Skip to content

scepter-quickstart.md

Latest commit

 

History

History
72 lines (49 loc) · 3.75 KB

scepter-quickstart.md

File metadata and controls

72 lines (49 loc) · 3.75 KB
Raw

Setting up SCEPTER simulations

Install dependencies

  1. Follow the instructions in the SCEPTER repo to set up fortran and other SCEPTER dependencies on your machine.

  2. Install python package dependencies in the ew-workflows/environment.yaml file.

Update hard-coded paths

  1. Update the lines to append your ew-workflows/run_scepter path in the following scripts within the SCEPTER repository.

    • rock_buff_minPH_multiyear.py
    • tunespin_4_newton_inert_buff_v2_defaultdict.py
    • singlerun_postproc_detached.py
    • rock_app_singleRun-szn.py
    • rock_buff_dust-ts_multiyear.py

  2. Update the lines to append your ew-workflows/run_scepter path in the following scripts within the ew-workflows/scripts/scepter directory.

    • */setup/batch/make_batch_input_grainsize+apprate.py
    • */setup/batch/make_batch_input_grainsize+apprate_multiyear.py
    • */setup/batch/make_batch_input_grainsize+apprate_cdrPot.py

  3. Update the default directories for the following functions (or update the code that runs the functions to input the correct directory).

    • in ew-workflows/run_scepter/argo_helper_fxns.py
      • maindir in run_multiple function should be /path/to/your/ew-workflows
      • maindir in retry_failed_runs function should be /path/to/your/ew-workflows
    • in ew-workflows/scripts/scepter/run/run-multiple.py
      • maindir in ahf.retry_failed_runs call should be /path/to/your/ew-workflows

Create a batch file

  1. Open ew-workflows/scripts/scepter/setup/batch/make_batch_input_grainsize+apprate.py.

    • This file defines values that will override the defaults defined in the SCEPTER/defaults/dict_singlerun.py

  2. Update system path line if you haven't already.

  3. Update the savepath_batch variable to your ew-workflows/inputs/scepter/batch.

  4. Within const_dict (dictionary of values held constant for all batch simulations) update your aws_save and aws_bucket options. If aws_save = None then the results will be saved locally in SCEPTER/scepter_output/<runname>.

(Argo only)

The following sections only apply if you are using Argo workflows to run batches of SCEPTER simulations.

Update argo parameters

  1. Open ew-workflows/inputs/scepter/params/batch-meanAnnliming_fert_fixedRate-base.yaml and update the following directories (you should only have to update these directories once):

    • control-script-dir
    • batch-input-dir
    • model-dir

  2. Update the batch-input variable to the name of the batch file you created in the last step.

  3. Update the default-dict variable to the name of the default dictionary you want to use in SCEPTER/defaults/dict_singlerun.py (or review the current default dictionary to make sure it's in line with what you want).

Note: we had the most success with argo if we imposed a delay (bleed_delay in the run_multiple function in ew-workflows/run_scepter/argo_helper_fxns.py) between each run being submitted.

Run with argo

  1. In terminal, navigate to ew-workflows/scripts/scepter/run. Activate the virtual environment you created with the ew-workflows/environment.yaml file.

  2. Open run-multiple.py and confirm the run_pars variable is set to the batch-*.yaml file you updated previously.

  3. run python3 run-multiple.py

Postprocessing

Comparing two batch runs

  1. Open ew-workflows/scripts/scepter/process/cdr_feedstock_compare-apprate+psize.ipynb.

  2. Update sys.path.append to your ew-workflows/run_scepter path.

  3. Update dur, fertLevel, dustsp, sitename, multiyear*, tag*, and version* to construct the appropriate values to generate the batch names and filenames.

  4. Update the outdir, savepath, and csv_loc to the appropriate paths.