Reuse enonce groups if availabel in annotating_by_label

annotating_by_label.py

@@ -125,11 +125,80 @@ def _initial_label_groups(labels: list[str]) -> str:
     return "".join(",".join(items) + "\n" for items in groups.values())
 
 
+def _gemini_label_groups(
+    workspace: EvaluationWorkspace, labels: list[str]
+) -> list[list[str]] | None:
+    source = workspace.gemini_exam_items_file
+    if not source.is_file():
+        return None
+
+    groups: list[list[str]] = []
+    current: list[str] = []
+    try:
+        source_lines = source.read_text(encoding="utf-8").splitlines()
+    except (OSError, UnicodeError) as exc:
+        print(f"Warning: could not read Gemini question groups from {source}: {exc}")
+        return None
+    for raw_line in source_lines:
+        line = raw_line.strip()
+        if not line or line.startswith("#"):
+            continue
+        if line == "---":
+            if current:
+                groups.append(current)
+                current = []
+            continue
+        if " ### " not in line:
+            continue
+        label = line.split(" ### ", 1)[0].strip()
+        if label and label != "CONTEXT":
+            current.append(label)
+    if current:
+        groups.append(current)
+
+    flattened = [label for group in groups for label in group]
+    known = set(labels)
+    if (
+        not flattened
+        or len(flattened) != len(set(flattened))
+        or set(flattened) != known
+    ):
+        missing = sorted(known.difference(flattened), key=natural_key)
+        unknown = sorted(set(flattened).difference(known), key=natural_key)
+        details = []
+        if missing:
+            details.append("missing: " + ", ".join(missing))
+        if unknown:
+            details.append("unknown: " + ", ".join(unknown))
+        if len(flattened) != len(set(flattened)):
+            details.append("duplicate labels")
+        print(
+            f"Warning: ignoring incompatible Gemini question groups in {source}"
+            + (f" ({'; '.join(details)})" if details else "")
+        )
+        return None
+    return groups
+
+
+def _serialize_label_groups(groups: list[list[str]]) -> str:
+    return "".join(",".join(group) + "\n" for group in groups)
+
+
 def _load_label_groups(workspace: EvaluationWorkspace, labels: list[str]) -> list[list[str]]:
-    label_groups = workspace.root / "label_groups"
+    label_groups = workspace.label_groups_file
     if not label_groups.exists():
-        atomic_write_text(label_groups, _initial_label_groups(labels))
-        print(f"Created {label_groups}; review the groups before continuing.")
+        gemini_groups = _gemini_label_groups(workspace, labels)
+        if gemini_groups is not None:
+            initial_content = _serialize_label_groups(gemini_groups)
+            source_description = "the groups selected in gemini_for_enonce.py"
+        else:
+            initial_content = _initial_label_groups(labels)
+            source_description = "the label-prefix fallback"
+        atomic_write_text(label_groups, initial_content)
+        print(
+            f"Created {label_groups} from {source_description}; "
+            "review the groups before continuing."
+        )
         utils.edit_file_and_enter(label_groups)
     known_labels = set(labels)
     groups: list[list[str]] = []