feat: limit source exports to current year
All checks were successful
CI/CD Pipeline / Code Quality Checks (push) Successful in 3m16s
CI/CD Pipeline / Run Tests (push) Successful in 4m58s
CI/CD Pipeline / Build and Push Dev Images (push) Successful in 32s
CI/CD Pipeline / Deploy Dev via Compose (push) Successful in 33s

This commit is contained in:
2026-08-04 23:25:36 +02:00
parent c2a09a0403
commit 2b9ecc2681
8 changed files with 316 additions and 30 deletions

View File

@@ -8,6 +8,7 @@ from django.apps import apps as django_apps
from django.conf import settings
from django.core.cache import cache
from django.test import TestCase, override_settings
from django.utils import timezone
from django_celery_beat.models import PeriodicTask
@@ -36,6 +37,7 @@ class SourceRecordExportArtifactsTaskTest(TestCase):
self.assertEqual(result["status"], "success")
self.assertEqual(result["artifacts_count"], 25)
self.assertEqual(result["export_year"], timezone.localdate().year)
self.assertIsNone(cache.get(settings.SOURCE_RECORD_EXPORT_LOCK_KEY))
def test_refresh_task_skips_when_generation_lock_is_held(self):

View File

@@ -2,19 +2,24 @@
import json
import zipfile
from datetime import UTC, datetime
from io import BytesIO, StringIO
from tempfile import TemporaryDirectory
from unittest.mock import patch
from apps.external_data.source_record_export import (
ORGANIZATION_EXPORT_FIELDS,
SOURCE_GROUP_EXPORT_SPECS,
SOURCE_RECORD_EXPORT_FIELDS,
SourceRecordExportArtifactsUnavailable,
_source_group_headers,
build_source_record_export_artifacts,
build_source_records_export_archive,
load_current_source_record_export_generation,
)
from django.core.management import call_command
from django.test import override_settings
from django.utils import timezone
from openpyxl import load_workbook
from rest_framework import status
from rest_framework.test import APITestCase
@@ -28,6 +33,7 @@ from tests.apps.external_data.factories import (
IndustrialProductFactory,
ManufacturerRegistryEntryFactory,
ProsecutorCheckFactory,
PublicProcurementFactory,
)
from tests.apps.organization.factories import OrganizationFactory
from tests.apps.user.factories import UserFactory
@@ -71,23 +77,38 @@ class SourceRecordExportApiTest(APITestCase):
self.assertEqual(response["Retry-After"], "3600")
def test_generation_builds_full_matrix_from_normalized_tables(self):
current_date = timezone.localdate()
organization = OrganizationFactory.create(
full_name='Акционерное общество "Экспорт"',
okpo="12345678",
)
IndustrialProductFactory.create(organization=organization)
IndustrialCertificateFactory.create(organization=organization)
IndustrialCertificateFactory.create(
organization=organization,
issue_date=current_date,
)
ManufacturerRegistryEntryFactory.create(organization=organization)
ProsecutorCheckFactory.create(organization=organization)
arbitration_case = ArbitrationCaseFactory.create(organization=organization)
ProsecutorCheckFactory.create(
organization=organization,
start_date=current_date,
)
arbitration_case = ArbitrationCaseFactory.create(
organization=organization,
decision_date=current_date,
)
report = FinancialReportFactory.create(organization=organization)
FinancialReportLineFactory.create(report=report, line_code="1600")
FinancialReportLineFactory.create(
report=report,
line_code="1600",
year=current_date.year,
)
generation = build_source_record_export_artifacts()
self.assertEqual(generation.artifacts_count, 25)
self.assertEqual(generation.files_count, 25)
self.assertEqual(generation.records_count, 6)
self.assertEqual(generation.export_year, current_date.year)
expected_prefix = [*ORGANIZATION_EXPORT_FIELDS, *SOURCE_RECORD_EXPORT_FIELDS]
for source_spec in SOURCE_GROUP_EXPORT_SPECS.values():
self.assertEqual(
@@ -229,14 +250,104 @@ class SourceRecordExportApiTest(APITestCase):
generation = load_current_source_record_export_generation()
self.assertEqual(generation.artifacts_count, 25)
self.assertEqual(generation.export_year, timezone.localdate().year)
self.assertIn('"artifacts_count": 25', command_output.getvalue())
def test_generation_contains_only_records_from_its_calendar_year(self):
export_year = 2026
generated_at = datetime(export_year, 8, 4, 6, 0, tzinfo=UTC)
organization = OrganizationFactory.create()
current_procurement = PublicProcurementFactory.create(
organization=organization,
contract_date=datetime(2026, 3, 1, tzinfo=UTC).date(),
)
PublicProcurementFactory.create(
organization=organization,
contract_date=datetime(2025, 3, 1, tzinfo=UTC).date(),
)
current_product = IndustrialProductFactory.create(organization=organization)
old_product = IndustrialProductFactory.create(organization=organization)
current_product.__class__.objects.filter(pk=current_product.pk).update(
created_at=generated_at
)
old_product.__class__.objects.filter(pk=old_product.pk).update(
