Rename pipeline.sh > calibrate.sh

docs/src/experiment_setup_guide.md

@@ -13,10 +13,10 @@ For the example experiment, `sphere_held_suarez_rhoe_equilmoist`, this is done b
 `sbatch experiments/sphere_held_suarez_rhoe_equilmoist/generate_observations.sbatch`. This script runs the model, passes the output through the observation map, and saves the result.
 
 Once the observations have been processed and saved, the actual calibration pipeline can be run via
-`bash pipeline.sh sphere_held_suarez_rhoe_equilmoist -n 10 -c 8`.
+`bash calibrate.sh sphere_held_suarez_rhoe_equilmoist -n 10 -c 8`.
 
 !!! note
-    The command line interface for `pipeline.sh` will change. For now, the first entry is the experiment id and the second is the number of tasks to use per ensemble member.
+    The command line interface for `calibrate.sh` will change. To obtain detailed usage information, run `bash calibrate.sh --help`.
 
 This consists of three `sbatch` scripts:
 - `initialize.sbatch` initializes the Julia project and the ensemble

docs/src/quickstart.md

@@ -12,7 +12,7 @@ By default, it runs 10 ensemble members for 3 iterations.
 To run this experiment:
 1. Log onto the Caltech HPC
 2. Clone CalibrateAtmos.jl and `cd` into the repository.
-3. Run: `bash pipeline.sh -n 10 -c 8 sphere_held_suarez_rhoe_equilmoist`. This will run the `sphere_held_suarez_rhoe_equilmoist` experiment with 10 tasks per ensemble member.
+3. Run: `bash calibrate.sh -n 10 -c 8 sphere_held_suarez_rhoe_equilmoist`. This will run the `sphere_held_suarez_rhoe_equilmoist` experiment with 10 tasks per ensemble member.
 
 ## Local Machine
 To run an experiment on your local machine, you can use the `pipeline.jl` script. This is recommended for more lightweight experiments, such as the `surface_fluxes_perfect_model` experiment, which uses the [SurfaceFluxes.jl](https://github.com/CliMA/SurfaceFluxes.jl) package to generate a physical model that calculates the Monin Obukhov turbulent surface fluxes based on idealized atmospheric and surface conditions. Since this is a "perfect model" example, the same model is used to generate synthetic observations using its default parameters and a small amount of noise. These synthetic observations are considered to be the ground truth, which is used to assess the model ensembles' performance when parameters are drawn from the prior parameter distributions. To run this experiment, you can use the following command from terminal to run an interactive run:
@@ -24,7 +24,7 @@ julia -i pipeline.jl surface_fluxes_perfect_model
 This pipeline mirrors the pipeline of the bash srcipts, and the same example can be run on the HPC cluster if needed:
 
 ```bash
-bash pipeline.sh surface_fluxes_perfect_model 8
+bash calibrate.sh surface_fluxes_perfect_model 8
 ```
 
 The experiments (such as `surface_fluxes_perfect_model`) can be equally defined within the component model repos (in this case, `SurfaceFluxes.jl`), so that the internals of `CalibrateAtmos.jl` do not explicitly depend on component models.

slurm/parse_commandline.sh

@@ -5,7 +5,7 @@ slurm_cpus_per_task="1"
 slurm_gpus_per_task="0"
 generate_data=false
 help_message="Usage:
-    ./pipeline.sh [options] experiment_id
+    ./calibrate.sh [options] experiment_id
 
 Options:
     -t, --time=HH:MM:SS: Set max wallclock time (default: 2:00:00).