</code>, <code>--seir_modifier_scenarios</code></td><td>Names of the intervention scenarios to run.</td><td><code>'all'</code></td><td><em>pending</em></td><td><code>flepimop-inference-slot.R</code>, <code>flepimop-inference-main.R</code>, <code>inference_job_launcher.py</code></td></tr><tr><td><code>FLEPI_SLOT_INDEX</code></td><td><code>-i</code>, <code>--this_slot</code></td><td>Index for current slots.</td><td><code>1</code></td><td><code>int</code></td><td><code>flepimop-inference-slot.R</code>, several <code>batch/</code> files</td></tr><tr><td><code>FS_RESULTS_PATH</code></td><td><code>-R</code>, <code>--results-path</code></td><td>A path to the model results.</td><td>--</td><td><code>your/path/to/model_results</code></td><td><code>prune_by_llik.py</code>, <code>prune_by_llik_and_proj.py</code>, several <code>postprocessing/</code> files, several <code>batch/</code> files, <code>model_output_notebook.Rmd</code></td></tr><tr><td><code>FULL_FIT</code></td><td><code>-F</code>, <code>--full-fit</code></td><td>Whether or not to process the full fit.</td><td><code>FALSE</code></td><td><code>--full-fit TRUE</code> for <code>TRUE</code>, <code>--full-fit</code> or no mention for <code>FALSE</code></td><td><code>run_sim_processing_SLURM.R</code></td></tr><tr><td><code>GT_DATA_SOURCE</code></td><td><code>-s</code>, <code>--gt_data_source</code></td><td>Sources of groundtruth data.</td><td><code>'csse_case, fluview_death, hhs_hosp'</code></td><td>See default</td><td><code>build_covid_data.R</code></td></tr><tr><td><code>GT_END_DATE</code></td><td><code>--ground_truth_end</code></td><td>Last date to include ground truth for.</td><td>--</td><td><code>YYYY-MM-DD</code> format</td><td><code>flepimop-inference-slot.R</code>, <code>flepimop-inference-main.R</code></td></tr><tr><td><code>GT_START_DATE</code></td><td><code>--ground_truth_start</code></td><td>First date to include ground truth for.</td><td>--</td><td><code>YYYY-MM-DD</code> format</td><td><code>flepimop-inference-slot.R</code>, <code>flepimop-inference-main.R</code></td></tr><tr><td><code>IMM_ESC_PROP</code></td><td><code>--imm_esc_prop</code></td><td>Annual percent of immune escape.</td><td><code>0.35</code></td><td><code>float</code> between <code>0.00 - 1.00</code></td><td>several <code>preprocessing/</code> files</td></tr><tr><td><code>INCL_AGGR_LIKELIHOOD</code></td><td><code>-a</code>, <code>--incl_aggr_likelihood</code></td><td>Whether or not the likelihood should be calculated with aggregate estimates.</td><td><code>FALSE</code></td><td><code>--incl_aggr_likelihood TRUE</code> for <code>TRUE</code>, <code>--incl_aggr_likelihood</code> or no mention for <code>FALSE</code></td><td><code>flepimop-inference-slot.R</code></td></tr><tr><td><code>IN_FILENAME</code></td><td>Not a CLI option.</td><td>Name of input files.</td><td>N/A</td><td><code>file_1.csv</code> <code>file_2.csv</code>, etc.</td><td>several <code>batch/</code> files</td></tr><tr><td><code>INIT_FILENAME</code></td><td><code>--init_file_name</code></td><td>Initial file global intermediate name.</td><td>--</td><td><code>file.csv</code></td><td><code>seir_init_immuneladder.R</code>, <code>inference_job.run</code>, several <code>preprocessing/</code> files</td></tr><tr><td><code>INTERACTIVE_RUN</code></td><td><code>-I</code>, <code>--is-interactive</code></td><td>Whether or not the current run is interactive.</td><td><code>FALSE</code></td><td><code>--is-interactive TRUE</code> for <code>TRUE</code>, <code>--is-interactive</code> or no mention for <code>FALSE</code></td><td><code>flepimop-inference-slot.R</code>, <code>flepimop-inference-main.R</code></td></tr><tr><td><code>JOB_NAME</code></td><td><code>--job-name</code></td><td>Unique job name (intended for use when submitting to SLURM).</td><td>--</td><td>Convention: <code>{config['name']}-{timestamp}</code> (str)</td><td>several <code>batch/</code> files</td></tr><tr><td><code>LAST_JOB_OUTPUT</code></td><td>Not a CLI option.</td><td>Path to output of last job.</td><td>N/A</td><td><code>path/to/last_job/output</code></td><td><code>utils.py</code>, several <code>batch/</code> files</td></tr><tr><td><code>OLD_FLEPI_RUN_INDEX</code></td><td>Not a CLI option.</td><td>Run ID of old flepiMoP run.</td><td>N/A</td><td><code>int</code></td><td>several <code>batch/</code> files</td></tr><tr><td><code>OUT_FILENAME</code></td><td>Not a CLI option.</td><td>Name of output files.