feat: harden datasource materialization payloads

This commit is contained in:
2026-07-29 19:38:34 +02:00
parent 5f0847b8ca
commit f3288dbd28
9 changed files with 1460 additions and 179 deletions
+133
View File
@@ -17,11 +17,14 @@ from govoplan_core.core.datasources import (
DatasourcePublicationRequest,
DatasourceReadRequest,
DatasourceStageInput,
DatasourceUnavailableError,
DatasourceValidationError,
)
from govoplan_core.db.base import Base, utcnow
from govoplan_datasources.backend.db.models import (
DatasourceMaterializationRecord,
DatasourcePayloadRecord,
DatasourcePayloadRowRecord,
DatasourcePublicationRecord,
DatasourceRecord,
DatasourceStageRecord,
@@ -32,6 +35,12 @@ from govoplan_datasources.backend.service import (
STAGE_WRITE_SCOPE,
SqlDatasourceProvider,
)
from govoplan_datasources.backend.payloads import (
create_database_rows_payload,
finalize_payload_deletion,
mark_unreferenced_payload_for_deletion,
verify_payload_integrity,
)
def principal(
@@ -137,6 +146,8 @@ class DatasourceLifecycleTests(unittest.TestCase):
self.engine,
tables=[
DatasourceRecord.__table__,
DatasourcePayloadRecord.__table__,
DatasourcePayloadRowRecord.__table__,
DatasourceMaterializationRecord.__table__,
DatasourceStageRecord.__table__,
DatasourcePublicationRecord.__table__,
@@ -157,6 +168,8 @@ class DatasourceLifecycleTests(unittest.TestCase):
DatasourceStageRecord.__table__,
DatasourcePublicationRecord.__table__,
DatasourceMaterializationRecord.__table__,
DatasourcePayloadRowRecord.__table__,
DatasourcePayloadRecord.__table__,
DatasourceRecord.__table__,
],
)
@@ -235,6 +248,18 @@ class DatasourceLifecycleTests(unittest.TestCase):
current.datasource.schema_version,
)
self.assertEqual(frozen.ref, frozen_result.materialization.ref)
first_record = self.session.get(
DatasourceMaterializationRecord,
first.ref.removeprefix("materialization:"),
)
frozen_record = self.session.get(
DatasourceMaterializationRecord,
frozen.ref.removeprefix("materialization:"),
)
self.assertIsNotNone(first_record)
self.assertIsNotNone(frozen_record)
self.assertEqual(first_record.payload_id, frozen_record.payload_id)
self.assertEqual([], first_record.rows)
def test_live_reads_origin_and_cached_refresh_is_explicit(self) -> None:
live = self.provider.register_origin(
@@ -445,6 +470,114 @@ class DatasourceLifecycleTests(unittest.TestCase):
),
)
def test_payload_preview_is_paged_and_metadata_mismatch_is_rejected(self) -> None:
stage = self.provider.create_stage(
self.session,
principal(),
stage=DatasourceStageInput(
name="Paged",
source_name="paged",
kind="upload",
mode="static",
shape="tabular",
rows=tuple({"id": index} for index in range(20)),
),
)
datasource, materialization = self.provider.promote_stage(
self.session,
principal(),
stage_ref=stage.ref,
)
self.session.commit()
result = self.provider.read_datasource(
self.session,
principal(),
request=DatasourceReadRequest(
datasource_ref=datasource.ref,
offset=7,
limit=3,
),
)
self.assertEqual([{"id": 7}, {"id": 8}, {"id": 9}], list(result.rows))
record = self.session.get(
DatasourceMaterializationRecord,
materialization.ref.removeprefix("materialization:"),
)
self.assertIsNotNone(record)
self.assertEqual([], record.rows)
self.assertEqual(
20,
self.session.query(DatasourcePayloadRowRecord)
.filter(DatasourcePayloadRowRecord.payload_id == record.payload_id)
.count(),
)
payload = self.session.get(DatasourcePayloadRecord, record.payload_id)
self.assertIsNotNone(payload)
verify_payload_integrity(self.session, payload)
payload.row_count += 1
self.session.flush()
with self.assertRaises(DatasourceUnavailableError):
self.provider.read_datasource(
self.session,
principal(),
request=DatasourceReadRequest(datasource_ref=datasource.ref),
)
def test_payload_deletion_is_staged_and_reference_safe(self) -> None:
payload = create_database_rows_payload(
self.session,
tenant_id="tenant-1",
rows=({"id": 1}, {"id": 2}),
actor_id="account-1",
)
self.session.commit()
self.assertTrue(
mark_unreferenced_payload_for_deletion(self.session, payload)
)
self.assertEqual("deleting", payload.state)
self.session.commit()
finalize_payload_deletion(self.session, payload)
self.session.commit()
self.assertEqual(0, self.session.query(DatasourcePayloadRecord).count())
self.assertEqual(
0,
