feat: move report artifacts to OSS storage

This commit is contained in:
Wyndham ARR
2026-08-04 12:37:26 +08:00
parent 727643f1f2
commit a2b86cdf10
45 changed files with 1288 additions and 1417 deletions

View File

@@ -6,7 +6,7 @@ from pathlib import Path
from company_reports.contracts import COMPANY_NAMES, REPORT_HEADERS, WarningCode
from company_reports.core import build_all_company_reports
from company_reports.publishing import ArtifactToolBuilder
from company_reports.publishing import OpenpyxlWorkbookBuilder
from tests.company_reports_acceptance_fixture import (
AS_OF_DATE,
REPORT_MONTH,
@@ -15,10 +15,6 @@ from tests.company_reports_acceptance_fixture import (
)
PROJECT_ROOT = Path(__file__).resolve().parents[1]
BUILDER_SCRIPT = PROJECT_ROOT / "company_reports" / "xlsx" / "build_workbook.mjs"
def acceptance_reports():
return build_all_company_reports(
REPORT_YEAR,
@@ -88,10 +84,7 @@ class CompanyReportAcceptanceCoreTests(unittest.TestCase):
class CompanyReportAcceptanceWorkbookTests(unittest.TestCase):
def test_real_builder_exports_and_reopens_all_five_workbooks(self):
builder = ArtifactToolBuilder(
BUILDER_SCRIPT,
timeout_seconds=180,
)
builder = OpenpyxlWorkbookBuilder()
with tempfile.TemporaryDirectory(prefix="company-report-acceptance-") as temp_dir:
root = Path(temp_dir)
for report in acceptance_reports():

View File

@@ -9,7 +9,7 @@ from datetime import date
from pathlib import Path
from company_reports.publishing import (
ArtifactToolBuilder,
OpenpyxlWorkbookBuilder,
AtomicReportPublisher,
BuiltWorkbook,
sha256_file,
@@ -19,10 +19,6 @@ from company_reports.service import CompanyReportService, RunRequest
from tests.test_company_reports_service import FakeRepository, synthetic_snapshot
PROJECT_ROOT = Path(__file__).resolve().parents[1]
BUILDER_SCRIPT = PROJECT_ROOT / "company_reports" / "xlsx" / "build_workbook.mjs"
class CompanyReportIntegrationTests(unittest.TestCase):
def test_missing_and_unmatched_group_codes_export_blank_booking_room(self):
class CapturingBuilder:
@@ -46,7 +42,7 @@ class CompanyReportIntegrationTests(unittest.TestCase):
)
)
builder = CapturingBuilder(
ArtifactToolBuilder(BUILDER_SCRIPT, timeout_seconds=90)
OpenpyxlWorkbookBuilder()
)
publisher = AtomicReportPublisher(project_root, output_root)
service = CompanyReportService(
@@ -114,7 +110,7 @@ class CompanyReportIntegrationTests(unittest.TestCase):
project_root.mkdir(parents=True)
repository = StableVersionRepository(synthetic_snapshot())
builder = RebuildingBuilder(
ArtifactToolBuilder(BUILDER_SCRIPT, timeout_seconds=90)
OpenpyxlWorkbookBuilder()
)
publisher = AtomicReportPublisher(project_root, output_root)
service = CompanyReportService(

View File

@@ -19,7 +19,7 @@ from company_reports.contracts import (
)
from company_reports.core import build_company_report
from company_reports.publishing import (
ArtifactToolBuilder,
OpenpyxlWorkbookBuilder,
AtomicReportPublisher,
BuiltWorkbook,
PublicationError,
@@ -112,14 +112,11 @@ def synthetic_built_workbook(
class CompanyReportPublishingTests(unittest.TestCase):
def test_builder_creates_xlsx_without_private_artifact_tool_runtime(self):
def test_builder_creates_xlsx_with_openpyxl_only(self):
with tempfile.TemporaryDirectory(prefix="company-report-builder-test-") as temp_dir:
root = Path(temp_dir)
report = workbook_report()
builder = ArtifactToolBuilder(
Path("/private/artifact-tool/build_workbook.mjs"),
node_binary="/definitely/not/node",
)
builder = OpenpyxlWorkbookBuilder()
built = builder.build(report, root)