</td><td>N/A</td><td><code>file_1.csv</code> <code>file_2.csv</code>, etc.</td><td>several <code>batch/</code> files</td></tr><tr><td><code>OUT_FILENAME_DIR</code></td><td>Not a CLI option.</td><td>Directory for output files.</td><td>N/A</td><td><code>path/to/output/files</code></td><td><code>SLURM_inference_job.run</code></td></tr><tr><td><code>OUTPUTS</code></td><td><code>-o</code>, <code>--select-outputs</code></td><td>A list of outputs to plot.</td><td><code>'hosp, hnpi, snpi, llik'</code></td><td><code>hosp, hnpi, snpi, llik</code></td><td><code>postprocess_snapshot.R</code></td></tr><tr><td><code>PARQUET_TYPES</code></td><td>Not a CLI option.</td><td>Parquet files.</td><td><code>'seed spar snpi seir hpar hnpi hosp llik init'</code></td><td><code>seed spar snpi seir hpar hnpi hosp llik init</code></td><td><code>AWS_postprocess_runner.sh</code>, <code>SLURM_inference_job.run</code>, <code>AWS_inference_runner.sh</code></td></tr><tr><td><code>PATH</code></td><td>Not a CLI option.</td><td>Path relating to AWS installation. Used during SLURM runs.</td><td>N/A</td><td>set with <code>export PATH=~/aws-cli/bin:$PATH</code> in <code>SLURM_inference_job.run</code></td><td><code>schema.yml</code>, <code>utils.py</code>, <code>info.py</code>, <code>AWS_postprocess_runner.sh</code>, <code>SLURM_inference_job.run</code></td></tr><tr><td><code>PROCESS</code></td><td><code>-r</code>, <code>--run-processing</code></td><td>Whether or not to process the run.</td><td><code>FALSE</code></td><td><code>--run-processing TRUE</code> for <code>TRUE</code>, <code>--run-processing</code> or no mention for <code>FALSE</code></td><td><code>run_sim_processing_SLURM.R</code></td></tr><tr><td><code>PROJECT_PATH</code></td><td><code>-d</code>, <code>--data_path</code></td><td>Path to the folder with configs and model output.</td><td>--</td><td><code>path/to/configs_and_model-output</code></td><td><code>base.py</code>, <code>_cli.py</code>, <code>calibrate.py</code>, several <code>postprocessing/</code> files, several <code>batch/</code> files</td></tr><tr><td><code>PULL_GT</code></td><td><code>-g</code>, <code>--pull-gt</code></td><td>Whether or not to pull ground truth data.</td><td><code>FALSE</code></td><td><code>--pull-gt TRUE</code> for <code>TRUE</code>, <code>--pull-gt</code> or no mention for <code>FALSE</code></td><td><code>run_sm_processing_SLURM.R</code></td></tr><tr><td><code>PYTHON_PATH</code></td><td><code>-y</code>, <code>--python</code></td><td>Path to Python executable.</td><td><code>'python3'</code></td><td><code>path/to/your_python</code></td><td><code>flepimop-inference-slot.R</code>, <code>flepimop-inference-main.R</code></td></tr><tr><td><code>RESUMED_CONFIG_PATH</code></td><td><code>--res_config</code></td><td>Path to previous config file, if using resumes.</td><td><code>NA</code></td><td><code>path/to/past_config</code></td><td><code>seir_init_immuneladder.R</code>, several <code>preprocessing/</code> files</td></tr><tr><td><code>RESUME_DISCARD_SEEDING</code></td><td><code>--resume-discard-seeding</code>, <code>--resume-carry-seeding</code></td><td>Whether or not to keep seeding in resume runs.</td><td><code>FALSE</code></td><td><code>--resume-carry-seeding TRUE</code> for <code>TRUE</code>, <code>--resume-carry-seeding</code> or no mention for <code>FALSE</code></td><td>several <code>batch/</code> files</td></tr><tr><td><code>RESUME_LOCATION</code></td><td><code>-r</code>, <code>--restart-from-location</code></td><td>The location (folder or an S3 bucket) where the previous run is stored.</td><td>--</td><td><code>path/to/last_job/output</code></td><td><code>built_initial_seeding.R</code>, <code>calibrate.py</code>, <code>slurm_init.sh</code>, <code>hpc_init</code>, <code>inference_job_launcher.py</code></td></tr><tr><td><code>RESUME_RUN</code></td><td><code>-R</code>, <code>--is-resume</code></td><td>Whether or not this run is a resume.</td><td><code>FALSE</code></td><td><code>--is-a-resume TRUE</code> for <code>TRUE</code>, <code>--is-a-resume</code> or no mention for <code>FALSE</code></td><td><code>flepimop-inference-slot.R</code>, <code>flepimop-inference-main.R</code></td></tr><tr><td><code>RESUME_RUN_INDEX</code></td><td>Not a CLI option.</td><td>Index of resumed run.</td><td>set by <code>OLD_FLEPI_RUN_INDEX</code></td><td><code>int</code></td><td><code>SLURM_inference_job.run</code></td></tr><tr><td><code>RSCRIPT_PATH</code></td><td><code>-r</code>, <code>--rpath</code></td><td>Path to R executable.