From fb2243f0d48c705e0255be64f96cc857e07f6e56 Mon Sep 17 00:00:00 2001
From: Christian <christian@blaahund.dk>
Date: Fri, 3 Apr 2026 00:50:34 +0200
Subject: [PATCH] Preserve email body in auto-created cases and backfill
 missing content

---
 app/services/email_workflow_service.py        | 38 +++++++++++++-
 .../160_backfill_autocreated_case_body.sql    | 51 +++++++++++++++++++
 2 files changed, 87 insertions(+), 2 deletions(-)
 create mode 100644 migrations/160_backfill_autocreated_case_body.sql

diff --git a/app/services/email_workflow_service.py b/app/services/email_workflow_service.py
index 8e419c6..b583da4 100644
--- a/app/services/email_workflow_service.py
+++ b/app/services/email_workflow_service.py
@@ -12,6 +12,7 @@ import json
 import hashlib
 import shutil
 import io
+import html
 from pathlib import Path
 from decimal import Decimal
 from uuid import uuid4
@@ -965,13 +966,45 @@ class EmailWorkflowService:
 
         return "\n".join(cleaned_lines).strip()
 
+    def _html_to_text(self, body_html: str) -> str:
+        """Convert HTML email bodies to readable plain text fallback."""
+        if not body_html:
+            return ""
+
+        text = str(body_html)
+        text = re.sub(r'<(style|script)[^>]*>.*?</\1>', '', text, flags=re.IGNORECASE | re.DOTALL)
+        text = re.sub(r'<\s*br\s*/?>', '\n', text, flags=re.IGNORECASE)
+        text = re.sub(r'</\s*(p|div|li|tr|h[1-6])\s*>', '\n', text, flags=re.IGNORECASE)
+        text = re.sub(r'<[^>]+>', ' ', text)
+        text = html.unescape(text)
+        text = text.replace('\r\n', '\n').replace('\r', '\n')
+        text = re.sub(r'\n{3,}', '\n\n', text)
+        text = re.sub(r'[ \t]{2,}', ' ', text)
+        return text.strip()
+
+    def _extract_primary_email_body(self, email_data: Dict) -> str:
+        """Return best-effort email body text for case/comment persistence."""
+        raw_text = (email_data.get('body_text') or '').strip()
+        if raw_text:
+            cleaned = self._strip_quoted_email_text(raw_text)
+            if cleaned:
+                return cleaned
+
+        html_fallback = self._html_to_text(email_data.get('body_html') or '')
+        if html_fallback:
+            cleaned = self._strip_quoted_email_text(html_fallback)
+            if cleaned:
+                return cleaned
+
+        return ""
+
     def _add_helpdesk_comment(self, sag_id: int, email_data: Dict) -> None:
         email_id = email_data.get('id')
         sender = email_data.get('sender_email') or 'ukendt'
         subject = email_data.get('subject') or '(ingen emne)'
         received = email_data.get('received_date')
         received_str = received.isoformat() if hasattr(received, 'isoformat') else str(received or '')
-        body_text = self._strip_quoted_email_text((email_data.get('body_text') or '').strip())
+        body_text = self._extract_primary_email_body(email_data)
 
         email_meta_line = f"Email-ID: {email_id}\n" if email_id else ""
 
@@ -995,12 +1028,13 @@ class EmailWorkflowService:
     def _create_sag_from_email(self, email_data: Dict, customer_id: int) -> Dict[str, Any]:
         sender = email_data.get('sender_email') or 'ukendt'
         subject = (email_data.get('subject') or '').strip() or f"Email fra {sender}"
+        body_text = self._extract_primary_email_body(email_data)
 
         description = (
             f"Auto-oprettet fra email\n"
             f"Fra: {sender}\n"
             f"Message-ID: {email_data.get('message_id') or ''}\n\n"
-            f"{(email_data.get('body_text') or '').strip()}"
+            f"{body_text}"
         )
 
         rows = execute_query(
diff --git a/migrations/160_backfill_autocreated_case_body.sql b/migrations/160_backfill_autocreated_case_body.sql
new file mode 100644
index 0000000..59d5bb9
--- /dev/null
+++ b/migrations/160_backfill_autocreated_case_body.sql
@@ -0,0 +1,51 @@
+-- Migration 160: Backfill missing body text in auto-created SAG descriptions
+-- Fills existing cases where description only contains metadata and no actual email content.
+
+WITH candidates AS (
+    SELECT
+        s.id AS sag_id,
+        s.beskrivelse,
+        COALESCE(
+            NULLIF(BTRIM(em.body_text), ''),
+            NULLIF(
+                BTRIM(
+                    REGEXP_REPLACE(
+                        REGEXP_REPLACE(COALESCE(em.body_html, ''), '<(style|script)[^>]*>.*?</\1>', ' ', 'gis'),
+                        '<[^>]+>',
+                        ' ',
+                        'g'
+                    )
+                ),
+                ''
+            )
+        ) AS recovered_body
+    FROM sag_sager s
+    JOIN sag_emails se ON se.sag_id = s.id
+    JOIN email_messages em ON em.id = se.email_id
+    WHERE s.beskrivelse LIKE 'Auto-oprettet fra email%'
+      AND COALESCE(
+            NULLIF(
+                BTRIM(
+                    REGEXP_REPLACE(
+                        COALESCE(s.beskrivelse, ''),
+                        '(?s)^Auto-oprettet fra email\s*\nFra:[^\n]*\nMessage-ID:[^\n]*\n*',
+                        ''
+                    )
+                ),
+                ''
+            ),
+            ''
+          ) = ''
+), dedup AS (
+    SELECT DISTINCT ON (sag_id)
+        sag_id,
+        recovered_body
+    FROM candidates
+    WHERE recovered_body IS NOT NULL
+      AND recovered_body <> ''
+    ORDER BY sag_id
+)
+UPDATE sag_sager s
+SET beskrivelse = CONCAT_WS(E'\n\n', BTRIM(COALESCE(s.beskrivelse, '')), dedup.recovered_body)
+FROM dedup
+WHERE s.id = dedup.sag_id;