fix(parsers): make source jobs resumable

tests/apps/organizations/test_api_v2_source_extensions.py

@@ -6,6 +6,7 @@ from django.core.cache import cache
 from django.urls import reverse
 from organizations.models import (
     ArbitrationExtension,
+    ElectronicDocumentExchangeExtension,
     Organization,
     OrganizationSourceRecord,
     PlannedInspectionExtension,
@@ -335,6 +336,58 @@ class OrganizationSourceExtensionsApiV2Test(APITestCase):
         self.assertIsNone(ascending.data["data"][-1]["record_date"])
         self.assertIsNone(descending.data["data"][-1]["record_date"])
 
+    def test_flat_gosedo_records_support_frontend_sorting_fields(self):
+        records = []
+        for external_id, name, full_name, attestation_status in (
+            ("GOSEDO-BETA", "Организация Бета", "Полное имя Бета", "not_attested"),
+            ("GOSEDO-ALPHA", "Организация Альфа", "Полное имя Альфа", "attested"),
+        ):
+            organization = create_frontend_organization(
+                name=name,
+                full_name=full_name,
+                inn="7707083815" if external_id == "GOSEDO-BETA" else "7707083816",
+                ogrn=(
+                    "1027700132015" if external_id == "GOSEDO-BETA" else "1027700132016"
+                ),
+            )
+            extension = ElectronicDocumentExchangeExtension.objects.create(
+                organization=organization,
+                title="Электронный документооборот",
+            )
+            records.append(
+                OrganizationSourceRecord.objects.create(
+                    extension=extension,
+                    record_type="participant",
+                    source="gosedo_address_directory",
+                    external_id=external_id,
+                    payload={"attestation_status": attestation_status},
+                )
+            )
+
+        endpoint = reverse("api_v2:organizations:organization-source-records-list")
+        params = {
+            "source_group": "electronic_document_exchange",
+            "source": "gosedo_address_directory",
+            "record_type": "participant",
+        }
+        expected_ascending = [records[1].external_id, records[0].external_id]
+        expected_descending = list(reversed(expected_ascending))
+
+        for ordering, expected in (
+            ("extension__organization__full_name", expected_ascending),
+            ("-extension__organization__full_name", expected_descending),
+            ("payload__attestation_status", expected_ascending),
+            ("-payload__attestation_status", expected_descending),
+        ):
+            with self.subTest(ordering=ordering):
+                response = self.client.get(endpoint, {**params, "ordering": ordering})
+
+                self.assertEqual(response.status_code, status.HTTP_200_OK)
+                self.assertEqual(
+                    [item["external_id"] for item in response.data["data"]],
+                    expected,
+                )
+
     def test_flat_source_records_rejects_invalid_date_range_and_ordering(self):
         response = self.client.get(
             reverse("api_v2:organizations:organization-source-records-list"),

tests/apps/parsers/test_gosedo_media_news.py

@@ -2,9 +2,11 @@ from __future__ import annotations
 
 from datetime import datetime, timedelta
 from io import BytesIO
+from tempfile import TemporaryDirectory
 from types import SimpleNamespace
 from unittest.mock import patch
 
+from apps.core.models import BackgroundJob, JobStatus
 from apps.parsers.gosedo import (
     GosedoNotModified,
     GosedoParseResult,
@@ -22,6 +24,9 @@ from apps.parsers.media_news import (
 )
 from apps.parsers.models import ParserLoadLog, ParserSourceArtifact, ParserStagedRecord
 from apps.parsers.source_artifacts import cleanup_parser_source_artifacts
+from apps.parsers.tasks import parse_media_news
+from django.core.files.base import ContentFile
+from django.core.files.storage import default_storage
 from django.core.files.uploadedfile import SimpleUploadedFile
 from django.test import TestCase
 from django.urls import reverse
@@ -519,6 +524,55 @@ class MediaNewsImportTest(TestCase):
         )
 
