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

@@ -20,6 +20,7 @@ class Command(BaseAppCommand):
{
"generation_id": generation.generation_id,
"generated_at": generation.generated_at,
"export_year": generation.export_year,
"artifacts_count": generation.artifacts_count,
"files_count": generation.files_count,
"records_count": generation.records_count,

View File

@@ -0,0 +1,34 @@
"""Index the remaining date fields used by current-year exports."""
from django.db import migrations, models
class Migration(migrations.Migration):
"""Add indexes for nullable certificate and security-registry dates."""
dependencies = [
("external_data", "0007_seed_nightly_source_record_exports"),
]
operations = [
migrations.AlterField(
model_name="industrialcertificate",
name="issue_date",
field=models.DateField(
blank=True,
db_index=True,
null=True,
verbose_name="дата выдачи",
),
),
migrations.AlterField(
model_name="informationsecurityregistryentry",
name="issued_at",
field=models.DateField(
blank=True,
db_index=True,
null=True,
verbose_name="дата выдачи",
),
),
]

View File

@@ -44,7 +44,12 @@ class IndustrialCertificate(UUIDPrimaryKeyMixin, TimestampMixin, models.Model):
certificate_number = models.CharField(
_("номер сертификата"), max_length=100, db_index=True
)
issue_date = models.DateField(_("дата выдачи"), null=True, blank=True)
issue_date = models.DateField(
_("дата выдачи"),
null=True,
blank=True,
db_index=True,
)
expiry_date = models.DateField(_("дата окончания"), null=True, blank=True)
certificate_file_url = models.TextField(
_("ссылка на файл сертификата"), blank=True, default=""
@@ -254,7 +259,12 @@ class InformationSecurityRegistryEntry(
blank=True,
default="",
)
issued_at = models.DateField(_("дата выдачи"), null=True, blank=True)
issued_at = models.DateField(
_("дата выдачи"),
null=True,
blank=True,
db_index=True,
)
expires_at = models.DateField(_("дата окончания"), null=True, blank=True)
class Meta:

