NeteaseUrl/core/db.py

# -*- coding: utf-8 -*-
"""异步 SQLite 访问层（aiosqlite + WAL）。

核心目标：**读写同时进行**。
- WAL 模式：读事务与写事务互不阻塞（读快照隔离），这是 SQLite 层并发的根本；
- 读路径：每次操作独立连接、无锁、可并发（`fetch_all`/`fetch_one`/`fetch_val`）；
- 写路径：SQLite 全库同一时刻只允许一个写者，用 `asyncio.Lock` 串行化写事务，
  但写者不阻塞读者（WAL）；
- `busy_timeout`：多进程/多连接下遇到短暂写锁自旋等待而非立刻 SQLITE_BUSY。

与旧实现的差异：旧 `auth_system` 用 `threading.RLock` 把**全部**读写串行化，
读也被写挡住；本层只串行化写，读放行并发，即「读写同时进行」。
"""

import asyncio
import os
from collections.abc import AsyncGenerator, Iterable, Sequence
from contextlib import asynccontextmanager
from pathlib import Path
from typing import Any

import aiosqlite

# 站点数据目录（可经环境变量覆盖，便于测试隔离）。
# 用 `or 'data'` 而不是 get 的默认值：get 的默认只在变量**缺失**时生效，
# 而 `SITE_DATA_DIR=`（空串）会让 Path('') 退化成当前工作目录，
# 数据库静默落到 CWD 而不是 data/。
# 注意：auth_system.db.DATA_DIR 必须与此处保持同一条表达式（两者指向同一个库）
DATA_DIR = Path(os.environ.get('SITE_DATA_DIR') or 'data')
DB_PATH = DATA_DIR / 'site.db'

# 锁按事件循环隔离：asyncio.Lock 绑定创建它的 loop，跨 loop 复用会抛
# "bound to a different event loop"（应用主 loop + 同步桥 loop + 测试各自
# 的 loop 并存，必须分开持有）。
_LOOP_LOCKS: dict[tuple[int, str], asyncio.Lock] = {}


def _loop_lock(kind: str) -> asyncio.Lock:
    key = (id(asyncio.get_running_loop()), kind)
    lock = _LOOP_LOCKS.get(key)
    if lock is None:
        lock = asyncio.Lock()
        _LOOP_LOCKS[key] = lock
    return lock


def _write_lock() -> asyncio.Lock:
    """写事务串行锁（每事件循环一个）。"""
    return _loop_lock('write')


def _wal_lock() -> asyncio.Lock:
    """WAL 切换专用锁（不能用 _write_lock：execute() 已持它再调
    connect→_configure 会自锁死）。"""
    return _loop_lock('wal')


# 已确认切到 WAL 的库路径（首连设置一次；后续连接不再重复执行 PRAGMA 争排他锁）
_WAL_READY_FOR: str | None = None


async def _configure(conn: aiosqlite.Connection) -> None:
    global _WAL_READY_FOR
    conn.row_factory = aiosqlite.Row
    # busy_timeout 先设：让后续语句遇锁自旋等待，而不是立刻报 database is locked
    await conn.execute('PRAGMA busy_timeout=15000')
    key = str(DB_PATH)
    if _WAL_READY_FOR != key:
        # 切换日志模式需要排他锁，且 `_WAL_READY_FOR` 的检查不是原子的——
        # 并发连接须在专用锁内串行完成「切换 + 标记」，并二次复查，否则批量并发
        # 连接会同时执行 PRAGMA journal_mode=WAL 争锁（实测 database is locked）
        async with _wal_lock():
            if _WAL_READY_FOR != key:
                await conn.execute('PRAGMA journal_mode=WAL')
                _WAL_READY_FOR = key
    await conn.execute('PRAGMA synchronous=NORMAL')  # WAL 下兼顾安全与吞吐
    await conn.execute('PRAGMA foreign_keys=ON')


async def connect() -> aiosqlite.Connection:
    """新建一个已配置好的连接（调用方负责 close）。"""
    DATA_DIR.mkdir(parents=True, exist_ok=True)  # noqa: ASYNC240 - 幂等的微秒级一次目录创建，无需线程池
    conn = await aiosqlite.connect(str(DB_PATH), timeout=15)
    await _configure(conn)
    return conn


# ============ 读路径（并发，无锁） ============


async def fetch_all(sql: str, params: Sequence = ()) -> list[aiosqlite.Row]:
    conn = await connect()
    try:
        cur = await conn.execute(sql, params)
        return list(await cur.fetchall())
    finally:
        await conn.close()


async def fetch_one(sql: str, params: Sequence = ()) -> aiosqlite.Row | None:
    conn = await connect()
    try:
        cur = await conn.execute(sql, params)
        return await cur.fetchone()
    finally:
        await conn.close()


async def fetch_val(sql: str, params: Sequence = (), default: Any = None) -> Any:
    row = await fetch_one(sql, params)
    return row[0] if row is not None else default


# ============ 写路径（串行，不阻塞读） ============


async def execute(sql: str, params: Sequence = ()) -> int:
    """执行单条写语句，返回 lastrowid（INSERT）或 rowcount（UPDATE/DELETE）。"""
    async with _write_lock():
        conn = await connect()
        try:
            cur = await conn.execute(sql, params)
            await conn.commit()
            return cur.lastrowid if cur.lastrowid else cur.rowcount
        finally:
            await conn.close()


async def execute_many(sql: str, seq: Iterable[Sequence]) -> int:
    """批量写，返回受影响行数。"""
    async with _write_lock():
        conn = await connect()
        try:
            cur = await conn.executemany(sql, seq)
            await conn.commit()
            return cur.rowcount
        finally:
            await conn.close()


@asynccontextmanager
async def write_transaction() -> AsyncGenerator[aiosqlite.Connection, None]:
    """写事务上下文：多条写语句原子提交（串行锁内）。

    用法：
        async with write_transaction() as conn:
            await conn.execute(...)
            await conn.execute(...)
        # 退出即 commit；异常则 rollback
    """
    async with _write_lock():
        conn = await connect()
        try:
            yield conn
            await conn.commit()
        except Exception:
            await conn.rollback()
            raise
        finally:
            await conn.close()


@asynccontextmanager
async def read_connection() -> AsyncGenerator[aiosqlite.Connection, None]:
    """只读连接上下文（无锁、可并发）；需要在一个连接上做多次读时使用。"""
    conn = await connect()
    try:
        yield conn
    finally:
        await conn.close()