 
+class MediaNewsTaskTest(TestCase):
+    def test_uploaded_workbook_task_completes_background_job(self):
+        Organization.objects.create(
+            name="АО Задача СМИ",
+            inn="0012345678",
+            ogrn="1027700132195",
+            okpo="00123456",
+            directory_imported_at=timezone.now(),
+        )
+        workbook = _media_workbook(
+            [
+                [
+                    "00123456",
+                    "0012345678",
+                    "2026-07-01",
+                    "СМИ",
+                    "https://example.test/news",
+                    "Источник\nЗаголовок\nЛид\nСтрока 4\nПолный текст",
+                    "Положительная",
+                ]
+            ]
+        )
+
+        with TemporaryDirectory() as media_root, self.settings(MEDIA_ROOT=media_root):
+            file_path = default_storage.save(
+                "parser_uploads/media.xlsx",
+                ContentFile(workbook.getvalue()),
+            )
+            task_result = parse_media_news.apply(
+                kwargs={
+                    "file_path": file_path,
+                    "original_name": "media.xlsx",
+                },
+                task_id="media-news-task-test",
+            )
+
+            self.assertTrue(task_result.successful())
+            self.assertFalse(default_storage.exists(file_path))
+
+        job = BackgroundJob.objects.get(task_id="media-news-task-test")
+        self.assertEqual(job.status, JobStatus.SUCCESS)
+        self.assertEqual(job.progress, 100)
+        self.assertEqual(task_result.result["published"], 1)
+        self.assertEqual(
+            ParserLoadLog.objects.get(source=ParserLoadLog.Source.MEDIA_NEWS).status,
+            ParserLoadLog.Status.SUCCESS,
+        )
+
+
 class MediaNewsPermissionsTest(APITestCase):
     def setUp(self):
         self.user = UserFactory.create_user()
@@ -536,8 +590,11 @@ class MediaNewsPermissionsTest(APITestCase):
 
         self.client.force_authenticate(self.admin)
         with patch(
+            "apps.parsers.views._save_uploaded_parser_file",
+            return_value="parser_uploads/media.xlsx",
+        ), patch(
             "apps.parsers.tasks.parse_media_news.apply_async",
-            return_value=SimpleNamespace(id="media-task-1"),
+            side_effect=lambda **kwargs: SimpleNamespace(id=kwargs["task_id"]),
         ):
             response = self.client.post(
                 self.url,
@@ -546,8 +603,21 @@ class MediaNewsPermissionsTest(APITestCase):
             )
 
         self.assertEqual(response.status_code, status.HTTP_202_ACCEPTED)
-        self.assertEqual(response.data["data"]["task_id"], "media-task-1")
-        self.assertEqual(response.data["data"]["task_ids"], ["media-task-1"])
+        task_id = response.data["data"]["task_id"]
+        self.assertEqual(response.data["data"]["task_ids"], [task_id])
+
+        card_response = self.client.get(
+            reverse(
+                "api_v1:sources:source-cards-detail",
+                kwargs={"slug": "media-mentions"},
+            )
+        )
+        self.assertEqual(card_response.status_code, status.HTTP_200_OK)
+        self.assertEqual(card_response.data["data"]["status"], "in_progress")
+        self.assertEqual(
+            card_response.data["data"]["active_tasks"][0]["task_id"],
+            task_id,
+        )
 
     def test_gosedo_manual_run_is_admin_only_and_returns_task_ids(self):
         url = reverse(

tests/apps/parsers/test_source_cards_service.py

@@ -441,6 +441,31 @@ class SourceCardServiceDatabaseTest(TestCase):
     def setUp(self):
         SourceCardService.clear_cache()
 
+    def test_enqueue_vacancy_refresh_reuses_fresh_active_job(self):
+        existing = BackgroundJob.objects.create(
+            task_id="active-vacancies",
+            task_name="apps.parsers.tasks.parse_trudvsem_vacancies",
+            status=JobStatus.STARTED,
+        )
+        task = MagicMock()
+
+        result = SourceCardService._enqueue_task(
+            task=task,
+            task_name="apps.parsers.tasks.parse_trudvsem_vacancies",
+            requested_by_id=5,
+            meta={"source_card": "labor-vacancies"},
+            kwargs={"requested_by_id": 5},
+        )
+
+        self.assertEqual(
+            result,
+            {
+                "task_id": existing.task_id,
+                "task_name": existing.task_name,
+            },
+        )
+        task.apply_async.assert_not_called()
+
     def test_defense_unreliable_suppliers_counts_unique_generic_organizations(self):
         _save_source_record(
             source=ParserLoadLog.Source.UNFAIR_SUPPLIERS,