</td><td><code>'Rscript'</code></td><td><code>path/to/your_R</code></td><td><code>build_initial_seeding.R</code>, <code>flepimop-inference-slot.R</code>, <code>flepimop-inference-main.R</code></td></tr><tr><td><code>RUN_INTERACTIVE</code></td><td><code>-I</code>, <code>--is-interactive</code></td><td>Whether or not the current run is interactive.</td><td><code>FALSE</code></td><td><code>--is-interactive</code> for <code>TRUE</code>, <code>--is-interactive</code> or no mention for <code>FALSE</code></td><td><code>flepimop-inference-slot.R</code>, <code>flepimop-inference-main.R</code></td></tr><tr><td><code>SAVE_HOSP</code></td><td><code>-H</code>, <code>--save_hosp</code></td><td>Whether or not the HOSP output files should be saved for each iteration.</td><td><code>TRUE</code></td><td><code>--save_hosp FALSE</code> for <code>FALSE</code>, <code>--save_hosp</code> or no mention for <code>TRUE</code></td><td><code>flepimop-inference-slot.R</code>, <code>flepimop-inference-main.R</code></td></tr><tr><td><code>SAVE_SEIR</code></td><td><code>-S</code>, <code>--save_seir</code></td><td>Whether or not the SEIR output files should be saved for each iteration.</td><td><code>FALSE</code></td><td><code>--save_seir TRUE</code> for <code>TRUE</code>, <code>--save_seir</code> or no mention for <code>FALSE</code></td><td><code>flepimop-inference-slot.R</code>, <code>flepimop-inference-main.R</code></td></tr><tr><td><code>SEED_VARIANTS</code></td><td><code>-s</code>, <code>--seed_variants</code></td><td>Whether or not to add variants/subtypes to outcomes in seeding.</td><td>--</td><td><code>FALSE</code>, <code>TRUE</code></td><td><code>create_seeding.R</code></td></tr><tr><td><code>SIMS_PER_JOB</code></td><td>Not a CLI option.</td><td>Simulations per job.</td><td>N/A</td><td><code>int >=1</code></td><td><code>AWS_postprocess_runner.sh</code>, <code>inference_job_launcher.py</code>, <code>AWS_inference_runner.sh</code></td></tr><tr><td><code>SLACK_CHANNEL</code></td><td><code>-s</code>, <code>--slack-channel</code></td><td>Slack channel, either 'csp-production' or 'debug'; or 'noslack' to disable slack.</td><td>--</td><td><code>csp-production</code>, <code>debug</code>, or <code>noslack</code></td><td><code>postrpocess_auto.py</code>, <code>postprocessing-scripts.sh</code>, <code>inference_job_launcher.py</code></td></tr><tr><td><code>SLACK_TOKEN</code></td><td><code>-s</code>, <code>--slack-token</code></td><td>Slack token.</td><td>--</td><td><a href="https://api.slack.com/tutorials/tracks/getting-a-token">How to get a Slack token</a></td><td><code>postprocess_auto.py</code>, <code>SLURM_postprocess_runner.run</code></td></tr><tr><td><code>SUBPOP_LENGTH</code></td><td><code>-g</code>, <code>--subpop_len</code></td><td>Number of digits in subpops.</td><td><code>5</code></td><td><code>int</code></td><td><code>flepimop-inference-slot.R</code>, <code>flepimop-inference-main.R</code></td></tr><tr><td><code>S3_MODEL_PROJECT_PATH</code></td><td>Not a CLI option.</td><td>Location in S3 bucket with the code, data, and dvc pipeline.</td><td>N/A</td><td><code>path/to/code_data_dvc</code></td><td>several <code>batch/</code> files</td></tr><tr><td><code>S3_RESULTS_PATH</code></td><td>Not a CLI option.</td><td>Location in S3 to store results.</td><td>N/A</td><td><code>path/to/s3/results</code></td><td>several <code>batch/</code> files</td></tr><tr><td><code>S3_UPLOAD</code></td><td>Not a CLI option.</td><td>Whether or not we also save runs to S3 for slurm runs</td><td><code>TRUE</code></td><td><code>TRUE</code>, <code>FALSE</code></td><td><code>SLURM_postprocess_runner.run</code>, <code>SLURM_inference_job.run</code>, <code>inference_job_launcher.py</code></td></tr><tr><td><code>VALIDATION_DATE</code></td><td><code>--validation-end-date</code></td><td>First date of projection/forecast (first date without ground truth data).</td><td><code>date.today()</code></td><td><code>YYYY-MM-DD</code> format</td><td><code>data_setup_source.R</code>, <code>DataUtils.R</code>, <code>groundtruth_source.R</code>, <code>slurm_init.sh</code>, <code>hpc_init</code>, <code>inference_job_launcher.py</code></td></tr></tbody></table>


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://iddynamics.gitbook.io/flepimop/how-to-run/environment-variables.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.