fix(parsers): search registry vacancies across job boards

tests/apps/parsers/test_tasks.py

@@ -2191,6 +2191,160 @@ 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 = {}
+
+        class _Provider:
+            def __init__(self, source_name, *, supports_company_inn):
+                self.source_name = source_name
+                self.supports_company_inn = supports_company_inn
+
+            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"},
+                        )
+                    ]
+
+                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(
+            set(
+                GenericParserRecord.objects.values_list(
+                    "external_id",
+                    "source",
+                    "registry_organization_id",
+                )
+            ),
+            {
+                ("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")
     def test_parse_trudvsem_vacancies_uses_combined_vacancies_client(self):
         captured_kwargs = {}