tests/apps/parsers/test_source_registry.py

@@ -2,6 +2,7 @@ from dataclasses import asdict
 
 from apps.parsers import tasks
 from apps.parsers.clients.common.structured import MAX_FILE_SIZE_BYTES
+from apps.parsers.serializers import ParserRunRequestSerializer
 from apps.parsers.source_registry import PARSER_SOURCES
 from apps.parsers.views import TASKS_BY_NAME
 from django.conf import settings
@@ -53,3 +54,19 @@ class ParserSourceRegistryPresentationTest(SimpleTestCase):
         )
 
         self.assertNotIn("checko", public_metadata.lower())
+
+    def test_vacancy_source_metadata_names_only_trudvsem(self):
+        source = PARSER_SOURCES["trudvsem"]
+
+        self.assertEqual(source.agency, "Работа России")
+        self.assertEqual(source.parser_strategy, "incremental_trudvsem_api")
+        self.assertNotIn("HH", source.data_scope)
+        self.assertNotIn("SuperJob", source.data_scope)
+
+    def test_vacancy_run_request_rejects_disabled_sources(self):
+        serializer = ParserRunRequestSerializer(
+            data={"vacancy_sources": ["hh", "superjob"]}
+        )
+
+        self.assertFalse(serializer.is_valid())
+        self.assertIn("vacancy_sources", serializer.errors)

tests/apps/parsers/test_tasks.py

@@ -16,6 +16,7 @@ from types import SimpleNamespace
 from unittest.mock import patch
 from urllib.parse import urlparse
 
+from apps.core.models import BackgroundJob, JobStatus
 from apps.core.services import BackgroundJobService
 from apps.parsers import tasks as parser_tasks
 from apps.parsers.clients.base import HTTPError
@@ -37,7 +38,11 @@ from apps.parsers.models import (
     FinancialReport,
     ParserLoadLog,
 )
-from apps.parsers.services import FNSReportService, ParserLoadLogService
+from apps.parsers.services import (
+    FNSReportService,
+    GenericParserRecordService,
+    ParserLoadLogService,
+)
 from apps.parsers.tasks import (
     FNS_API_DEFAULT_REGISTRY_ORGANIZATION_LIMIT,
     FNS_API_MAX_REGISTRY_ORGANIZATION_LIMIT,
@@ -2869,6 +2874,19 @@ class TaskHelpersTestCase(TestCase):
 
 
 class ParseVacanciesTaskTestCase(TestCase):
+    def _make_registry_organization(self, *, inn: int, name: str) -> Organization:
+        organization = OrganizationFactory(
+            pn_name=name,
+            mn_inn=inn,
+            mn_ogrn=inn + 1000000000000,
+        )
+        RegistryMembershipPeriodFactory(organization=organization, ended_at=None)
+        return organization
+
+    def test_vacancy_task_requeues_after_worker_loss(self):
+        self.assertTrue(parse_trudvsem_vacancies.acks_late)
+        self.assertTrue(parse_trudvsem_vacancies.reject_on_worker_lost)
+
     def test_parse_trudvsem_vacancies_without_filters_fetches_active_registry_orgs(
         self,
     ):
@@ -3052,170 +3070,28 @@ class ParseVacanciesTaskTestCase(TestCase):
             {"trudvsem:7701000102"},
         )
 
-    @override_settings(SUPERJOB_APP_ID="test-superjob-app-id")
-    def test_parse_trudvsem_vacancies_matches_job_boards_by_employer_name(self):
-        organization = OrganizationFactory(
-            pn_name='Общество с ограниченной ответственностью "Ромашка"',
-            mn_inn=7701000301,
-            mn_ogrn=1027700000301,
-        )
-        RegistryMembershipPeriodFactory(organization=organization, ended_at=None)
-        captured_client_kwargs = {}
-        captured_text_queries = {}
+    def test_parse_trudvsem_vacancies_rejects_disabled_sources(self):
+        _ensure_directory_organization(inn="7701000401")
 
