Update Argument Parser to use action="store_true" instead of 0/1 for …

…boolean arguments. (#72)

I have replaced the parser.add_argument where the standard has been to
have an integer 0=False, 1=True with an action="store_true" instead.

Example:
OLD:
    parser.add_argument(
        "--restore_opt",
        type=int,
        default=0,
        help="If optimizer state should be restored with model "
        "(default: 0 (false))",
    )
NEW:
    parser.add_argument(
        "--restore_opt",
        action="store_true",
        help="If optimizer state should be restored with model "
        "(default: false)",
    )

This will save some time and characters when running the scripts from
the command line as well as being easier to understand as the parsed
variables are supposed to be booleans.

CHANGELOG.md

@@ -35,7 +35,11 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Changed
 
-  Optional multi-core/GPU support for statistics calculation in `create_parameter_weights.py`
+- Argument Parser updated to use action="store_true" instead of 0/1 for boolean arguments.
+  (https://github.com/mllam/neural-lam/pull/72)
+  @ErikLarssonDev
+
+-  Optional multi-core/GPU support for statistics calculation in `create_parameter_weights.py`
   [\#22](https://github.com/mllam/neural-lam/pull/22)
   @sadamov
 

README.md

@@ -118,7 +118,7 @@ Run `python -m neural_lam.create_mesh` with suitable options to generate the gra
 The graphs used for the different models in the [paper](https://arxiv.org/abs/2309.17370) can be created as:
 
 * **GC-LAM**: `python -m neural_lam.create_mesh --graph multiscale`
-* **Hi-LAM**: `python -m neural_lam.create_mesh --graph hierarchical --hierarchical 1` (also works for Hi-LAM-Parallel)
+* **Hi-LAM**: `python -m neural_lam.create_mesh --graph hierarchical --hierarchical` (also works for Hi-LAM-Parallel)
 * **L1-LAM**: `python -m neural_lam.create_mesh --graph 1level --levels 1`
 
 The graph-related files are stored in a directory called `graphs`.

neural_lam/create_mesh.py

@@ -169,10 +169,9 @@ def main(input_args=None):
     )
     parser.add_argument(
         "--plot",
-        type=int,
-        default=0,
+        action="store_true",
         help="If graphs should be plotted during generation "
-        "(default: 0 (false))",
+        "(default: False)",
     )
     parser.add_argument(
         "--levels",
@@ -182,9 +181,8 @@ def main(input_args=None):
     )
     parser.add_argument(
         "--hierarchical",
-        type=int,
-        default=0,
-        help="Generate hierarchical mesh graph (default: 0, no)",
+        action="store_true",
+        help="Generate hierarchical mesh graph (default: False)",
     )
     args = parser.parse_args(input_args)
 

neural_lam/create_parameter_weights.py

@@ -156,9 +156,8 @@ def main():
     )
     parser.add_argument(
         "--distributed",
-        type=int,
-        default=0,
-        help="Run the script in distributed mode (1) or not (0) (default: 0)",
+        action="store_true",
+        help="Run the script in distributed mode (default: False)",
     )
     args = parser.parse_args()
     distributed = bool(args.distributed)

neural_lam/train_model.py

@@ -41,10 +41,9 @@ def main(input_args=None):
     )
     parser.add_argument(
         "--subset_ds",
-        type=int,
-        default=0,
+        action="store_true",
         help="Use only a small subset of the dataset, for debugging"
-        "(default: 0=false)",
+        "(default: false)",
     )
     parser.add_argument(
         "--seed", type=int, default=42, help="random seed (default: 42)"
@@ -71,10 +70,9 @@ def main(input_args=None):
     )
     parser.add_argument(
         "--restore_opt",
-        type=int,
-        default=0,
+        action="store_true",
         help="If optimizer state should be restored with model "
-        "(default: 0 (false))",
+        "(default: false)",
     )
     parser.add_argument(
         "--precision",
@@ -118,11 +116,10 @@ def main(input_args=None):
     )
     parser.add_argument(
         "--output_std",
-        type=int,
-        default=0,
+        action="store_true",
         help="If models should additionally output std.-dev. per "
         "output dimensions "
-        "(default: 0 (no))",
+        "(default: False (no))",
     )
 
     # Training options
@@ -135,10 +132,9 @@ def main(input_args=None):
     )
     parser.add_argument(
         "--control_only",
-        type=int,
-        default=0,
+        action="store_true",
         help="Train only on control member of ensemble data "
-        "(default: 0 (False))",
+        "(default: False)",
     )
     parser.add_argument(
         "--loss",
@@ -233,7 +229,7 @@ def main(input_args=None):
             pred_length=args.ar_steps,
             split="train",
             subsample_step=args.step_length,
-            subset=bool(args.subset_ds),
+            subset=args.subset_ds,
             control_only=args.control_only,
         ),
         args.batch_size,
@@ -247,7 +243,7 @@ def main(input_args=None):
             pred_length=max_pred_length,
             split="val",
             subsample_step=args.step_length,
-            subset=bool(args.subset_ds),
+            subset=args.subset_ds,
             control_only=args.control_only,
         ),
         args.batch_size,
@@ -313,7 +309,7 @@ def main(input_args=None):
                     pred_length=max_pred_length,
                     split="test",
                     subsample_step=args.step_length,
-                    subset=bool(args.subset_ds),
+                    subset=args.subset_ds,
                 ),
                 args.batch_size,
                 shuffle=False,

plot_graph.py

@@ -38,9 +38,8 @@ def main():
     )
     parser.add_argument(
         "--show_axis",
-        type=int,
-        default=0,
-        help="If the axis should be displayed (default: 0 (No))",
+        action="store_true",
+        help="If the axis should be displayed (default: False)",
     )
 
     args = parser.parse_args()

tests/test_mllam_dataset.py

@@ -118,7 +118,7 @@ def test_load_reduced_meps_dataset(meps_example_reduced_filepath):
 def test_create_graph_reduced_meps_dataset():
     args = [
         "--graph=hierarchical",
-        "--hierarchical=1",
+        "--hierarchical",
         "--data_config=data/meps_example_reduced/data_config.yaml",
         "--levels=2",
     ]