|
import os |
|
from datasets import load_dataset |
|
from huggingface_hub import hf_hub_download |
|
|
|
def download_data(repo_name: str = "Persius/imr-appeals", output_dir="./data"): |
|
"""Download adjudication data from HF hub, and save it locally in the format expected by HICRIC codebase.""" |
|
train = load_dataset(repo_name, split="train") |
|
test = load_dataset(repo_name, split="test") |
|
train.to_json(os.path.join(output_dir, "outcomes", "train_backgrounds_suff.jsonl")) |
|
test.to_json(os.path.join(output_dir, "outcomes", "test_backgrounds_suff.jsonl")) |
|
|
|
_path = hf_hub_download(repo_id="Persius/imr-appeals", |
|
filename="case-backgrounds.jsonl", |
|
local_dir=output_dir, |
|
subfolder="annotated", |
|
repo_type="dataset") |
|
return None |
|
|
|
if __name__ == "__main__": |
|
download_data() |