-        class _Provider:
-            def __init__(self, source_name, *, supports_company_inn):
-                self.source_name = source_name
-                self.supports_company_inn = supports_company_inn
+        with self.assertRaisesMessage(
+            ValueError,
+            "Поддерживается только источник trudvsem",
+        ):
+            parse_trudvsem_vacancies(
+                limit=25,
+                region_code="1",
+                text="инженер",
+                proxies=[],
+                vacancy_sources=["hh", "superjob"],
+            )
 
-            def fetch_vacancies(self, **kwargs):
-                if self.source_name == "trudvsem":
-                    return [
-                        GenericParserItem(
-                            source=ParserLoadLog.Source.TRUDVSEM,
-                            external_id="trudvsem:romashka",
-                            inn=kwargs["company_inn"],
-                            title="Работа России",
-                            payload={"vacancy_source": "trudvsem"},
-                        )
-                    ]
+        self.assertFalse(OrganizationSourceRecord.objects.exists())
 
-                captured_text_queries[self.source_name] = kwargs["text"]
-                return [
-                    GenericParserItem(
-                        source=ParserLoadLog.Source.TRUDVSEM,
-                        external_id=f"{self.source_name}:romashka",
-                        organisation_name='ООО "Ромашка"',
-                        title=f"{self.source_name} matching vacancy",
-                        payload={"vacancy_source": self.source_name},
-                    ),
-                    GenericParserItem(
-                        source=ParserLoadLog.Source.TRUDVSEM,
-                        external_id=f"{self.source_name}:other",
-                        organisation_name='ООО "Лютик"',
-                        title=f"{self.source_name} unrelated vacancy",
-                        payload={"vacancy_source": self.source_name},
-                    ),
-                ]
-
-        class _VacanciesClient:
-            def __init__(self, **kwargs):
-                captured_client_kwargs.update(kwargs)
-
-            def __enter__(self):
-                return self
-
-            def __exit__(self, exc_type, exc_val, exc_tb):
-                return None
-
-            def fetch_vacancies(self, **kwargs):
-                return [
-                    GenericParserItem(
-                        source=ParserLoadLog.Source.TRUDVSEM,
-                        external_id="trudvsem:romashka",
-                        inn=kwargs["company_inn"],
-                        title="Работа России",
-                        payload={"vacancy_source": "trudvsem"},
-                    )
-                ]
-
-            def iter_source_clients(self):
-                return [
-                    ("trudvsem", _Provider("trudvsem", supports_company_inn=True)),
-                    ("hh", _Provider("hh", supports_company_inn=False)),
-                    ("superjob", _Provider("superjob", supports_company_inn=False)),
-                ]
-
-        original_client = parser_tasks.VacanciesClient
-        parser_tasks.VacanciesClient = _VacanciesClient
-        try:
-            result = parse_trudvsem_vacancies(limit=50, proxies=[])
-        finally:
-            parser_tasks.VacanciesClient = original_client
-
-        self.assertEqual(result["status"], "success")
-        self.assertIsNone(captured_client_kwargs["sources"])
-        self.assertEqual(
-            captured_text_queries,
-            {
-                "hh": "ромашка",
-                "superjob": "ромашка",
-            },
-        )
-        self.assertEqual(result["saved"], 3)
-        self.assertEqual(
-            {
-                (
-                    record.external_id,
-                    record.source,
-                    record.registry_organization_id,
-                )
-                for record in OrganizationSourceRecord.objects.all()
-            },
-            {
-                ("trudvsem:romashka", "trudvsem", organization.id),
-                ("hh:romashka", "hh", organization.id),
-                ("superjob:romashka", "superjob", organization.id),
-            },
-        )
-
-    def test_registry_job_board_matching_fetches_only_first_text_search_page(self):
-        organization = OrganizationFactory(
-            pn_name='Общество с ограниченной ответственностью "Ромашка"',
-            mn_inn=7701000302,
-            mn_ogrn=1027700000302,
-        )
-        RegistryMembershipPeriodFactory(organization=organization, ended_at=None)
-        captured_offsets = []
-
-        class _Provider:
-            supports_company_inn = False
-
-            def fetch_vacancies(self, **kwargs):
-                captured_offsets.append(kwargs["offset"])
-                return [
-                    GenericParserItem(
-                        source=ParserLoadLog.Source.TRUDVSEM,
-                        external_id=f"hh:romashka:{kwargs['offset']}",
-                        organisation_name='ООО "Ромашка"',
-                        title="HeadHunter",
-                        payload={"vacancy_source": "hh"},
-                    )
-                ]
-
-        class _VacanciesClient:
-            def __init__(self, **kwargs):
-                pass
-
-            def __enter__(self):
-                return self
-
-            def __exit__(self, exc_type, exc_val, exc_tb):
-                return None
-
-            def iter_source_clients(self):
-                return [("hh", _Provider())]
-
-        original_client = parser_tasks.VacanciesClient
-        parser_tasks.VacanciesClient = _VacanciesClient
-        try:
-            result = parse_trudvsem_vacancies(limit=1, proxies=[])
-        finally:
-            parser_tasks.VacanciesClient = original_client
-
-        self.assertEqual(result["status"], "success")
-        self.assertEqual(result["saved"], 1)
-        self.assertEqual(captured_offsets, [0])
-
-    @override_settings(
-        SUPERJOB_APP_ID="test-superjob-app-id",
-        HH_USER_AGENT="Mostovik/1.0 (ops@example.test)",
-    )
-    def test_parse_trudvsem_vacancies_uses_combined_vacancies_client(self):
-        for inn in ("7701000401", "7701000402", "7701000403"):
-            _ensure_directory_organization(inn=inn)
+    def test_registry_run_persists_a_chunk_and_queues_continuation(self):
+        self._make_registry_organization(inn=7701000501, name="АО Первый пакет")
+        self._make_registry_organization(inn=7701000502, name="АО Второй пакет")
         captured_kwargs = {}
-        captured_fetch_kwargs = {}
+        captured_fetches = []
 