View File

@@ -1,119 +0,0 @@
from __future__ import annotations
import json
import os
import shutil
import subprocess
import tempfile
import unittest
from pathlib import Path
PROJECT_ROOT = Path(__file__).resolve().parents[1]
BUILDER = PROJECT_ROOT / "company_reports" / "xlsx" / "build_workbook.mjs"
FIXTURE = (
PROJECT_ROOT
/ "tests"
/ "fixtures"
/ "company_reports"
/ "synthetic_qbd_payload.json"
)
ARTIFACT_PACKAGE = (
PROJECT_ROOT
/ "company_reports"
/ "xlsx"
/ "node_modules"
/ "@oai"
/ "artifact-tool"
)
def artifact_tool_available() -> bool:
configured = os.environ.get("COMPANY_REPORT_ARTIFACT_TOOL_MODULE", "").strip()
return ARTIFACT_PACKAGE.exists() or bool(configured and Path(configured).is_file())
def node_binary() -> str:
configured = os.environ.get("COMPANY_REPORT_NODE_BINARY", "").strip()
return configured or shutil.which("node") or ""
@unittest.skipUnless(artifact_tool_available(), "artifact-tool dependency is not installed")
class CompanyReportXlsxTests(unittest.TestCase):
def test_builder_exports_reopens_checks_and_renders_all_sheets(self):
node = node_binary()
if not node:
self.skipTest("Node.js is unavailable")
with tempfile.TemporaryDirectory(prefix="company-report-xlsx-test-") as temp_dir:
root = Path(temp_dir)
output = root / "QBD-July-2026.xlsx"
previews = root / "previews"
summary_path = root / "summary.json"
completed = subprocess.run(
[
node,
str(BUILDER),
str(FIXTURE),
str(output),
str(previews),
str(summary_path),
],
cwd=PROJECT_ROOT,
capture_output=True,
text=True,
timeout=90,
check=False,
)
self.assertEqual(completed.returncode, 0, completed.stderr)
self.assertTrue(output.is_file())
self.assertFalse(Path(f"{output}.inspect.ndjson").exists())
self.assertEqual(output.stat().st_mode & 0o777, 0o600)
summary = json.loads(summary_path.read_text(encoding="utf-8"))
self.assertEqual(summary["status"], "success")
self.assertEqual(summary["row_counts"], [2, 1, 0])
self.assertEqual(summary["formula_count"], 0)
self.assertEqual(summary["preview_count"], 3)
self.assertEqual(len(list(previews.glob("sheet-*.png"))), 3)
def test_invalid_payload_fails_with_only_a_stable_error_code(self):
node = node_binary()
if not node:
self.skipTest("Node.js is unavailable")
with tempfile.TemporaryDirectory(prefix="company-report-xlsx-test-") as temp_dir:
root = Path(temp_dir)
invalid = root / "invalid.json"
invalid.write_text(
json.dumps({"schema_version": "invalid", "synthetic_marker": "SYN-ONLY"}),
encoding="utf-8",
)
completed = subprocess.run(
[
node,
str(BUILDER),
str(invalid),
str(root / "invalid.xlsx"),
str(root / "previews"),
str(root / "summary.json"),
],
cwd=PROJECT_ROOT,
capture_output=True,
text=True,
timeout=30,
check=False,
)
self.assertEqual(completed.returncode, 4)
self.assertEqual(
json.loads(completed.stderr),
{
"status": "failed",
"code": "COMPANY_REPORT_OUTPUT_VALIDATION_FAILED",
},
)
self.assertNotIn("SYN-ONLY", completed.stderr)
self.assertFalse((root / "invalid.xlsx").exists())
if __name__ == "__main__":
unittest.main()

View File

@@ -33,6 +33,9 @@ class DeploymentEntrypointTests(unittest.TestCase):
root_requirements = (PROJECT_ROOT / "requirements.txt").read_text(
encoding="utf-8"
)
monthly_requirements = (
PROJECT_ROOT / "requirements-monthly-reports.txt"
).read_text(encoding="utf-8")
company_requirements = (
PROJECT_ROOT / "requirements-company-reports.txt"
).read_text(encoding="utf-8")
@@ -51,13 +54,11 @@ class DeploymentEntrypointTests(unittest.TestCase):
self.assertNotIn("basic_auth", caddy)
self.assertNotIn("requirements-agent-integration", root_requirements)
self.assertNotIn("requirements-arr-mcp", root_requirements)
self.assertIn("openpyxl", monthly_requirements)
self.assertIn("openpyxl", company_requirements)
company_package = json.loads(
(PROJECT_ROOT / "company_reports" / "xlsx" / "package.json").read_text(
encoding="utf-8"
)
)
self.assertNotIn("@oai/artifact-tool", json.dumps(company_package))
self.assertFalse((PROJECT_ROOT / "company_reports" / "xlsx" / "package.json").exists())
self.assertFalse((PROJECT_ROOT / "monthly_reports" / "xlsx" / "package.json").exists())
self.assertNotIn("artifact-tool", compose)
if __name__ == "__main__":

