refactor(utils): 重构数据处理脚本

2025-07-15 13:10:38 +08:00
parent e6f5b26a75
commit e77fb56c45
5 changed files with 20 additions and 21 deletions
@@ -1,14 +1,18 @@
+from datasets import load_dataset
 from utils import process_dataset

 if __name__ == "__main__":
    max_chunk_num = 10
    chunk_size = 1000000
-    item_size = max_chunk_num * chunk_size
+    
+    dataset = load_dataset(
+        "opencsg/chinese-cosmopedia", 
+        data_files=[f"0000{i}.parquet" for i in range(5)]
+    )
    
    process_dataset(
-        dataset_name="opencsg/chinese-cosmopedia",
+        dataset_dict=dataset,
        output_subdir="chinese-wiki",
-        data_files=[f"0000{i}.parquet" for i in range(5)],
        max_chunk_num=max_chunk_num,
        chunk_size=chunk_size,
    )