         class _VacanciesClient:
             def __init__(self, **kwargs):
@@ -3228,70 +3104,175 @@ class ParseVacanciesTaskTestCase(TestCase):
                 return None
 
             def fetch_vacancies(self, **kwargs):
-                captured_fetch_kwargs.update(kwargs)
+                captured_fetches.append(kwargs["company_inn"])
                 return [
                     GenericParserItem(
                         source=ParserLoadLog.Source.TRUDVSEM,
-                        external_id="trudvsem:1",
-                        inn="7701000401",
+                        external_id=f"trudvsem:{kwargs['company_inn']}",
+                        inn=kwargs["company_inn"],
                         title="Работа России",
                         payload={"vacancy_source": "trudvsem"},
-                    ),
-                    GenericParserItem(
-                        source=ParserLoadLog.Source.TRUDVSEM,
-                        external_id="hh:1",
-                        inn="7701000402",
-                        title="HeadHunter",
-                        payload={"vacancy_source": "hh"},
-                    ),
-                    GenericParserItem(
-                        source=ParserLoadLog.Source.TRUDVSEM,
-                        external_id="superjob:1",
-                        inn="7701000403",
-                        title="SuperJob",
-                        payload={"vacancy_source": "superjob"},
-                    ),
+                    )
                 ]
 
         original_client = parser_tasks.VacanciesClient
         parser_tasks.VacanciesClient = _VacanciesClient
         try:
-            result = parse_trudvsem_vacancies(
-                limit=25,
-                offset=5,
-                region_code="1",
-                text="инженер",
-                proxies=[],
-                vacancy_sources=["trudvsem", "hh", "superjob"],
-            )
+            with patch.object(
+                parser_tasks,
+                "VACANCY_REGISTRY_ORGANIZATIONS_PER_TASK",
+                1,
+                create=True,
+            ), patch.object(
+                parser_tasks.parse_trudvsem_vacancies,
+                "apply_async",
+                return_value=SimpleNamespace(id="vacancy-continuation"),
+            ) as apply_async_mock:
+                first_result = parse_trudvsem_vacancies(limit=25, proxies=[])
+                continuation_kwargs = apply_async_mock.call_args.kwargs["kwargs"]
+
+                self.assertEqual(first_result["status"], "in_progress")
+                self.assertEqual(captured_fetches, ["7701000501"])
+                self.assertEqual(OrganizationSourceRecord.objects.count(), 1)
+
+                job = BackgroundJob.objects.get(
+                    task_name="apps.parsers.tasks.parse_trudvsem_vacancies"
+                )
+                load = ParserLoadLog.objects.get(
+                    source=ParserLoadLog.Source.TRUDVSEM,
+                    batch_id=job.meta["batch_id"],
+                )
+                self.assertEqual(job.status, JobStatus.STARTED)
+                self.assertEqual(job.progress, 50)
+                self.assertEqual(job.meta["next_offset"], 1)
+                self.assertEqual(load.status, ParserLoadLog.Status.IN_PROGRESS)
+                self.assertEqual(load.records_count, 1)
+
+                final_result = parse_trudvsem_vacancies(**continuation_kwargs)
         finally:
             parser_tasks.VacanciesClient = original_client
 
