From 46c0e4cc18daa991821f0133e44561bcc8f360a6 Mon Sep 17 00:00:00 2001
From: Joshua Laymon <joshlaymon@icloud.com>
Date: Sat, 16 Aug 2025 15:54:10 +0000
Subject: [PATCH] Add web/core/normalize_subjects.py

---
 web/core/normalize_subjects.py | 47 ++++++++++++++++++++++++++++++++++
 1 file changed, 47 insertions(+)
 create mode 100644 web/core/normalize_subjects.py

diff --git a/web/core/normalize_subjects.py b/web/core/normalize_subjects.py
new file mode 100644
index 0000000..3fd2253
--- /dev/null
+++ b/web/core/normalize_subjects.py
@@ -0,0 +1,47 @@
+from django.core.management.base import BaseCommand
+from core.models import Entry
+
+class Command(BaseCommand):
+    help = "Normalize subjects (split by commas, unify delimiters like ; and -)"
+
+    def add_arguments(self, parser):
+        parser.add_argument(
+            "--dry-run",
+            action="store_true",
+            help="Preview changes without saving"
+        )
+
+    def handle(self, *args, **options):
+        dry_run = options["dry_run"]
+        changes = 0
+
+        for entry in Entry.objects.all():
+            subj = entry.subject or ""
+            normalized = self.normalize_subjects(subj)
+
+            if subj != normalized:
+                changes += 1
+                self.stdout.write(
+                    f"Would change: {subj} → {normalized}"
+                )
+                if not dry_run:
+                    entry.subject = normalized
+                    entry.save(update_fields=["subject"])
+
+        if dry_run:
+            self.stdout.write(self.style.WARNING(f"[DRY RUN] {changes} subjects would be changed."))
+        else:
+            self.stdout.write(self.style.SUCCESS(f"Updated {changes} subjects."))
+
+    def normalize_subjects(self, text: str) -> str:
+        if not text:
+            return ""
+
+        # Unify delimiters into commas
+        cleaned = text.replace(";", ",").replace("|", ",").replace("/", ",").replace(" - ", ",").replace("-", ",")
+
+        # Split, strip spaces, remove empties
+        parts = [p.strip() for p in cleaned.split(",") if p.strip()]
+
+        # Rejoin with a standard ", " separator
+        return ", ".join(parts)
\ No newline at end of file