View File

@@ -24,6 +24,7 @@ from apps.external_data.models import (
BankruptcyProcedure,
DefenseUnreliableSupplier,
FinancialReport,
FinancialReportLine,
IndustrialCertificate,
IndustrialProduct,
InformationSecurityRegistryEntry,
@@ -34,7 +35,7 @@ from apps.external_data.models import (
)
from django.conf import settings
from django.core.cache import cache
from django.db.models import QuerySet, prefetch_related_objects
from django.db.models import Prefetch, Q, QuerySet, prefetch_related_objects
from django.utils import timezone
from openpyxl import Workbook
@@ -43,7 +44,7 @@ EXPORT_FORMAT_XLSX = "xlsx"
EXPORT_FORMAT_JSON = "json"
EXPORT_FORMATS = (EXPORT_FORMAT_CSV, EXPORT_FORMAT_XLSX, EXPORT_FORMAT_JSON)
FINANCIAL_SOURCE_GROUP = "financial_indicators"
EXPORT_MANIFEST_VERSION = 1
EXPORT_MANIFEST_VERSION = 2
CURRENT_EXPORT_MANIFEST_FILE_NAME = "current.json"
GENERATION_MANIFEST_FILE_NAME = "manifest.json"
GENERATION_DIRECTORY_NAME = "generations"
@@ -117,6 +118,7 @@ class SourceModelExportSpec:
payload_aliases: tuple[tuple[str, str], ...] = ()
payload_export_fields: tuple[str, ...] | None = None
prefetch_related: tuple[str, ...] = ()
export_year_lookup: str | None = None
@dataclass(frozen=True)
@@ -153,6 +155,7 @@ SOURCE_GROUP_EXPORT_SPECS: dict[str, SourceGroupExportSpec] = {
status_field="status",
load_batch_field="load_batch",
prefetch_related=("lines",),
export_year_lookup="lines__year",
),
),
),
@@ -436,6 +439,7 @@ class SourceRecordExportGeneration:
generation_id: str
generated_at: str
export_year: int
artifacts: tuple[SourceRecordExportArtifact, ...]
records_count: int
@@ -514,6 +518,7 @@ def build_source_record_export_artifacts(
generations_directory.mkdir(parents=True, exist_ok=True)
generated_at_datetime = _normalize_generation_datetime(now or timezone.now())
export_year = _export_year(generated_at_datetime)
generation_id = (
f"{generated_at_datetime.strftime('%Y%m%dT%H%M%SZ')}-{uuid4().hex[:8]}"
)
@@ -530,6 +535,7 @@ def build_source_record_export_artifacts(
headers, records_count = _spool_source_group_rows(
source_spec=source_spec,
output_path=row_spool_path,
export_year=export_year,
)
source_record_counts[source_group] = records_count
@@ -569,6 +575,7 @@ def build_source_record_export_artifacts(
generation = SourceRecordExportGeneration(
generation_id=generation_id,
generated_at=generated_at_datetime.isoformat(),
export_year=export_year,
artifacts=tuple(artifacts),
records_count=sum(source_record_counts.values()),
)
@@ -622,6 +629,7 @@ def build_source_records_export_archive(
source_groups: Sequence[str],
export_format: str,
export_directory: str | Path | None = None,
requested_at: datetime | None = None,
) -> SourceRecordExportArchive:
"""Package selected prepared files without querying external-data tables."""
@@ -629,6 +637,13 @@ def build_source_records_export_archive(
generation = load_current_source_record_export_generation(
export_directory=root_directory,
)
requested_at_datetime = _normalize_generation_datetime(
requested_at or timezone.now()
)
if generation.export_year != _export_year(requested_at_datetime):
raise SourceRecordExportArtifactsUnavailable(
"Prepared source-record export belongs to a different calendar year."
)
artifacts_by_key: dict[
tuple[str, str],
list[SourceRecordExportArtifact],
@@ -654,10 +669,11 @@ def build_source_records_export_archive(
sorted(artifacts, key=lambda artifact: artifact.part_number)
)
generated_at = datetime.fromisoformat(generation.generated_at)
timestamp = generated_at.strftime("%Y%m%d_%H%M%S")
return SourceRecordExportArchive(
archive_name=f"organization_source_records_export_{timestamp}.zip",
archive_name=_build_source_records_archive_name(
source_groups=source_groups,
requested_at=requested_at_datetime,
),
archive_chunks=_stream_zip_archive(selected_artifacts),
files_count=len(selected_artifacts),
generated_at=generation.generated_at,
@@ -671,9 +687,11 @@ def create_source_record_export_download_ticket(
) -> SourceRecordExportDownloadTicket:
"""Validate prepared files and cache a short-lived download capability."""
requested_at = _normalize_generation_datetime(timezone.now())
package = build_source_records_export_archive(
source_groups=source_groups,
export_format=export_format,
requested_at=requested_at,
)
expires_in = max(
1,
@@ -685,7 +703,11 @@ def create_source_record_export_download_ticket(
)
),
)
payload = {"sources": list(source_groups), "format": export_format}
payload = {
"sources": list(source_groups),
"format": export_format,
"requested_at": requested_at.isoformat(),
}
for _attempt in range(3):
ticket = secrets.token_urlsafe(32)
if cache.add(
@@ -718,6 +740,7 @@ def consume_source_record_export_download_ticket(
try:
source_groups = payload["sources"]
export_format = payload["format"]
requested_at_value = payload["requested_at"]
if (
not isinstance(source_groups, list)
or not source_groups
