formatted code, added setup.py file to install packages

.riahub/workflows/workflow.yaml

@@ -36,6 +36,7 @@ jobs:
         run: |
           python -m pip install --upgrade pip
           pip install h5py numpy
+          pip install -e 
 
       - name: 1. Build HDF5 Dataset
         run: |

scripts/produce_dataset.py

@@ -21,6 +21,7 @@ info_dtype = np.dtype(
     ]
 )
 
+
 def write_hdf5_file(records, output_path, dataset_name="data"):
     """
     Writes a list of records to an HDF5 file.
@@ -74,6 +75,7 @@ def write_hdf5_file(records, output_path, dataset_name="data"):
 
     return output_path
 
+
 def split_recording(recording_list):
     """
     Splits a list of recordings into smaller chunks.

scripts/split_dataset.py

@@ -18,14 +18,12 @@ def split(dataset, train_frac=0.8, seed=42):
 
     by_rec = defaultdict(list)
     for i, (_, md) in enumerate(dataset):
-        by_rec[md['rec_id']].append(i)
-    
+        by_rec[md["rec_id"]].append(i)
 
     rec_ids = list(by_rec.keys())
     random.seed(seed)
     random.shuffle(rec_ids)
 
-    
     train_set = set()
     count = 0
     for rec_id in rec_ids:
@@ -34,8 +32,6 @@ def split(dataset, train_frac=0.8, seed=42):
             train_set.update(index)
             count += len(index)
 
-
-    
     validation_set = set(range(N)) - train_set
 
     print(f"Train set :{len(train_set)}")
@@ -45,5 +41,3 @@ def split(dataset, train_frac=0.8, seed=42):
     val_records = [dataset[i] for i in sorted(validation_set)]
 
     return train_records, val_records
-
-

setup.py

@@ -0,0 +1,10 @@
+from setuptools import setup, find_packages
+
+setup(
+    name="modrec_workflow",
+    version="0.1",
+    packages=find_packages(),  # this will pick up `utils/` (so utils/__init__.py must exist)
+    install_requires=[
+        # runtime dependencies go here (if any)
+    ],
+)