fix(session): use atomic writes and add corrupt-file repair

SessionManager.save() previously used bare open("w") which could
truncate the JSONL file if the process crashed mid-write. Now writes
to a .tmp file and atomically replaces via os.replace(), matching the
pattern already used in qq.py.

_load() now attempts _repair() before returning None, recovering
valid lines from partially-written files. 12 new tests cover atomic
save correctness, temp-file cleanup on failure, and repair of
truncated/corrupt JSONL.

cowork-with:opencode(glm-5.1)
This commit is contained in:
aiguozhi123456
2026-04-20 00:17:50 +08:00
committed by Xubin Ren
parent 5d976d79ff
commit efb04a1712
2 changed files with 303 additions and 13 deletions
+83 -13
View File
@@ -1,6 +1,7 @@
"""Session management for conversation history.""" """Session management for conversation history."""
import json import json
import os
import shutil import shutil
from dataclasses import dataclass, field from dataclasses import dataclass, field
from datetime import datetime from datetime import datetime
@@ -187,24 +188,93 @@ class SessionManager:
) )
except Exception as e: except Exception as e:
logger.warning("Failed to load session {}: {}", key, e) logger.warning("Failed to load session {}: {}", key, e)
repaired = self._repair(key)
if repaired is not None:
logger.info("Recovered session {} from corrupt file ({} messages)", key, len(repaired.messages))
return repaired
def _repair(self, key: str) -> Session | None:
"""Attempt to recover a session from a corrupt JSONL file."""
path = self._get_session_path(key)
if not path.exists():
return None
try:
messages: list[dict[str, Any]] = []
metadata: dict[str, Any] = {}
created_at: datetime | None = None
updated_at: datetime | None = None
last_consolidated = 0
skipped = 0
with open(path, encoding="utf-8") as f:
for line in f:
line = line.strip()
if not line:
continue
try:
data = json.loads(line)
except json.JSONDecodeError:
skipped += 1
continue
if data.get("_type") == "metadata":
metadata = data.get("metadata", {})
if data.get("created_at"):
try:
created_at = datetime.fromisoformat(data["created_at"])
except (ValueError, TypeError):
pass
if data.get("updated_at"):
try:
updated_at = datetime.fromisoformat(data["updated_at"])
except (ValueError, TypeError):
pass
last_consolidated = data.get("last_consolidated", 0)
else:
messages.append(data)
if skipped:
logger.warning("Skipped {} corrupt lines in session {}", skipped, key)
if not messages and not metadata:
return None
return Session(
key=key,
messages=messages,
created_at=created_at or datetime.now(),
updated_at=updated_at or datetime.now(),
metadata=metadata,
last_consolidated=last_consolidated
)
except Exception as e:
logger.warning("Repair failed for session {}: {}", key, e)
return None return None
def save(self, session: Session) -> None: def save(self, session: Session) -> None:
"""Save a session to disk.""" """Save a session to disk atomically."""
path = self._get_session_path(session.key) path = self._get_session_path(session.key)
tmp_path = path.with_suffix(".jsonl.tmp")
with open(path, "w", encoding="utf-8") as f: try:
metadata_line = { with open(tmp_path, "w", encoding="utf-8") as f:
"_type": "metadata", metadata_line = {
"key": session.key, "_type": "metadata",
"created_at": session.created_at.isoformat(), "key": session.key,
"updated_at": session.updated_at.isoformat(), "created_at": session.created_at.isoformat(),
"metadata": session.metadata, "updated_at": session.updated_at.isoformat(),
"last_consolidated": session.last_consolidated "metadata": session.metadata,
} "last_consolidated": session.last_consolidated
f.write(json.dumps(metadata_line, ensure_ascii=False) + "\n") }
for msg in session.messages: f.write(json.dumps(metadata_line, ensure_ascii=False) + "\n")
f.write(json.dumps(msg, ensure_ascii=False) + "\n") for msg in session.messages:
f.write(json.dumps(msg, ensure_ascii=False) + "\n")
os.replace(tmp_path, path)
except BaseException:
tmp_path.unlink(missing_ok=True)
raise
self._cache[session.key] = session self._cache[session.key] = session
+220
View File
@@ -0,0 +1,220 @@
"""Tests for atomic session save and corrupt-file repair."""
import json
from datetime import datetime
from pathlib import Path
from nanobot.session.manager import Session, SessionManager
class TestAtomicSave:
def test_save_creates_valid_jsonl(self, tmp_path: Path):
mgr = SessionManager(tmp_path)
session = Session(key="test:1")
session.add_message("user", "hello")
session.add_message("assistant", "hi")
mgr.save(session)
path = mgr._get_session_path("test:1")
lines = path.read_text(encoding="utf-8").strip().split("\n")
assert len(lines) == 3
meta = json.loads(lines[0])
assert meta["_type"] == "metadata"
assert meta["key"] == "test:1"
msg1 = json.loads(lines[1])
assert msg1["role"] == "user"
assert msg1["content"] == "hello"
def test_no_tmp_file_left_after_successful_save(self, tmp_path: Path):
mgr = SessionManager(tmp_path)
session = Session(key="test:clean")
mgr.save(session)
tmp_files = list(mgr.sessions_dir.glob("*.tmp"))
assert tmp_files == []
def test_tmp_file_cleaned_up_on_write_failure(self, tmp_path: Path):