View File

@@ -0,0 +1,31 @@
from __future__ import annotations
import unittest
from pathlib import Path
PROJECT_ROOT = Path(__file__).resolve().parents[1]
class MonthlyOssArtifactMigrationTests(unittest.TestCase):
def test_forward_migration_allows_oss_and_keeps_local_compatibility(self):
sql = (PROJECT_ROOT / "database" / "016_monthly_report_oss_artifacts.sql").read_text(
encoding="utf-8"
)
self.assertIn("current_database() <> 'booking_test'", sql)
self.assertIn("COALESCE(workbook_provider, '') NOT IN ('oss', 's3', 'local')", sql)
self.assertIn("COALESCE(result_provider, '') NOT IN ('oss', 's3', 'local')", sql)
self.assertIn("CREATE OR REPLACE FUNCTION reporting.validate_monthly_run_publication()", sql)
self.assertIn("existing private OSS", sql)
def test_down_migration_refuses_published_nonlocal_artifacts(self):
sql = (
PROJECT_ROOT / "database" / "016_monthly_report_oss_artifacts.down.sql"
).read_text(encoding="utf-8")
self.assertIn("artifact.storage_provider <> 'local'", sql)
self.assertIn("rollback refused", sql)
self.assertIn("workbook_provider IS DISTINCT FROM 'local'", sql)
if __name__ == "__main__":
unittest.main()