-        self.assertEqual(result["status"], "success")
-        self.assertEqual(result["saved"], 3)
-        self.assertEqual(captured_kwargs["superjob_app_id"], "test-superjob-app-id")
-        self.assertEqual(
-            captured_kwargs["hh_user_agent"],
-            "Mostovik/1.0 (ops@example.test)",
+        self.assertEqual(final_result["status"], "success")
+        self.assertEqual(final_result["saved"], 2)
+        self.assertEqual(captured_fetches, ["7701000501", "7701000502"])
+        self.assertEqual(captured_kwargs["sources"], ["trudvsem"])
+        self.assertEqual(OrganizationSourceRecord.objects.count(), 2)
+        job.refresh_from_db()
+        load.refresh_from_db()
+        self.assertEqual(job.status, JobStatus.SUCCESS)
+        self.assertEqual(job.progress, 100)
+        self.assertEqual(load.status, ParserLoadLog.Status.SUCCESS)
+        self.assertEqual(load.records_count, 2)
+
+    def test_registry_run_preserves_and_resumes_records_after_revoke(self):
+        self._make_registry_organization(inn=7701000601, name="АО До остановки")
+        self._make_registry_organization(inn=7701000602, name="АО После остановки")
+        captured_fetches = []
+
+        class _VacanciesClient:
+            def __init__(self, **kwargs):
+                pass
+
+            def __enter__(self):
+                return self
+
+            def __exit__(self, exc_type, exc_val, exc_tb):
+                return None
+
+            def fetch_vacancies(self, **kwargs):
+                captured_fetches.append(kwargs["company_inn"])
+                return [
+                    GenericParserItem(
+                        source=ParserLoadLog.Source.TRUDVSEM,
+                        external_id=f"trudvsem:{kwargs['company_inn']}",
+                        inn=kwargs["company_inn"],
+                        title="Работа России",
+                        payload={"vacancy_source": "trudvsem"},
+                    )
+                ]
+
+        original_client = parser_tasks.VacanciesClient
+        original_save_records = GenericParserRecordService.save_records
+
+        def save_then_revoke(records, batch_id, *, source, chunk_size=500):
+            saved = original_save_records(
+                records,
+                batch_id,
+                source=source,
+                chunk_size=chunk_size,
+            )
+            job = BackgroundJob.objects.get(
+                task_name="apps.parsers.tasks.parse_trudvsem_vacancies"
+            )
+            job.revoke()
+            return saved
+
+        parser_tasks.VacanciesClient = _VacanciesClient
+        try:
+            with patch.object(
+                GenericParserRecordService,
+                "save_records",
+                side_effect=save_then_revoke,
+            ):
+                stopped_result = parse_trudvsem_vacancies(limit=25, proxies=[])
+
+            stopped_job = BackgroundJob.objects.get(
+                task_name="apps.parsers.tasks.parse_trudvsem_vacancies"
+            )
+            stopped_load = ParserLoadLog.objects.get(
+                source=ParserLoadLog.Source.TRUDVSEM,
+                batch_id=stopped_job.meta["batch_id"],
+            )
+            self.assertEqual(stopped_result["status"], "revoked")
+            self.assertEqual(stopped_job.meta["next_offset"], 1)
+            self.assertEqual(stopped_load.status, ParserLoadLog.Status.SKIPPED)
+            self.assertEqual(stopped_load.records_count, 1)
+            self.assertEqual(OrganizationSourceRecord.objects.count(), 1)
+
+            resumed_result = parse_trudvsem_vacancies(limit=25, proxies=[])
+        finally:
+            parser_tasks.VacanciesClient = original_client
+
+        self.assertEqual(resumed_result["status"], "success")
+        self.assertEqual(resumed_result["batch_id"], stopped_result["batch_id"])
+        self.assertTrue(resumed_result["resumed"])
+        self.assertEqual(captured_fetches, ["7701000601", "7701000602"])
+        self.assertEqual(OrganizationSourceRecord.objects.count(), 2)
+
+    def test_registry_run_fails_when_trudvsem_fails_for_every_organization(self):
+        self._make_registry_organization(inn=7701000701, name="АО Ошибка источника")
+
+        class _VacanciesClient:
+            def __init__(self, **kwargs):
+                pass
+
+            def __enter__(self):
+                return self
+
+            def __exit__(self, exc_type, exc_val, exc_tb):
+                return None
+
+            def fetch_vacancies(self, **kwargs):
+                raise RuntimeError("external 503")
+
+        original_client = parser_tasks.VacanciesClient
+        parser_tasks.VacanciesClient = _VacanciesClient
+        try:
+            result = parse_trudvsem_vacancies(limit=25, proxies=[])
+        finally:
+            parser_tasks.VacanciesClient = original_client
+
+        job = BackgroundJob.objects.get(
+            task_name="apps.parsers.tasks.parse_trudvsem_vacancies"
         )
-        self.assertEqual(captured_kwargs["sources"], ["trudvsem", "hh", "superjob"])
-        self.assertEqual(captured_fetch_kwargs["limit"], 25)
-        self.assertEqual(captured_fetch_kwargs["offset"], 5)
-        self.assertEqual(captured_fetch_kwargs["region_code"], "1")
-        self.assertEqual(captured_fetch_kwargs["text"], "инженер")
-        self.assertEqual(
-            set(
-                OrganizationSourceRecord.objects.values_list(
-                    "external_id",
-                    "source",
-                )
-            ),
-            {
-                ("trudvsem:1", "trudvsem"),
-                ("hh:1", "hh"),
-                ("superjob:1", "superjob"),
-            },
+        load = ParserLoadLog.objects.get(
+            source=ParserLoadLog.Source.TRUDVSEM,
+            batch_id=job.meta["batch_id"],
         )
+        self.assertEqual(result["status"], "failure")
+        self.assertEqual(job.status, JobStatus.FAILURE)
+        self.assertEqual(load.status, ParserLoadLog.Status.FAILED)
+        self.assertFalse(OrganizationSourceRecord.objects.exists())
 
