Speed up: cache repo file list; Dasheng dropdown exclude fold*

dataset_loader.py

@@ -16,10 +16,21 @@ REPO_TYPE = "dataset"
 ROOT_PREFIX = "batch_outputs/"
 DASHENG_PREFIX = "batch_outputs_dasheng/"
 
+# Cache full repo file list so we only call list_repo_files once per process (major speedup)
+_cached_repo_files: Optional[list[str]] = None
+
+
+def _get_repo_files() -> list[str]:
+    """Return full list of repo file paths, cached after first call."""
+    global _cached_repo_files
+    if _cached_repo_files is None:
+        _cached_repo_files = list_repo_files(REPO_ID, repo_type=REPO_TYPE)
+    return _cached_repo_files
+
 
 def _get_sample_ids(prefix: str = ROOT_PREFIX) -> list[str]:
     """List sample IDs (e.g. 07_003277) under given prefix in repo."""
-    files = list_repo_files(REPO_ID, repo_type=REPO_TYPE)
+    files = _get_repo_files()
     seen = set()
     pat = re.escape(prefix.rstrip("/")) + r"/([^/]+)/"
     for f in files:
@@ -62,13 +73,14 @@ def list_samples() -> list[str]:
 
 
 def list_samples_dasheng() -> list[str]:
-    """Return only sample IDs that exist in batch_outputs_dasheng (no fold* from UrbanSound8k)."""
-    return _get_sample_ids(DASHENG_PREFIX)
+    """Return only sample IDs for Dasheng view: from batch_outputs_dasheng, excluding fold* (UrbanSound8k)."""
+    ids = _get_sample_ids(DASHENG_PREFIX)
+    return sorted([x for x in ids if not x.startswith("fold")])
 
 
 def _find_files(inner: str) -> list[str]:
-    """List all repo files under inner path."""
-    files = list_repo_files(REPO_ID, repo_type=REPO_TYPE)
+    """List all repo files under inner path (uses cached repo file list)."""
+    files = _get_repo_files()
     return [f for f in files if f.startswith(inner + "/")]