View File

@@ -1,9 +1,5 @@
from __future__ import annotations
import json
import os
import shutil
import subprocess
import tempfile
import unittest
from datetime import date
@@ -19,21 +15,10 @@ from monthly_reports.contracts import (
MonthlySnapshot,
)
from monthly_reports.core import build_monthly_report
from monthly_reports.publishing import OpenpyxlWorkbookBuilder
PROJECT_ROOT = Path(__file__).resolve().parents[1]
BUILDER = PROJECT_ROOT / "monthly_reports" / "xlsx" / "build_workbook.mjs"
def artifact_tool_module() -> str:
return os.environ.get("MONTHLY_REPORT_ARTIFACT_TOOL_MODULE", "").strip()
def node_binary() -> str:
return os.environ.get("MONTHLY_REPORT_NODE_BINARY", "").strip() or shutil.which("node") or ""
def report_payload() -> dict:
def report_fixture():
source = MonthlyFact(
daily_record_id=1,
daily_version_id=100,
@@ -60,7 +45,7 @@ def report_payload() -> dict:
total_price=Decimal("901.00"),
kb_amount=None,
)
report = build_monthly_report(
return build_monthly_report(
2026,
7,
date(2026, 7, 8),
@@ -70,59 +55,28 @@ def report_payload() -> dict:
(ChannelObservation(date(2026, 7, 8), 100, "QBD", 1, 1),),
),
)
return report.to_workbook_payload()
@unittest.skipUnless(
bool(artifact_tool_module() and Path(artifact_tool_module()).is_file()),
"artifact-tool module is not configured",
)
class MonthlyReportsXlsxTests(unittest.TestCase):
def test_builder_exports_reopens_scans_and_renders_every_channel(self):
node = node_binary()
if not node:
self.skipTest("Node.js is unavailable")
def test_openpyxl_builder_reopens_checks_formulas_and_semantic_identity(self):
report = report_fixture()
with tempfile.TemporaryDirectory(prefix="monthly-report-xlsx-") as temp_dir:
root = Path(temp_dir)
payload_path = root / "payload.json"
payload_path.write_text(
json.dumps(report_payload(), ensure_ascii=False),
encoding="utf-8",
)
output = root / "monthly.xlsx"
previews = root / "previews"
summary_path = root / "summary.json"
completed = subprocess.run(
[
node,
str(BUILDER),
str(payload_path),
str(output),
str(previews),
str(summary_path),
],
cwd=PROJECT_ROOT,
capture_output=True,
text=True,
timeout=120,
check=False,
env=os.environ.copy(),
)
built = OpenpyxlWorkbookBuilder().build(report, Path(temp_dir))
self.assertEqual(completed.returncode, 0, completed.stderr)
summary = json.loads(summary_path.read_text(encoding="utf-8"))
self.assertEqual(summary["sheet_names"][:5], [
self.assertTrue(built.path.is_file())
self.assertEqual(built.path.stat().st_mode & 0o777, 0o600)
self.assertEqual(built.summary["sheet_names"][:5], [
"LIANTAI-GROUP", "LIANTAI-FIT", "QBD", "DY-AI-Easy-KB", "FENGRUN"
])
self.assertEqual(summary["row_counts"], [0, 0, 1, 0, 0])
self.assertEqual(summary["formula_count"], 1)
self.assertEqual(summary["preview_count"], 5)
self.assertEqual(len(list(previews.glob("sheet-*.png"))), 5)
self.assertEqual(output.stat().st_mode & 0o777, 0o600)
self.assertFalse(Path(f"{output}.inspect.ndjson").exists())
workbook = load_workbook(output, data_only=False, read_only=True)
self.assertEqual(built.summary["row_counts"], [0, 0, 1, 0, 0])
self.assertEqual(built.summary["formula_count"], 1)
self.assertEqual(built.summary["preview_count"], 5)
self.assertEqual(len(built.summary["semantic_sha256"]), 64)
workbook = load_workbook(built.path, data_only=False, read_only=True)
try:
self.assertEqual(workbook["QBD"]["S2"].value, "=R2*C2*G2")
self.assertEqual(workbook["QBD"]["F2"].value, "'=SYN-FORMULA-LIKE")
formulas = [
cell.value
for sheet in workbook.worksheets
@@ -134,43 +88,6 @@ class MonthlyReportsXlsxTests(unittest.TestCase):
finally:
workbook.close()
def test_invalid_payload_returns_only_stable_error_code(self):
node = node_binary()
if not node:
self.skipTest("Node.js is unavailable")
with tempfile.TemporaryDirectory(prefix="monthly-report-xlsx-") as temp_dir:
root = Path(temp_dir)
invalid = root / "invalid.json"
invalid.write_text(
json.dumps({"schema_version": "bad", "sensitive": "SYN-SECRET"}),
encoding="utf-8",
)
completed = subprocess.run(
[
node,
str(BUILDER),
str(invalid),
str(root / "invalid.xlsx"),
str(root / "previews"),
str(root / "summary.json"),
],
cwd=PROJECT_ROOT,
capture_output=True,
text=True,
timeout=30,
check=False,
env=os.environ.copy(),
)
self.assertEqual(completed.returncode, 4)
self.assertEqual(
json.loads(completed.stderr),
{
"status": "failed",
"code": "MONTHLY_REPORT_OUTPUT_VALIDATION_FAILED",
},
)
self.assertNotIn("SYN-SECRET", completed.stderr)
if __name__ == "__main__":
unittest.main()

