find latest wf directory

Author: anwurl

examples/analyse_study_classification.ipynb

@@ -62,6 +62,7 @@
     "from octopus.predict import TaskPredictorTest\n",
     "from octopus.predict.notebook_utils import (\n",
     "    display_table,\n",
+    "    find_latest_study,\n",
     "    show_aucroc_plots,\n",
     "    show_confusionmatrix,\n",
     "    show_overall_fi_plot,\n",
@@ -92,8 +93,10 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "# INPUT: Select study\n",
-    "study_directory = \"../studies/wf_octo_mrmr_octo/\""
+    "# INPUT: Select study by name prefix (automatically picks the latest timestamped run)\n",
+    "study_name_prefix = \"wf_octo_mrmr_octo\"\n",
+    "study_directory = find_latest_study(\"../studies\", study_name_prefix)\n",
+    "print(f\"Using study: {study_directory}\")"
    ]
   },
   {
@@ -374,6 +377,9 @@
    "cell_metadata_filter": "vscode,-all",
    "main_language": "python",
    "notebook_metadata_filter": "-all"
+  },
+  "language_info": {
+   "name": "python"
   }
  },
  "nbformat": 4,

octopus/predict/notebook_utils.py

@@ -21,6 +21,7 @@
 
 __all__ = [
     "display_table",
+    "find_latest_study",
     "show_aucroc_plots",
     "show_confusionmatrix",
     "show_overall_fi_plot",
@@ -31,6 +32,44 @@
     "show_testset_performance",
 ]
 
+def find_latest_study(studies_root: str | UPath, prefix: str) -> str:
+    """Find the latest study directory matching a name prefix.
+
+    Study directories are named ``<prefix>-YYYYMMDD_HHMMSS``.  This function
+    finds all directories matching the given *prefix* and returns the one with
+    the most recent timestamp (lexicographic sort).  Falls back to an exact
+    match (no timestamp suffix) when no timestamped directories are found.
+
+    Args:
+        studies_root: Path to the parent directory containing study directories.
+        prefix: The study name prefix, e.g. ``"wf_octo_mrmr_octo"``.
+
+    Returns:
+        Path string to the latest matching study directory.
+
+    Raises:
+        FileNotFoundError: If no matching study directory is found.
+
+    Example:
+        >>> from octopus.predict.notebook_utils import find_latest_study
+        >>> study_dir = find_latest_study("./studies", "wf_octo_mrmr_octo")
+    """
+    root = UPath(studies_root)
+    # Match timestamped directories: prefix-YYYYMMDD_HHMMSS
+    candidates = sorted(
+        [d for d in root.glob(f"{prefix}-*") if d.is_dir()],
+        key=lambda p: p.name,
+        reverse=True,
+    )
+    if candidates:
+        return str(candidates[0])
+    # Fallback: exact match without timestamp
+    exact = root / prefix
+    if exact.is_dir():
+        return str(exact)
+    raise FileNotFoundError(f"No study directory found for prefix '{prefix}' in {root}")
+
+
 try:
     from IPython.display import display as ipython_display
 except ImportError: