Datasets:
Modalities:
3D
Size:
10K<n<100K
commit files to HF hub
Browse files- windtunnel_dataset.py +63 -0
windtunnel_dataset.py
ADDED
@@ -0,0 +1,63 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import json
|
2 |
+
import os
|
3 |
+
import datasets
|
4 |
+
|
5 |
+
class WindtunnelDataset(datasets.GeneratorBasedBuilder):
|
6 |
+
"""Dataset for loading simulation data with JSON and mesh files"""
|
7 |
+
|
8 |
+
def _info(self):
|
9 |
+
return datasets.DatasetInfo(
|
10 |
+
features=datasets.Features({
|
11 |
+
'coeff': datasets.Value('dict'), # JSON file as a dictionary
|
12 |
+
'input': datasets.Value('dict'), # JSON file as a dictionary
|
13 |
+
'input_mesh': datasets.Value('binary'), # OBJ mesh file as binary
|
14 |
+
'openfoam_mesh': datasets.Value('binary'), # OBJ mesh file as binary
|
15 |
+
'pressure_field_mesh': datasets.Value('binary'),# VTK file as binary
|
16 |
+
'streamlines_mesh': datasets.Value('binary'), # PLY file as binary
|
17 |
+
})
|
18 |
+
)
|
19 |
+
|
20 |
+
def _split_generators(self, dl_manager):
|
21 |
+
"""Define the splits for the dataset."""
|
22 |
+
metadata_path = os.path.join(self.config.data_dir, 'metadata.json')
|
23 |
+
with open(metadata_path, 'r') as f:
|
24 |
+
metadata = json.load(f)
|
25 |
+
|
26 |
+
# Split the dataset into train, validation, and test
|
27 |
+
train_ids, val_ids, test_ids = self._train_val_test_split(metadata)
|
28 |
+
|
29 |
+
return [
|
30 |
+
datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={'metadata': {id: metadata[id] for id in train_ids}}),
|
31 |
+
datasets.SplitGenerator(name=datasets.Split.VALIDATION, gen_kwargs={'metadata': {id: metadata[id] for id in val_ids}}),
|
32 |
+
datasets.SplitGenerator(name=datasets.Split.TEST, gen_kwargs={'metadata': {id: metadata[id] for id in test_ids}}),
|
33 |
+
]
|
34 |
+
|
35 |
+
def _generate_examples(self, metadata):
|
36 |
+
"""Generate examples for each split."""
|
37 |
+
for sim_id, files in metadata.items():
|
38 |
+
yield sim_id, {
|
39 |
+
'coeff': self._load_json(files['coeff']),
|
40 |
+
'input': self._load_json(files['input']),
|
41 |
+
'input_mesh': self._load_binary(os.path.join(self.config.data_dir, sim_id, 'input_mesh.obj')),
|
42 |
+
'openfoam_mesh': self._load_binary(os.path.join(self.config.data_dir, sim_id, 'openfoam_mesh.obj')),
|
43 |
+
'pressure_field_mesh': self._load_binary(os.path.join(self.config.data_dir, sim_id, 'pressure_field_mesh.vtk')),
|
44 |
+
'streamlines_mesh': self._load_binary(os.path.join(self.config.data_dir, sim_id, 'streamlines_mesh.ply')),
|
45 |
+
}
|
46 |
+
|
47 |
+
def _load_json(self, file_path):
|
48 |
+
"""Load a JSON file and return it as a dictionary."""
|
49 |
+
with open(file_path, 'r') as f:
|
50 |
+
return json.load(f)
|
51 |
+
|
52 |
+
def _load_binary(self, file_path):
|
53 |
+
"""Load a binary file and return its contents."""
|
54 |
+
with open(file_path, 'rb') as f:
|
55 |
+
return f.read()
|
56 |
+
|
57 |
+
def _train_val_test_split(self, metadata):
|
58 |
+
"""Implement logic for splitting metadata into train, validation, and test sets."""
|
59 |
+
keys = list(metadata.keys())
|
60 |
+
train_ids = keys[:int(0.7 * len(keys))]
|
61 |
+
val_ids = keys[int(0.7 * len(keys)):int(0.85 * len(keys))]
|
62 |
+
test_ids = keys[int(0.85 * len(keys)):]
|
63 |
+
return train_ids, val_ids, test_ids
|