self.session.query(DatasourcePayloadRowRecord).count(),
)
def test_rolled_back_materialization_leaves_no_payload_rows(self) -> None:
stage = self.provider.create_stage(
self.session,
principal(),
stage=DatasourceStageInput(
name="Rollback",
source_name="rollback",
kind="upload",
mode="static",
shape="tabular",
rows=({"id": 1},),
),
)
self.provider.promote_stage(
self.session,
principal(),
stage_ref=stage.ref,
)
self.session.rollback()
self.assertEqual(0, self.session.query(DatasourcePayloadRecord).count())
self.assertEqual(
0,
self.session.query(DatasourcePayloadRowRecord).count(),
)
self.assertEqual(
0,
self.session.query(DatasourceMaterializationRecord).count(),
)
if __name__ == "__main__":
unittest.main()
+3 -1
View File
@@ -24,13 +24,15 @@ class DatasourceMigrationTests(unittest.TestCase):
try:
with engine.connect() as connection:
self.assertIn(
"c4e9f1a7d2b6",
"d5f0a2b8c3e7",
set(MigrationContext.configure(connection).get_current_heads()),
)
self.assertEqual(
{
"datasource_catalogue",
"datasource_materializations",
"datasource_payload_rows",
"datasource_payloads",
"datasource_publications",
"datasource_stages",
},
@@ -0,0 +1,122 @@
from __future__ import annotations
import os
import threading
import unittest
import uuid
from concurrent.futures import ThreadPoolExecutor
from sqlalchemy import create_engine, text
from sqlalchemy.orm import Session
from govoplan_core.db.base import Base
from govoplan_datasources.backend.db.models import (
DatasourceMaterializationRecord,
DatasourcePayloadRecord,
DatasourcePayloadRowRecord,
DatasourceRecord,
)
from govoplan_datasources.backend.service import _append_materialization
from govoplan_datasources.backend.tabular import (
encoded_size,
field_payload,
fingerprint_rows,
infer_schema,
)
@unittest.skipUnless(
os.environ.get("GOVOPLAN_DATASOURCES_TEST_POSTGRES_URL"),
"set GOVOPLAN_DATASOURCES_TEST_POSTGRES_URL for PostgreSQL concurrency checks",
)
class DatasourceMaterializationPostgresTests(unittest.TestCase):
def setUp(self) -> None:
database_url = os.environ["GOVOPLAN_DATASOURCES_TEST_POSTGRES_URL"]
self.schema = f"datasource_revision_race_{uuid.uuid4().hex}"
self.admin_engine = create_engine(database_url)
with self.admin_engine.begin() as connection:
connection.execute(text(f'CREATE SCHEMA "{self.schema}"'))
self.engine = create_engine(
database_url,
connect_args={"options": f"-c search_path={self.schema}"},
)
self.tables = [
DatasourceRecord.__table__,
DatasourcePayloadRecord.__table__,
DatasourcePayloadRowRecord.__table__,
DatasourceMaterializationRecord.__table__,
]
Base.metadata.create_all(self.engine, tables=self.tables)
with Session(self.engine) as session:
session.add(
DatasourceRecord(
id="datasource-1",
tenant_id="tenant-1",
source_name="concurrent",
name="Concurrent",
kind="custom",
mode="static",
shape="tabular",
)
)
session.commit()
def tearDown(self) -> None:
try:
Base.metadata.drop_all(
self.engine,
tables=list(reversed(self.tables)),
)
finally:
self.engine.dispose()
with self.admin_engine.begin() as connection:
connection.execute(text(f'DROP SCHEMA "{self.schema}"'))
self.admin_engine.dispose()
def test_competing_publishers_receive_distinct_monotonic_revisions(self) -> None:
barrier = threading.Barrier(2)
def publish(value: int) -> int:
rows = ({"id": value},)
schema = infer_schema(rows)
with Session(self.engine) as session:
datasource = session.get(DatasourceRecord, "datasource-1")
self.assertIsNotNone(datasource)
barrier.wait(timeout=10)
materialization = _append_materialization(
session,
datasource=datasource,
rows=rows,
schema=[field_payload(field) for field in schema],
fingerprint=fingerprint_rows(rows, schema),
byte_count=encoded_size(rows),
actor_id=f"worker-{value}",
set_current=True,
)
revision = materialization.revision
session.commit()
return revision
with ThreadPoolExecutor(max_workers=2) as executor:
revisions = tuple(executor.map(publish, (1, 2)))
self.assertEqual((1, 2), tuple(sorted(revisions)))
with Session(self.engine) as session:
self.assertEqual(
[1, 2],
list(
session.scalars(
DatasourceMaterializationRecord.__table__.select()
.with_only_columns(
DatasourceMaterializationRecord.revision
)
.order_by(
DatasourceMaterializationRecord.revision.asc()
)
)
),
)
if __name__ == "__main__":
unittest.main()