fix manual parser refresh reliability

tests/apps/parsers/test_gisp_client.py

@@ -49,6 +49,35 @@ class GispProductsClientTest(SimpleTestCase):
         ):
             GispProductsClient(max_records=10_001)
 
+    def test_retries_transient_server_error(self):
+        calls = 0
+
+        def handle_page(_request, _body: bytes) -> Response:
+            nonlocal calls
+            calls += 1
+            if calls == 1:
+                return Response(status=504, body=b"gateway timeout")
+            return Response(
+                body=b'{"ok":true,"total_count":1,"items":[]}',
+                headers={"Content-Type": "application/json"},
+            )
+
+        with HTTPTestServer() as server:
+            server.add_route("POST", "/pp719v2/pub/prod/b/", handle_page)
+            client = GispProductsClient(
+                base_url=server.base_url,
+                max_pages=1,
+                retry_backoff_seconds=0,
+                http_adapter=server.adapter,
+            )
+            with self.assertRaisesMessage(
+                GispProductsClientError,
+                "empty page before total_count",
+            ):
+                client.fetch_products()
+
+        self.assertEqual(calls, 2)
+
     def test_fetch_products_follows_skip_pagination_and_maps_items(self):
         requests: list[dict] = []
 

tests/apps/parsers/test_proverki_client.py

@@ -9,6 +9,7 @@ import tempfile
 import types
 from asyncio import events as asyncio_events
 from pathlib import Path
+from unittest.mock import patch
 from xml.etree import ElementPath as element_path
 from xml.etree import ElementTree as ET
 
@@ -563,6 +564,18 @@ class ProverkiParseXMLTest(SimpleTestCase):
         inspections = client._parse_xml_content(xml)
         self.assertEqual(len(inspections), 1)
 
+    def test_large_xml_streaming_does_not_build_sanitized_string(self):
+        xml = _xml_with_tag("INSPECTION", _inspection_attrs())
+        client = ProverkiClient()
+        client.STREAMING_THRESHOLD_BYTES = 1
+        with patch.object(
+            client,
+            "_sanitize_xml",
+            side_effect=AssertionError("large XML must stay as bytes"),
+        ):
+            inspections = client._parse_xml_content(xml)
+        self.assertEqual(len(inspections), 1)
+
     def test_parse_xml_record_missing_fields_returns_none(self):
         element = ET.fromstring("<INSPECTION />")  # noqa: S314
         client = ProverkiClient()

tests/apps/parsers/test_views.py

@@ -11,7 +11,7 @@ import zipfile
 from datetime import date
 from unittest.mock import Mock, patch
 
-from apps.core.models import BackgroundJob
+from apps.core.models import BackgroundJob, JobStatus
 from apps.parsers.models import (
     FinancialReport,
     FinancialReportLine,
@@ -1428,3 +1428,22 @@ class ParsersViewSetTest(APITestCase):
         queued_task_id = apply_async_mock.call_args.kwargs["task_id"]
         job = BackgroundJob.objects.get(task_id=queued_task_id)
         self.assertEqual(job.meta["source_key"], "mpt_products")
+
+    def test_run_parser_reuses_fresh_active_job(self):
+        self.client.force_authenticate(self.user)
+        existing = BackgroundJob.objects.create(
+            task_id="active-products",
+            task_name="apps.parsers.tasks.parse_industrial_products",
+            status=JobStatus.STARTED,
+        )
+        url = reverse("api_v1:parsers:run-parser", args=["industrial_products"])
+
+        with patch(
+            "apps.parsers.views.tasks.parse_industrial_products.apply_async"
+        ) as apply_async_mock:
+            response = self.client.post(url, {}, format="json")
+
+        self.assertEqual(response.status_code, status.HTTP_202_ACCEPTED)
+        self.assertEqual(response.data["data"]["task_id"], existing.task_id)
+        self.assertTrue(response.data["data"]["already_running"])
+        apply_async_mock.assert_not_called()