mgr = SessionManager(tmp_path)
session = Session(key="test:fail")
path = mgr._get_session_path("test:fail")
tmp_path_file = path.with_suffix(".jsonl.tmp")
path.parent.mkdir(parents=True, exist_ok=True)
tmp_path_file.write_text("stale")
class BadMessage:
def __init__(self, data):
self.data = data
original_dumps = json.dumps
def failing_dumps(obj, **kwargs):
if isinstance(obj, dict) and obj.get("role") == "assistant":
raise OSError("simulated disk full")
return original_dumps(obj, **kwargs)
session = Session(key="test:fail")
session.messages = [
{"role": "user", "content": "ok"},
{"role": "assistant", "content": "will fail"},
]
import unittest.mock
with unittest.mock.patch("nanobot.session.manager.json.dumps", side_effect=failing_dumps):
try:
mgr.save(session)
except OSError:
pass
assert not tmp_path_file.exists()
def test_overwrite_preserves_latest_data(self, tmp_path: Path):
mgr = SessionManager(tmp_path)
session = Session(key="test:overwrite")
session.add_message("user", "first")
mgr.save(session)
session.add_message("user", "second")
mgr.save(session)
mgr.invalidate("test:overwrite")
loaded = mgr.get_or_create("test:overwrite")
assert len(loaded.messages) == 2
assert loaded.messages[0]["content"] == "first"
assert loaded.messages[1]["content"] == "second"
def test_consecutive_saves_are_consistent(self, tmp_path: Path):
mgr = SessionManager(tmp_path)
session = Session(key="test:consistency")
for i in range(5):
session.add_message("user", f"msg{i}")
mgr.save(session)
mgr.invalidate("test:consistency")
loaded = mgr.get_or_create("test:consistency")
assert len(loaded.messages) == 5
for i in range(5):
assert loaded.messages[i]["content"] == f"msg{i}"
class TestRepairCorruptFile:
def _write_corrupt_jsonl(self, path: Path, lines: list[str]) -> None:
path.parent.mkdir(parents=True, exist_ok=True)
path.write_text("\n".join(lines) + "\n", encoding="utf-8")
def test_truncated_last_line_recovered(self, tmp_path: Path):
mgr = SessionManager(tmp_path)
path = mgr._get_session_path("test:trunc")
valid_meta = json.dumps({
"_type": "metadata",
"key": "test:trunc",
"created_at": datetime.now().isoformat(),
"updated_at": datetime.now().isoformat(),
"metadata": {},
"last_consolidated": 0,
})
valid_msg = json.dumps({"role": "user", "content": "hello"})
self._write_corrupt_jsonl(path, [
valid_meta,
valid_msg,
'{"role": "assistant", "content": "partial...',
])
session = mgr._load("test:trunc")
assert session is not None
assert len(session.messages) == 1
assert session.messages[0]["content"] == "hello"
def test_corrupt_metadata_line_skipped(self, tmp_path: Path):
mgr = SessionManager(tmp_path)
path = mgr._get_session_path("test:badmeta")
self._write_corrupt_jsonl(path, [
"NOT VALID JSON!!!",
'{"role": "user", "content": "survived"}',
])
session = mgr._load("test:badmeta")
assert session is not None
assert len(session.messages) == 1
assert session.messages[0]["content"] == "survived"
def test_all_corrupt_lines_returns_none(self, tmp_path: Path):
mgr = SessionManager(tmp_path)
path = mgr._get_session_path("test:allbad")
self._write_corrupt_jsonl(path, [
"garbage line 1",
"garbage line 2",
"{{invalid json",
])
session = mgr._load("test:allbad")
assert session is None
def test_empty_file_returns_empty_session(self, tmp_path: Path):
mgr = SessionManager(tmp_path)
path = mgr._get_session_path("test:empty")
path.parent.mkdir(parents=True, exist_ok=True)
path.write_text("", encoding="utf-8")
session = mgr._load("test:empty")
assert session is not None
assert session.messages == []
assert session.key == "test:empty"
def test_repair_preserves_valid_messages_amid_corruption(self, tmp_path: Path):
mgr = SessionManager(tmp_path)
path = mgr._get_session_path("test:mixed")
self._write_corrupt_jsonl(path, [
json.dumps({"_type": "metadata", "key": "test:mixed",
"created_at": datetime.now().isoformat(),
"updated_at": datetime.now().isoformat(),
"metadata": {}, "last_consolidated": 0}),
"BROKEN",
json.dumps({"role": "user", "content": "msg1"}),
'{"role": "assistant", "content": "broken',
json.dumps({"role": "user", "content": "msg2"}),
])
session = mgr._load("test:mixed")
assert session is not None
assert len(session.messages) == 2
assert session.messages[0]["content"] == "msg1"
assert session.messages[1]["content"] == "msg2"
def test_repair_with_bad_timestamp_uses_fallback(self, tmp_path: Path):
mgr = SessionManager(tmp_path)
path = mgr._get_session_path("test:badts")
self._write_corrupt_jsonl(path, [
json.dumps({"_type": "metadata", "key": "test:badts",
"created_at": "not-a-date",
"updated_at": "also-bad",
"metadata": {}, "last_consolidated": 5}),
json.dumps({"role": "user", "content": "hi"}),
])
session = mgr._load("test:badts")
assert session is not None
assert session.last_consolidated == 5
assert isinstance(session.created_at, datetime)
def test_get_or_create_returns_new_session_for_corrupt_file(self, tmp_path: Path):
mgr = SessionManager(tmp_path)
path = mgr._get_session_path("test:fallback")
self._write_corrupt_jsonl(path, ["{{{{"])
session = mgr.get_or_create("test:fallback")
assert session is not None
assert session.messages == []
assert session.key == "test:fallback"