@@ -728,14 +751,19 @@ def consume_source_record_export_download_ticket(
)
or len(source_groups) != len(set(source_groups))
or export_format not in EXPORT_FORMATS
or not isinstance(requested_at_value, str)
):
raise ValueError
requested_at = datetime.fromisoformat(requested_at_value)
if timezone.is_naive(requested_at):
raise ValueError
except (KeyError, TypeError, ValueError):
raise SourceRecordExportTicketInvalid from None
return build_source_records_export_archive(
source_groups=source_groups,
export_format=export_format,
requested_at=requested_at,
)
@@ -788,6 +816,24 @@ def _normalize_generation_datetime(value: datetime) -> datetime:
return value.astimezone(UTC)
def _export_year(value: datetime) -> int:
"""Return the calendar year in the configured service timezone."""
return timezone.localtime(value).year
def _build_source_records_archive_name(
*,
source_groups: Sequence[str],
requested_at: datetime,
) -> str:
source_name = "__".join(
SOURCE_GROUP_EXPORT_SPECS[source_group].file_stem
for source_group in source_groups
)
timestamp = timezone.localtime(requested_at).strftime("%Y%m%d_%H%M%S")
return f"{source_name}_{timestamp}.zip"
def _source_group_export_formats(source_group: str) -> tuple[str, ...]:
if source_group == FINANCIAL_SOURCE_GROUP:
return (EXPORT_FORMAT_JSON,)
@@ -808,25 +854,59 @@ def _build_source_group_file_name(*, source_group: str, file_format: str) -> str
return f"{SOURCE_GROUP_EXPORT_SPECS[source_group].file_stem}.{file_format}"
def _source_model_queryset(model_spec: SourceModelExportSpec) -> QuerySet:
return cast(
def _source_model_queryset(
model_spec: SourceModelExportSpec,
*,
export_year: int,
) -> QuerySet:
queryset = cast(
QuerySet,
model_spec.model.objects.select_related("organization").order_by(),
)
if model_spec.export_year_lookup:
return queryset.filter(
**{model_spec.export_year_lookup: export_year}
).distinct()
if model_spec.record_date_field:
return queryset.filter(
Q(**{f"{model_spec.record_date_field}__year": export_year})
| Q(
**{
f"{model_spec.record_date_field}__isnull": True,
"created_at__year": export_year,
}
)
)
return queryset.filter(created_at__year=export_year)
def _iter_source_model_records(
model_spec: SourceModelExportSpec,
*,
export_year: int,
) -> Iterator[Any]:
iterator = _source_model_queryset(model_spec).iterator(
chunk_size=SOURCE_RECORD_EXPORT_ITERATOR_CHUNK_SIZE
)
iterator = _source_model_queryset(
model_spec,
export_year=export_year,
).iterator(chunk_size=SOURCE_RECORD_EXPORT_ITERATOR_CHUNK_SIZE)
while True:
batch = list(islice(iterator, SOURCE_RECORD_EXPORT_ITERATOR_CHUNK_SIZE))
if not batch:
return
if model_spec.prefetch_related:
prefetch_related_objects(batch, *model_spec.prefetch_related)
prefetches = [
Prefetch(
related_name,
queryset=FinancialReportLine.objects.filter(
year=export_year
).order_by(),
)
if related_name == "lines"
and model_spec.export_year_lookup == "lines__year"
else related_name
for related_name in model_spec.prefetch_related
]
prefetch_related_objects(batch, *prefetches)
yield from batch
@@ -851,13 +931,17 @@ def _spool_source_group_rows(
*,
source_spec: SourceGroupExportSpec,
output_path: Path,
export_year: int,
) -> tuple[list[str], int]:
records_count = 0
with output_path.open("w", encoding="utf-8", newline="") as output:
output.write("[")
is_first_row = True
for model_spec in source_spec.models:
for record in _iter_source_model_records(model_spec):
for record in _iter_source_model_records(
model_spec,
export_year=export_year,
):
row = _build_record_row(
record,
source_spec=source_spec,
@@ -1149,6 +1233,7 @@ def _generation_manifest_payload(
"version": EXPORT_MANIFEST_VERSION,
"generation_id": generation.generation_id,
"generated_at": generation.generated_at,
"export_year": generation.export_year,
"records_count": generation.records_count,
"artifacts_count": generation.artifacts_count,
"files_count": generation.files_count,
@@ -1180,6 +1265,9 @@ def _generation_from_manifest(
generation_id = str(payload["generation_id"])
generated_at = str(payload["generated_at"])
datetime.fromisoformat(generated_at)
export_year = int(payload["export_year"])
if not 1 <= export_year <= 9999:
raise ValueError("Source-record export year is invalid.")
records_count = int(payload["records_count"])
artifact_payloads = payload["artifacts"]
if not isinstance(artifact_payloads, list):
@@ -1234,6 +1322,7 @@ def _generation_from_manifest(
return SourceRecordExportGeneration(
generation_id=generation_id,
generated_at=generated_at,
export_year=export_year,
artifacts=tuple(artifacts),
records_count=records_count,
)

View File

@@ -35,6 +35,7 @@ def refresh_source_record_export_artifacts(self) -> dict: # noqa: ARG001
"status": "success",
"generation_id": generation.generation_id,
"generated_at": generation.generated_at,
"export_year": generation.export_year,
"artifacts_count": generation.artifacts_count,
"files_count": generation.files_count,
"records_count": generation.records_count,