fix: enhance email threading and extraction logic in sag email handling

app/modules/sag/backend/router.py

@@ -1893,11 +1893,25 @@ async def add_sag_email_link(sag_id: int, payload: dict):
 async def get_sag_emails(sag_id: int):
     """Get emails linked to a case."""
     query = """
-        SELECT e.* 
+        WITH linked_emails AS (
-        FROM email_messages e
+            SELECT
-        JOIN sag_emails se ON e.id = se.email_id
+                e.*,
-        WHERE se.sag_id = %s
+                COALESCE(
-        ORDER BY e.received_date DESC
+                    NULLIF(REGEXP_REPLACE(TRIM(COALESCE(e.in_reply_to, '')), '[<>\\s]', '', 'g'), ''),
+                    NULLIF(REGEXP_REPLACE((REGEXP_SPLIT_TO_ARRAY(COALESCE(e.email_references, ''), E'[\\s,]+'))[1], '[<>\\s]', '', 'g'), ''),
+                    NULLIF(REGEXP_REPLACE(TRIM(COALESCE(e.message_id, '')), '[<>\\s]', '', 'g'), ''),
+                    CONCAT('email-', e.id::text)
+                ) AS thread_key
+            FROM email_messages e
+            JOIN sag_emails se ON e.id = se.email_id
+            WHERE se.sag_id = %s
+        )
+        SELECT
+            linked_emails.*,
+            COUNT(*) OVER (PARTITION BY linked_emails.thread_key) AS thread_message_count,
+            MAX(linked_emails.received_date) OVER (PARTITION BY linked_emails.thread_key) AS thread_last_received_date
+        FROM linked_emails
+        ORDER BY thread_last_received_date DESC NULLS LAST, received_date DESC
     """
     return execute_query(query, (sag_id,)) or []
 
@@ -1969,6 +1983,8 @@ async def upload_sag_email(sag_id: int, file: UploadFile = File(...)):
             
         email_data = {
             'message_id': msg.get('Message-ID', f"eml-{temp_id}"),
+            'in_reply_to': _decode_header_str(msg.get('In-Reply-To', '')),
+            'email_references': _decode_header_str(msg.get('References', '')),
             'subject': _decode_header_str(msg.get('Subject', 'No Subject')),
             'sender_email': _decode_header_str(msg.get('From', '')),
             'sender_name': _decode_header_str(msg.get('From', '')),

app/modules/sag/templates/detail.html

@@ -5837,20 +5837,40 @@
                 return;
             }
             setModuleContentState('emails', true);
-            container.innerHTML = emails.map(e => `
+            const threadMap = new Map();
-                <div class="list-group-item">
+            emails.forEach(e => {
-                    <div class="d-flex justify-content-between align-items-start">
+                const key = e.thread_key || `email-${e.id}`;
-                        <div class="text-truncate">
+                if(!threadMap.has(key)) threadMap.set(key, []);
-                            <i class="bi bi-envelope text-primary me-1"></i>
+                threadMap.get(key).push(e);
-                            <strong>${e.subject || '(Ingen emne)'}</strong>
+            });
-                            <div class="small text-muted text-truncate">${e.sender_email}</div>
+
-                        </div>
+            const threads = Array.from(threadMap.values());
-                        <button class="btn btn-sm btn-link text-danger p-0 ms-2" onclick="unlinkEmail(${e.id})">
+            container.innerHTML = threads.map((threadEmails, threadIndex) => {
-                            <i class="bi bi-link-45deg" style="text-decoration: line-through;"></i>
+                const labelEmail = threadEmails[0];
-                        </button>
+                const messageCount = labelEmail.thread_message_count || threadEmails.length;
+                return `
+                <div class="list-group-item p-0">
+                    <div class="px-3 py-2 border-bottom bg-light d-flex justify-content-between align-items-center">
+                        <span class="small fw-semibold text-secondary">Tråd ${threadIndex + 1}</span>
+                        <span class="badge bg-primary-subtle text-primary-emphasis">${messageCount} beskeder</span>
                     </div>
+                    ${threadEmails.map(e => `
+                    <div class="px-3 py-2 border-bottom">
+                        <div class="d-flex justify-content-between align-items-start">
+                            <div class="text-truncate">
+                                <i class="bi bi-envelope text-primary me-1"></i>
+                                <strong>${e.subject || '(Ingen emne)'}</strong>
+                                <div class="small text-muted text-truncate">${e.sender_email || '-'}</div>
+                            </div>
+                            <button class="btn btn-sm btn-link text-danger p-0 ms-2" onclick="unlinkEmail(${e.id})">
+                                <i class="bi bi-link-45deg" style="text-decoration: line-through;"></i>
+                            </button>
+                        </div>
+                    </div>
+                    `).join('')}
                 </div>
-            `).join('');
+                `;
+            }).join('');
         }
         
         async function unlinkEmail(emailId) {

app/services/email_workflow_service.py

@@ -221,6 +221,49 @@ class EmailWorkflowService:
                 return int(match.group(1))
         return None
 
+    def _normalize_message_id(self, value: Optional[str]) -> Optional[str]:
+        if not value:
+            return None
+        normalized = re.sub(r'[<>\s]', '', str(value)).lower().strip()
+        return normalized or None
+
+    def _extract_thread_message_ids(self, email_data: Dict) -> List[str]:
+        tokens: List[str] = []
+
+        in_reply_to = self._normalize_message_id(email_data.get('in_reply_to'))
+        if in_reply_to:
+            tokens.append(in_reply_to)
+
+        raw_references = (email_data.get('email_references') or '').strip()
+        if raw_references:
+            for ref in re.split(r'[\s,]+', raw_references):
+                normalized_ref = self._normalize_message_id(ref)
+                if normalized_ref:
+                    tokens.append(normalized_ref)
+
+        # De-duplicate while preserving order
+        return list(dict.fromkeys(tokens))
+
+    def _find_sag_id_from_thread_headers(self, email_data: Dict) -> Optional[int]:
+        thread_message_ids = self._extract_thread_message_ids(email_data)
+        if not thread_message_ids:
+            return None
+
+        placeholders = ','.join(['%s'] * len(thread_message_ids))
+        rows = execute_query(
+            f"""
+            SELECT se.sag_id
+            FROM sag_emails se
+            JOIN email_messages em ON em.id = se.email_id
+            WHERE em.deleted_at IS NULL
+              AND LOWER(REGEXP_REPLACE(COALESCE(em.message_id, ''), '[<>\\s]', '', 'g')) IN ({placeholders})
+            ORDER BY se.created_at DESC
+            LIMIT 1
+            """,
+            tuple(thread_message_ids)
+        )
+        return rows[0]['sag_id'] if rows else None
+
     def _find_customer_by_domain(self, domain: str) -> Optional[Dict[str, Any]]:
         if not domain:
             return None
@@ -309,6 +352,10 @@ class EmailWorkflowService:
             return None
 
         sag_id = self._extract_sag_id(email_data)
+        if not sag_id:
+            sag_id = self._find_sag_id_from_thread_headers(email_data)
+            if sag_id:
+                logger.info("🔗 Matched email %s to SAG-%s via thread headers", email_id, sag_id)
 
         # 1) Existing SAG via subject/headers
         if sag_id: