some shit

app.py

@@ -21,22 +21,18 @@ _dist_cache: dict = {}
 
 def _ensure_data():
     if not os.path.exists(DATA_PATH):
-        from huggingface_hub import hf_hub_download
-        from huggingface_hub.errors import EntryNotFoundError, RepositoryNotFoundError
+        from huggingface_hub import download_bucket_files
         os.makedirs(_DATA_DIR, exist_ok=True)
         filename = os.path.basename(DATA_PATH)
-        print(f"Downloading {filename} from HF Hub...")
+        print(f"Downloading {filename} from HF bucket...")
         try:
-            hf_hub_download(
-                repo_id=_HF_REPO_ID,
-                filename=filename,
-                repo_type="dataset",
-                local_dir=_DATA_DIR,
-                local_dir_use_symlinks=False,
+            download_bucket_files(
+                _HF_REPO_ID,
+                files=[(filename, DATA_PATH)],
                 token=os.environ.get("HF_TOKEN"),
             )
-        except (EntryNotFoundError, RepositoryNotFoundError):
-            print(f"{filename} not found on HF Hub. Fetching from Baseball Savant...")
+        except Exception:
+            print(f"{filename} not found in bucket. Fetching from Baseball Savant...")
             result = update_season_data()
             if result["status"] != "ok":
                 raise RuntimeError(f"Auto-fetch failed: {result['message']}")
@@ -188,16 +184,17 @@ def _load_cache(today: str, n_simulations: int) -> dict | None:
         return entry["result"]
 
     try:
-        from huggingface_hub import hf_hub_download
-        local_path = hf_hub_download(
-            repo_id=_HF_REPO_ID,
-            filename=_cache_hf_path(today, n_simulations),
-            repo_type="dataset",
-            token=os.environ.get("HF_TOKEN"),
-            force_download=True,
-        )
-        with open(local_path) as f:
-            stored = json.load(f)
+        from huggingface_hub import download_bucket_files
+        remote_path = _cache_hf_path(today, n_simulations)
+        with tempfile.TemporaryDirectory() as tmp_dir:
+            local_path = os.path.join(tmp_dir, "cache.json")
+            download_bucket_files(
+                _HF_REPO_ID,
+                files=[(remote_path, local_path)],
+                token=os.environ.get("HF_TOKEN"),
+            )
+            with open(local_path) as f:
+                stored = json.load(f)
         cached_at = datetime.fromisoformat(stored["cached_at"])
         if cached_at.tzinfo is None:
             cached_at = cached_at.replace(tzinfo=timezone.utc)
@@ -218,15 +215,13 @@ def _save_cache(today: str, n_simulations: int, result: dict) -> None:
     payload = {"cached_at": now.isoformat(), "result": result}
     tmp_path = None
     try:
-        from huggingface_hub import upload_file
+        from huggingface_hub import batch_bucket_files
         with tempfile.NamedTemporaryFile(mode="w", suffix=".json", delete=False) as f:
             json.dump(payload, f)
             tmp_path = f.name
-        upload_file(
-            path_or_fileobj=tmp_path,
-            path_in_repo=_cache_hf_path(today, n_simulations),
-            repo_id=_HF_REPO_ID,
-            repo_type="dataset",
+        batch_bucket_files(
+            _HF_REPO_ID,
+            add=[(tmp_path, _cache_hf_path(today, n_simulations))],
             token=os.environ.get("HF_TOKEN"),
         )
     except Exception as e:
@@ -347,18 +342,14 @@ def update_season_data(season_year: int = 2026) -> dict:
     print(f"Saved {len(df):,} rows to {local_path}")
 
     try:
-        from huggingface_hub import create_repo, upload_file
-        token = os.environ.get("HF_TOKEN")
-        create_repo(repo_id=_HF_REPO_ID, repo_type="dataset", token=token, exist_ok=True)
-        upload_file(
-            path_or_fileobj=local_path,
-            path_in_repo=filename,
-            repo_id=_HF_REPO_ID,
-            repo_type="dataset",
-            token=token,
+        from huggingface_hub import batch_bucket_files
+        batch_bucket_files(
+            _HF_REPO_ID,
+            add=[(local_path, filename)],
+            token=os.environ.get("HF_TOKEN"),
         )
     except Exception as e:
-        return {"status": "error", "message": f"HF Hub upload failed: {e}", "rows": len(df), "filename": filename}
+        return {"status": "error", "message": f"HF bucket upload failed: {e}", "rows": len(df), "filename": filename}
 
     global DATA_PATH
     DATA_PATH = local_path