Fix loader path for HF dataset

isonetpp_loader.py

@@ -21,8 +21,8 @@ def _folder_for_size(dataset_size: str) -> str:
 
 def _normalize_name(base_name: str, dataset_size: str) -> str:
     """
-    Accepts 'aids' or 'aids240k'.
-    If user passes bare name, append pairs; if they already passed '...80k/240k', keep as-is.
+    Accepts 'aids' or 'aids240k' (and similarly for other sets).
+    If bare name -> append pairs; if already has 80k/240k -> keep as-is.
     """
     pairs = _pairs_for_size(dataset_size)
     if base_name.endswith(("80k", "240k")):
@@ -31,11 +31,11 @@ def _normalize_name(base_name: str, dataset_size: str) -> str:
 
 def _mode_prefix_and_dir(mode: str) -> tuple[str, str]:
     """
-    File prefix uses 'test' when mode contains 'test' (your repo convention).
-    Directory only has train/val/test. Map Extra_test_300 => 'test'.
+    File prefix uses 'test' when mode contains 'test' (repo convention).
+    Directory has train/val/test. Map Extra_test_300 => 'test'.
     """
     prefix = "test" if "test" in mode.lower() else mode
-    mode_dir = "test" if "test" in mode.lower() else mode  # maps Extra_test_300 -> test
+    mode_dir = "test" if "test" in mode.lower() else mode
     return prefix, mode_dir
 
 # ----------------------------
@@ -45,23 +45,21 @@ def _mode_prefix_and_dir(mode: str) -> tuple[str, str]:
 def _ensure_paths(
     repo_id: str,
     mode: str,
-    dataset_name: str,      # can be 'aids' or 'aids240k'
+    dataset_name: str,      # 'aids' or 'aids240k'
     dataset_size: str,      # 'small' | 'large'
     local_root: Optional[str] = None,
 ) -> Dict[str, str]:
     """
     Download the three files needed into cache (or local_root if set):
-      - splits/<mode_dir>/<prefix>_<base>_query_subgraphs.pkl
-      - splits/<mode_dir>/<prefix>_<base>_rel_nx_is_subgraph_iso.pkl
-      - corpus/<base>_corpus_subgraphs.pkl
-    where <base> is the normalized dataset name (includes 80k/240k exactly once).
+      - large_dataset/splits/<mode_dir>/<prefix>_<base>_query_subgraphs.pkl
+      - large_dataset/splits/<mode_dir>/<prefix>_<base>_rel_nx_is_subgraph_iso.pkl
+      - large_dataset/corpus/<base>_corpus_subgraphs.pkl
+    where <base> is normalized (contains 80k/240k exactly once).
     """
-    pairs = _pairs_for_size(dataset_size)
-    folder = _folder_for_size(dataset_size)           # "large_dataset" or "small_dataset"
-    base = _normalize_name(dataset_name, dataset_size)  # e.g., "aids240k" (no double-append)
+    folder = _folder_for_size(dataset_size)                # "large_dataset" or "small_dataset"
+    base   = _normalize_name(dataset_name, dataset_size)   # e.g., "aids240k"
     prefix, mode_dir = _mode_prefix_and_dir(mode)
 
-    # exact filenames used in your repo
     query_fname  = f"{prefix}_{base}_query_subgraphs.pkl"
     rel_fname    = f"{prefix}_{base}_rel_nx_is_subgraph_iso.pkl"
     corpus_fname = f"{base}_corpus_subgraphs.pkl"
@@ -97,7 +95,7 @@ def load_isonetpp_benchmark(
     device: Optional[str] = None,
     download_root: Optional[str] = None,
 ):
-    # Map user mode to your class constants
+    # Map to class constants
     mode_map = {
         "train": TRAIN_MODE,
         "val": VAL_MODE,
@@ -115,21 +113,27 @@ def load_isonetpp_benchmark(
         local_root=download_root,
     )
 
-    # paths["query"] points to .../<folder>/splits/<mode_dir>/<file>
-    # dataset_base_path must be the parent "<folder>" so that subiso_dataset finds:
-    #   dataset_base_path/{splits/<mode_dir>/..., corpus/...}
-    splits_dir = os.path.dirname(paths["query"])           # .../<folder>/splits/<mode_dir>
-    folder_dir = os.path.dirname(splits_dir)               # .../<folder>
-    dataset_base_path = folder_dir
+    # paths["query"] = .../<folder>/splits/<mode_dir>/<file>
+    # We want dataset_base_path to be the **parent of <folder>** so that:
+    #   dataset_base_path / dataset_path_override / splits/<mode>/... exists
+    # Compute levels carefully:
+    #   file_dir   = .../<folder>/splits/<mode_dir>
+    #   splits_dir = .../<folder>/splits
+    #   folder_dir = .../<folder>
+    #   parent_dir = parent of <folder>
+    file_dir   = os.path.dirname(paths["query"])
+    splits_dir = os.path.dirname(file_dir)
+    folder_dir = os.path.dirname(splits_dir)
+    parent_dir = os.path.dirname(folder_dir)   # <-- this is the correct dataset_base_path
 
     dataset_config = dict(
         mode=mode_norm,
-        dataset_name=_normalize_name(dataset_name, dataset_size),  # ensure '...240k' once
+        dataset_name=_normalize_name(dataset_name, dataset_size),
         dataset_size=dataset_size,
         batch_size=batch_size,
         data_type=data_type,
-        dataset_base_path=dataset_base_path,
-        dataset_path_override=_folder_for_size(dataset_size),      # critical: "large_dataset"/"small_dataset"
+        dataset_base_path=parent_dir,                    # parent of <folder>
+        dataset_path_override=_folder_for_size(dataset_size),  # "large_dataset"/"small_dataset"
         experiment=None,
         device=device,
     )