View File

@@ -0,0 +1,182 @@
from __future__ import annotations
import hashlib
import tempfile
import unittest
from pathlib import Path
from arr_storage.filesystem import FilesystemObjectBackend
from arr_storage.store import ManagedObjectStore
from arr_web.downloads import (
ArtifactDescriptor,
ControlledProjectArtifactReader,
ManagedObjectArtifactReader,
RoutedArtifactReader,
)
from company_reports.publishing import AtomicReportPublisher as CompanyPublisher
from company_reports.repository import ReservedReport as CompanyReservation
from monthly_reports.publishing import (
AtomicReportPublisher as MonthlyPublisher,
OpenpyxlWorkbookBuilder,
private_staging_directory,
)
from monthly_reports.repository import ReservedReport as MonthlyReservation
from tests.test_company_reports_publishing import (
FakeRepository as CompanyFakeRepository,
empty_report,
synthetic_built_workbook,
)
from tests.test_monthly_reports_publishing import FakeRepository as MonthlyFakeRepository
from tests.test_monthly_reports_xlsx import report_fixture
XLSX_MIME = "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet"
class ReportArtifactStorageTests(unittest.TestCase):
def test_monthly_and_company_publications_upload_workbook_and_result_to_oss(self):
with tempfile.TemporaryDirectory(prefix="report-artifact-oss-") as temporary:
root = Path(temporary) / "project"
root.mkdir()
store = ManagedObjectStore(
FilesystemObjectBackend(root / "objects", create=True)
)
monthly_report = report_fixture()
monthly_work = root / "monthly-work"
monthly_work.mkdir()
monthly_built = OpenpyxlWorkbookBuilder().build(monthly_report, monthly_work)
monthly_outcome = MonthlyPublisher(
root,
root / "outputs" / "monthly_reports",
object_store=store,
).publish(
monthly_report,
MonthlyReservation(44, 2),
monthly_built,
MonthlyFakeRepository(),
monthly_work,
)
self.assertEqual(monthly_outcome.artifact.storage_provider, "oss")
self.assertEqual(monthly_outcome.result_json.storage_provider, "oss")
self.assertEqual(
store.inspect_committed(
monthly_outcome.artifact.storage_key,
monthly_report.filename,
).role,
"monthly_report",
)
self.assertEqual(
store.inspect_committed(
monthly_outcome.result_json.storage_key,
monthly_outcome.result_json.original_filename,
).role,
"result_json",
)
company_report = empty_report()
company_work = root / "company-work"
company_work.mkdir()
company_built = synthetic_built_workbook(
company_work,
company_report.filename,
b"synthetic-company-xlsx",
"a" * 64,
)
company_outcome = CompanyPublisher(
root,
root / "outputs" / "company_reports",
object_store=store,
).publish(
company_report,
CompanyReservation(45, 3, company_report.company),
company_built,
CompanyFakeRepository(),
company_work,
)
self.assertEqual(company_outcome.artifact.storage_provider, "oss")
self.assertEqual(company_outcome.result_json.storage_provider, "oss")
self.assertEqual(
store.inspect_committed(
company_outcome.artifact.storage_key,
company_report.filename,
).role,
"company_report",
)
retry_work = root / "company-retry-work"
retry_work.mkdir()
retry_built = synthetic_built_workbook(
retry_work,
company_report.filename,
b"synthetic-company-xlsx-rebuilt",
"a" * 64,
)
retry_outcome = CompanyPublisher(
root,
root / "outputs" / "company_reports",
object_store=store,
).publish(
company_report,
CompanyReservation(45, 3, company_report.company),
retry_built,
CompanyFakeRepository(),
retry_work,
)
self.assertEqual(retry_outcome.artifact.storage_provider, "oss")
self.assertEqual(
retry_outcome.artifact.sha256,
company_outcome.artifact.sha256,
)
self.assertEqual(retry_outcome.result_json.storage_provider, "oss")
def test_download_router_supports_oss_and_legacy_local_records(self):
with tempfile.TemporaryDirectory(prefix="report-download-") as temporary:
root = Path(temporary)
source = root / "monthly.xlsx"
source.write_bytes(b"monthly-oss-bytes")
store = ManagedObjectStore(
FilesystemObjectBackend(root / "objects", create=True)
)
stored = store.upload_committed(
job_id="monthly-download-001",
attempt_no=1,
role="monthly_report",
source=source,
original_filename="月报.xlsx",
)
oss_descriptor = ArtifactDescriptor(
"monthly_xlsx",
"月报.xlsx",
stored.object_key,
stored.sha256,
stored.byte_size,
stored.mime_type,
"oss",
"arr-private",
)
local_path = root / "outputs" / "legacy.xlsx"
local_path.parent.mkdir()
local_path.write_bytes(b"legacy-local-bytes")
local_descriptor = ArtifactDescriptor(
"monthly_xlsx",
"legacy.xlsx",
"outputs/legacy.xlsx",
hashlib.sha256(local_path.read_bytes()).hexdigest(),
local_path.stat().st_size,
XLSX_MIME,
)
reader = RoutedArtifactReader(
daily_reader=None,
local_reader=ControlledProjectArtifactReader(root),
oss_reader=ManagedObjectArtifactReader(store),
)
self.assertEqual(reader.read(oss_descriptor), source.read_bytes())
self.assertEqual(reader.read(local_descriptor), local_path.read_bytes())
if __name__ == "__main__":
unittest.main()