From 5077ecffca8d60bde73257680c4a8556d805a3fb Mon Sep 17 00:00:00 2001
From: Joshua Laymon <joshlaymon@icloud.com>
Date: Sat, 16 Aug 2025 16:46:16 +0000
Subject: [PATCH] Update web/core/views.py

---
 web/core/views.py | 72 +++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 72 insertions(+)

diff --git a/web/core/views.py b/web/core/views.py
index 6c97e34..6e36e59 100644
--- a/web/core/views.py
+++ b/web/core/views.py
@@ -14,6 +14,7 @@ from .forms import ImportForm, EntryForm
 from .utils import terms, has_wildcards, wildcard_to_regex, import_csv_bytes
 from .scripture_normalizer import normalize_scripture_field  # <-- NEW
 from .source_normalizer import normalize_source_field  # NEW
+from .subject_normalizer import normalize_subject_field  # NEW
 
 # Order + labels used in the Search UI
 FIELD_ORDER = [
@@ -638,4 +639,75 @@ def stats_page(request):
             "top_refs": top_refs,
             "book_distribution": book_distribution,
         },
+    )
+    @login_required
+@user_passes_test(is_admin)
+@require_http_methods(["GET", "POST"])
+def normalize_subjects(request):
+    """
+    GET  -> dry-run preview (summary + first 100 examples)
+    POST -> apply changes to all entries' subject (batched)
+    Optional ?limit= for preview subset.
+    """
+    apply = request.method == "POST"
+    limit = int(request.GET.get("limit", "0") or "0")
+
+    qs = Entry.objects.all().order_by("id")
+    if limit:
+        qs = qs[:limit]
+
+    changed = 0
+    warnings_total = 0
+    preview = []
+
+    if apply:
+        # write in batches to keep transactions short
+        from django.db import transaction
+        batch, pending = 500, []
+        for e in qs.iterator():
+            original = (e.subject or "").strip()
+            normalized, warns = normalize_subject_field(original)
+            warnings_total += len(warns)
+            if normalized != original:
+                changed += 1
+                preview.append((e.id, original, normalized))
+                e.subject = normalized
+                pending.append(e)
+                if len(pending) >= batch:
+                    with transaction.atomic():
+                        for obj in pending:
+                            obj.save(update_fields=["subject"])
+                    pending.clear()
+        if pending:
+            from django.db import transaction
+            with transaction.atomic():
+                for obj in pending:
+                    obj.save(update_fields=["subject"])
+    else:
+        # dry-run only
+        for e in qs.iterator():
+            original = (e.subject or "").strip()
+            normalized, warns = normalize_subject_field(original)
+            warnings_total += len(warns)
+            if normalized != original:
+                changed += 1
+                preview.append((e.id, original, normalized))
+
+    preview = preview[:100]  # keep table reasonable
+
+    messages.info(
+        request,
+        f"{'Applied' if apply else 'Dry-run'}: {changed} entries "
+        f"{'changed' if apply else 'would change'}; {warnings_total} warnings."
+    )
+    return render(
+        request,
+        "normalize_subjects_result.html",
+        {
+            "applied": apply,
+            "changed": changed,
+            "warnings_total": warnings_total,
+            "preview": preview,
+            "limit": limit,
+        },
     )
\ No newline at end of file