[IMP] datev_export_dtvf_nitrokey: Try to get the best references for …

…the financial advisor

datev_export_dtvf_nitrokey/models/datev_export_dtvf.py

@@ -6,16 +6,47 @@
 class DatevExportDtvfExport(models.Model):
     _inherit = "datev_export_dtvf.export"
 
+    def _get_matching_references(self, move, ref1, ref2):
+        # Incoming invoices will use the vendor reference
+        if move.move_type.startswith("in_"):
+            return move.ref, ref1
+
+        # Outgoing invoices will use the sale.order reference
+        if move.move_type == "out_invoice" and move.invoice_origin:
+            return move.invoice_origin, ref1
+
+        # Outgoing refunds will use the sale.order reference from the reversed invoice
+        if move.move_type == "out_refund" and move.reversed_entry_id.invoice_origin:
+            return move.reversed_entry_id.invoice_origin, ref1
+
+        # All non-entry will use the default
+        if move.move_type != "entry":
+            return ref1, ref2
+
+        # Try to find the reconciled invoices
+        matching = move.mapped("line_ids.full_reconcile_id.reconciled_line_ids")
+        if not matching:
+            return ref1, ref2
+
+        # First invoices if possible
+        domain = [("move_type", "=like", "%_invoice")]
+        invs = matching.filtered_domain(domain)
+        if len(invs) == 1:
+            return self._get_matching_references(invs, ref1, ref2)
+
+        # Fall back to refunds if possible
+        domain = [("move_type", "=like", "%_refund")]
+        invs = matching.filtered_domain(domain)
+        if len(invs) == 1:
+            return self._get_matching_references(
+                invs.reversed_entry_id or invs, ref1, ref2
+            )
+
+        # Fall back to the default
+        return ref1, ref2
+
     def _get_data_transaction(self, move):
-        """Try to export whatever looks like an SO number as "Belegfeld 1" """
         for data in super()._get_data_transaction(move):
-            for line in move.line_ids:
-                for field_name in ("ref", "name", "move_name"):
-                    for token in (line[field_name] or "").split():
-                        if (
-                            token.startswith("SO")
-                            or token.startswith("EK")
-                            or token.startswith("GSV")
-                        ):
-                            data["Belegfeld 1"] = token
+            ref1, ref2 = self._get_matching_references(move, move.name, "")
+            data.update({"Belegfeld 1": ref1, "Belegfeld 2": ref2})
             yield data