created_at=datetime(2025, 8, 4, 6, 0, tzinfo=UTC)
)
current_report = FinancialReportFactory.create(organization=organization)
old_report = FinancialReportFactory.create(organization=organization)
FinancialReportLineFactory.create(report=current_report, year=2025)
FinancialReportLineFactory.create(
report=current_report,
year=export_year,
line_code="1601",
)
FinancialReportLineFactory.create(report=old_report, year=2025)
generation = build_source_record_export_artifacts(now=generated_at)
self.assertEqual(generation.export_year, export_year)
procurements_path = next(
artifact.path
for artifact in generation.artifacts
if artifact.source_group == "government_procurements"
and artifact.file_format == "json"
)
procurement_rows = json.loads(procurements_path.read_text(encoding="utf-8"))
self.assertEqual(
[row["uid"] for row in procurement_rows],
[str(current_procurement.pk)],
)
industrial_path = next(
artifact.path
for artifact in generation.artifacts
if artifact.source_group == "industrial_production"
and artifact.file_format == "json"
)
industrial_rows = json.loads(industrial_path.read_text(encoding="utf-8"))
self.assertEqual(
[row["uid"] for row in industrial_rows],
[str(current_product.pk)],
)
financial_path = next(
artifact.path
for artifact in generation.artifacts
if artifact.source_group == "financial_indicators"
)
financial_rows = json.loads(financial_path.read_text(encoding="utf-8"))
self.assertEqual(
[row["uid"] for row in financial_rows],
[str(current_report.pk)],
)
self.assertEqual(
{line["year"] for line in financial_rows[0]["financial_lines"]},
{export_year},
)
def test_new_calendar_year_requires_a_new_prepared_generation(self):
build_source_record_export_artifacts(
now=datetime(2026, 12, 31, 23, 59, tzinfo=UTC)
)
with self.assertRaises(SourceRecordExportArtifactsUnavailable):
build_source_records_export_archive(
source_groups=["planned_inspections"],
export_format="json",
requested_at=datetime(2027, 1, 1, tzinfo=UTC),
)
def test_admin_streams_selected_prepared_files_without_database_queries(self):
self.client.force_authenticate(UserFactory.create_superuser())
organization = OrganizationFactory.create(okpo="87654321")
ProsecutorCheckFactory.create(organization=organization)
current_date = timezone.localdate()
ProsecutorCheckFactory.create(
organization=organization,
start_date=current_date,
)
report = FinancialReportFactory.create(organization=organization)
FinancialReportLineFactory.create(report=report)
FinancialReportLineFactory.create(report=report, year=current_date.year)
generation = build_source_record_export_artifacts()
with self.assertNumQueries(0):
@@ -274,7 +385,8 @@ class SourceRecordExportApiTest(APITestCase):
self.assertEqual(rows[1][4], "87654321")
def test_ticket_is_admin_only_and_can_be_consumed_once_without_auth(self):
build_source_record_export_artifacts()
requested_at = datetime(2026, 8, 4, 12, 34, 56, tzinfo=UTC)
build_source_record_export_artifacts(now=requested_at)
regular_user = UserFactory.create_user()
self.client.force_authenticate(regular_user)
forbidden_response = self.client.post(
@@ -285,7 +397,13 @@ class SourceRecordExportApiTest(APITestCase):
self.assertEqual(forbidden_response.status_code, status.HTTP_403_FORBIDDEN)
self.client.force_authenticate(UserFactory.create_superuser())
with self.assertNumQueries(0):
with (
patch(
"apps.external_data.source_record_export.timezone.now",
return_value=requested_at,
),
self.assertNumQueries(0),
):
ticket_response = self.client.post(
self.ticket_url,
{"sources": ["bankruptcy"], "format": "json"},
@@ -294,6 +412,10 @@ class SourceRecordExportApiTest(APITestCase):
self.assertEqual(ticket_response.status_code, status.HTTP_201_CREATED)
self.assertRegex(ticket_response.data["ticket"], r"^[A-Za-z0-9_-]{43}$")
self.assertEqual(ticket_response.data["expires_in"], 300)
self.assertEqual(
ticket_response.data["file_name"],
"bankruptcy-procedures_20260804_123456.zip",
)
self.client.force_authenticate(user=None)
with self.assertNumQueries(0):
@@ -303,6 +425,10 @@ class SourceRecordExportApiTest(APITestCase):
format="multipart",
)
self.assertEqual(download_response.status_code, status.HTTP_200_OK)
self.assertEqual(
download_response["Content-Disposition"],
'attachment; filename="bankruptcy-procedures_20260804_123456.zip"',
)
with zipfile.ZipFile(
BytesIO(self._response_body(download_response))
) as archive:
@@ -319,7 +445,11 @@ class SourceRecordExportApiTest(APITestCase):
@override_settings(SOURCE_RECORD_EXPORT_XLSX_ROWS_PER_FILE=2)
def test_xlsx_is_split_into_bounded_workbook_parts(self):
organization = OrganizationFactory.create()
ProsecutorCheckFactory.create_batch(3, organization=organization)
ProsecutorCheckFactory.create_batch(
3,
organization=organization,
start_date=timezone.localdate(),
)
generation = build_source_record_export_artifacts()