 
 class ParserLoadLogServiceTestCase(TestCase):

tests/apps/parsers/test_vacancy_clients.py

@@ -196,7 +196,7 @@ class SuperJobVacanciesClientTest(SimpleTestCase):
 
 
 class VacanciesClientTest(SimpleTestCase):
-    def test_fetch_vacancies_combines_enabled_sources(self):
+    def test_fetch_vacancies_uses_only_trudvsem_by_default(self):
         class _Provider:
             def __init__(self, source_name: str):
                 self.source_name = source_name
@@ -222,7 +222,7 @@ class VacanciesClientTest(SimpleTestCase):
 
         self.assertEqual(
             [record.payload["vacancy_source"] for record in records],
-            ["trudvsem", "hh", "superjob"],
+            ["trudvsem"],
         )
 
     def test_fetch_vacancies_skips_sources_without_company_inn_support(self):
@@ -278,8 +278,11 @@ class VacanciesClientTest(SimpleTestCase):
 
         self.assertEqual([record.external_id for record in records], ["trudvsem:1"])
 
-    def test_fetch_vacancies_requires_superjob_app_id_when_explicitly_selected(self):
-        client = VacanciesClient(sources=["superjob"], superjob_app_id="")
+    def test_fetch_vacancies_rejects_disabled_sources(self):
+        client = VacanciesClient(sources=["hh", "superjob"])
 
-        with self.assertRaises(VacanciesClientError):
+        with self.assertRaisesMessage(
+            VacanciesClientError,
+            "Disabled vacancy sources: hh, superjob",
+        ):
             client.fetch_vacancies()