|
|
|
|
|
|
|
import argparse |
|
import os |
|
|
|
from mattergen.common.data.dataset import CrystalDataset |
|
from mattergen.common.globals import PROJECT_ROOT |
|
|
|
if __name__ == "__main__": |
|
parser = argparse.ArgumentParser() |
|
parser.add_argument( |
|
"--csv-folder", |
|
type=str, |
|
required=True, |
|
help="Path to the folder containing the csv files. All csv files in the folder will be processed (e.g., 'train.csv', 'val.csv', 'test.csv') and the resulting datasets will be placed under {cache_path/dataset_name/filename_without_extension}, e.g, /path/to/project/dataset/mp_20/train.", |
|
) |
|
parser.add_argument( |
|
"--dataset-name", |
|
type=str, |
|
required=True, |
|
help="Name of the dataset (e.g. mp_20. Will be used to create a folder in the cache folder)", |
|
) |
|
parser.add_argument( |
|
"--cache-folder", |
|
type=str, |
|
required=True, |
|
default=f"{PROJECT_ROOT}/datasets", |
|
help="Path to the cache folder. Defaults to datasets folder in the project root.", |
|
) |
|
args = parser.parse_args() |
|
for file in os.listdir(f"{args.csv_folder}"): |
|
if file.endswith(".csv"): |
|
print(f"Processing {args.csv_folder}/{file}") |
|
CrystalDataset.from_csv( |
|
csv_path=f"{args.csv_folder}/{file}", |
|
cache_path=f"{args.cache_folder}/{args.dataset_name}/{file.split('.')[0]}", |
|
) |
|
|
