feat(pending): app_name field and date-sharded pending storage

- Add optional app_name to PendingSubmission and prepare_submit_job,
  passed through PrepareSubmitJobRequest for user-provided tracking.
- Rewrite PendingStore to shard records by UTC date under
  data/pending_jobs/YYYY-MM-DD.json instead of a single monolithic
  pending_jobs.json.
- Legacy single-file pending_jobs.json remains readable; first save()
  migrates it and renames the old file to avoid double reads.
- Tests cover date sharding, multi-date list, legacy read, legacy
  migration, and app_name round-trip.
This commit is contained in:
Claude
2026-06-26 14:36:24 +08:00
parent c44e9bcc55
commit da9ba657f3
7 changed files with 163 additions and 35 deletions
+86 -31
View File
@@ -6,6 +6,7 @@
import json
import os
import tempfile
from datetime import datetime
from pathlib import Path
from threading import Lock
@@ -14,62 +15,104 @@ from common.logging import logger
from spark_executor.models import PendingSubmission
DEFAULT_DATA_DIR = settings.data_dir
DEFAULT_FILE_NAME = "pending_jobs.json"
DEFAULT_DIR_NAME = "pending_jobs"
class PendingStore:
"""JSON-backed CRUD for PendingSubmission records.
"""JSON-backed CRUD for PendingSubmission records, sharded by UTC date.
File path: <DEFAULT_DATA_DIR>/<DEFAULT_FILE_NAME>.
Atomic writes (tempfile + os.replace), same as ConnectionStore.
Records live in <data_dir>/pending_jobs/YYYY-MM-DD.json, one file per
calendar date taken from PendingSubmission.created_at. The legacy single-file
pending_jobs.json is still readable and is migrated non-destructively on
the first save().
"""
def __init__(self, data_dir: str | None = None, file_name: str = DEFAULT_FILE_NAME) -> None:
def __init__(self, data_dir: str | None = None, dir_name: str = DEFAULT_DIR_NAME) -> None:
self._data_dir = data_dir or DEFAULT_DATA_DIR
self._file_name = file_name
self._dir_name = dir_name
self._lock = Lock()
@property
def path(self) -> Path:
return Path(self._data_dir) / self._file_name
def dir_path(self) -> Path:
return Path(self._data_dir) / self._dir_name
def _load(self) -> dict[str, PendingSubmission]:
if not self.path.exists():
logger.debug(f"PendingStore._load: file {self.path} absent, returning empty")
def _date_path(self, created_at: datetime) -> Path:
# created_at is datetime.utcnow(), so the shard date is a UTC date.
return self.dir_path / f"{created_at.date().isoformat()}.json"
def _legacy_path(self) -> Path:
return Path(self._data_dir) / "pending_jobs.json"
def _list_date_files(self) -> list[Path]:
if not self.dir_path.exists():
return []
return sorted(self.dir_path.glob("*.json"))
def _load_file(self, path: Path) -> dict[str, PendingSubmission]:
if not path.exists():
return {}
raw = json.loads(self.path.read_text(encoding="utf-8"))
logger.debug(f"PendingStore._load: loaded {len(raw)} records from {self.path}")
raw = json.loads(path.read_text(encoding="utf-8"))
return {pid: PendingSubmission.model_validate(p) for pid, p in raw.items()}
def _dump(self, records: dict[str, PendingSubmission]) -> None:
os.makedirs(self._data_dir, exist_ok=True)
def _dump_file(self, path: Path, records: dict[str, PendingSubmission]) -> None:
os.makedirs(path.parent, exist_ok=True)
payload = {pid: p.model_dump() for pid, p in records.items()}
fd, tmp_path = tempfile.mkstemp(
prefix=self._file_name + ".", dir=self._data_dir
)
fd, tmp_path = tempfile.mkstemp(prefix=path.name + ".", dir=path.parent)
try:
with os.fdopen(fd, "w", encoding="utf-8") as f:
json.dump(payload, f, indent=2, ensure_ascii=False, default=str)
os.replace(tmp_path, self.path)
logger.debug(f"PendingStore._dump: wrote {len(records)} records to {self.path}")
os.replace(tmp_path, path)
except Exception:
if os.path.exists(tmp_path):
os.unlink(tmp_path)
raise
def _load_all(self) -> dict[str, PendingSubmission]:
records: dict[str, PendingSubmission] = {}
for path in self._list_date_files():
records.update(self._load_file(path))
legacy = self._legacy_path()
if legacy.exists():
records.update(self._load_file(legacy))
return records
def list_all(self) -> list[PendingSubmission]:
with self._lock:
return list(self._load().values())
return list(self._load_all().values())
def get(self, pending_id: str) -> PendingSubmission | None:
with self._lock:
return self._load().get(pending_id)
# Newest first is the cheapest common case.
for path in reversed(self._list_date_files()):
record = self._load_file(path).get(pending_id)
if record is not None:
return record
legacy = self._legacy_path()
if legacy.exists():
return self._load_file(legacy).get(pending_id)
return None
def _migrate_legacy(self) -> None:
"""Move records out of the monolithic legacy file into date shards."""
legacy = self._legacy_path()
if not legacy.exists():
return
for p in self._load_file(legacy).values():
path = self._date_path(p.created_at)
records = self._load_file(path)
records[p.pending_id] = p
self._dump_file(path, records)
# Preserve the old file but stop reading it as a data file.
os.makedirs(self.dir_path, exist_ok=True)
legacy.rename(self.dir_path / "pending_jobs.legacy")
def save(self, pending: PendingSubmission) -> None:
with self._lock:
records = self._load()
self._migrate_legacy()
path = self._date_path(pending.created_at)
records = self._load_file(path)
records[pending.pending_id] = pending
self._dump(records)
self._dump_file(path, records)
logger.info(
f"pending saved pending_id={pending.pending_id} "
f"status={pending.status} connection={pending.connection}"
@@ -77,13 +120,25 @@ class PendingStore:
def delete(self, pending_id: str) -> bool:
with self._lock:
records = self._load()
if pending_id not in records:
return False
del records[pending_id]
self._dump(records)
logger.info(f"pending deleted pending_id={pending_id}")
return True
for path in self._list_date_files():
records = self._load_file(path)
if pending_id in records:
del records[pending_id]
if records:
self._dump_file(path, records)
else:
path.unlink()
logger.info(f"pending deleted pending_id={pending_id}")
return True
legacy = self._legacy_path()
if legacy.exists():
records = self._load_file(legacy)
if pending_id in records:
del records[pending_id]
self._dump_file(legacy, records)
logger.info(f"pending deleted pending_id={pending_id}")
return True
return False
# Module-level singleton; replaced in tests.