Update hf_api.py

hf_api.py

@@ -1,34 +1,56 @@
-```py
 import os
+import time
+import requests
+from dataclasses import dataclass
+from typing import Dict, Any, Optional, List, Tuple
+
 from huggingface_hub import HfApi
 
-def require_token(env_name: str = "HF_TOKEN") -> str:
-    token = os.getenv(env_name)
-    if not token:
-        raise RuntimeError(
-            f"Missing {env_name}. Set it in your shell (export {env_name}=...) "
-            f"or as a Hugging Face Space Secret named {env_name}."
-        )
-    return token
+HF_DATASETS_SERVER = "https://datasets-server.huggingface.co
+"
+
+@dataclass
+class HFUploadResult:
+repo_id: str
+path_in_repo: str
+commit_message: str
+commit_url: Optional[str] = None
+
+def require_token() -> str:
+token = os.getenv("HF_TOKEN", "").strip()
+if not token:
+raise RuntimeError(
+"Missing HF_TOKEN. Set it in Space Secrets: Settings → Secrets → New secret → Name=HF_TOKEN."
+)
+return token
 
-def api() -> HfApi:
-    return HfApi(token=require_token())
+def get_splits(dataset: str, token: Optional[str] = None) -> Dict[str, Any]:
+headers = {}
+if token:
+headers["Authorization"] = f"Bearer {token}"
+url = f"{HF_DATASETS_SERVER}/splits?dataset={requests.utils.quote(dataset, safe='')}"
+r = requests.get(url, headers=headers, timeout=30)
+r.raise_for_status()
+return r.json()
 
-def upload_delta_csv(repo_id: str, local_csv_path: str, target_config: str, commit_message: str) -> str:
-    """
-    Uploads a CSV delta into: data/uploads/<target_config>_<timestamp>.csv
-    """
-    import datetime
-    ts = datetime.datetime.utcnow().strftime("%Y%m%d_%H%M%S")
-    filename = f"{target_config}_{ts}.csv"
-    path_in_repo = f"data/uploads/{filename}"
+def get_first_rows(dataset: str, config: str, split: str, token: Optional[str] = None) -> Dict[str, Any]:
+headers = {}
+if token:
+headers["Authorization"] = f"Bearer {token}"
+url = (
+f"{HF_DATASETS_SERVER}/first-rows"
+f"?dataset={requests.utils.quote(dataset, safe='')}"
+f"&config={requests.utils.quote(config, safe='')}"
+f"&split={requests.utils.quote(split, safe='')}"
+)
+r = requests.get(url, headers=headers, timeout=30)
+r.raise_for_status()
+return r.json()
 
-    a = api()
-    a.upload_file(
-        path_or_fileobj=local_csv_path,
-        path_in_repo=path_in_repo,
-        repo_id=repo_id,
-        repo_type="dataset",
-        commit_message=commit_message or f"Upload delta CSV for {target_config}: {filename}",
-    )
-    return path_in_repo
+def upload_csv_delta_to_dataset_repo(
+dataset_repo_id: str,
+local_csv_path: str,
+target_config: str,
+token: Optional[str] = None,
+) -> HFUploadResult:
+"""