Upload maverix.py with huggingface_hub

maverix.py

@@ -1,3 +1,4 @@
+import os
 import subprocess
 import tarfile
 import zipfile
@@ -7,7 +8,6 @@ from urllib.request import urlretrieve
 
 import datasets
 import pandas as pd
-from datasets import config
 from huggingface_hub import hf_hub_download
 
 
@@ -202,8 +202,7 @@ def split_video(input_path: Path, output_dir: Path):
 class MaverixDatasetConfig(datasets.BuilderConfig):
     def __init__(
         self,
-        # metadata_url: Path = "maverix/test.jsonl",
-        metadata_url: str = "maverix/dev.jsonl",
+        metadata_url: Path = "maverix/test.jsonl",
         subtitles_url: str = "subtitle.tar.gz",
         cookies_file: Optional[Path] = None,
         **kwargs,
@@ -267,7 +266,7 @@ class MaverixDataset(datasets.GeneratorBasedBuilder):
         if self.config.metadata_url is None:
             raise ValueError("You must provide metadata_url")
 
-        base_cache_dir = dl_manager.download_config.cache_dir or config.HF_DATASETS_CACHE
+        base_cache_dir = os.path.expanduser(os.getenv("HF_HOME", "~/.cache/huggingface/"))
         cache_dir = Path(base_cache_dir) / "maverix"
         cache_dir.mkdir(parents=True, exist_ok=True)
         
@@ -298,7 +297,6 @@ class MaverixDataset(datasets.GeneratorBasedBuilder):
     # Example generation (lazy download)
     # =========================
     def _generate_examples(self, metadata_path, cache_dir):
-        # TODO: fix the rest of this and test end-to-end
         metadata = pd.read_json(metadata_path, lines=True)
             
         vid_full_dir = cache_dir / "vid_full"