feat(db): SQLite 全量迁移至 MySQL(config.yaml 配置化 + 迁移脚本 + 一致性校验)

- config.yaml database.mysql:host/port/db/user/charset 明文可提交;密码经 password_env
  引用 .env 的 MYSQL_PASSWORD(AGENT.md §33 密钥不进 git)
- config.py _build_mysql_url:URL 优先级 DATABASE_URL env > database.mysql 段 > sqlite 兜底
- pyproject 引入 pymysql>=1.1
- tests/conftest.py 强制每进程 /tmp SQLite(测试绝不触 MySQL 开发库);test_config 覆盖
  mysql 组装/密码可选/sqlite 兜底分支
- scripts/migrate_sqlite_to_mysql.py:sqlite→mysql 一次性迁移工具(keyset 分页 + chunk
  多值 INSERT + 攒批 commit + 幂等续传 + 低配 MySQL 节流 --throttle-sec + --verify-only
  行数与抽样一致性校验);已用于 data/quant.db 约 1605 万行迁移并经校验一致
This commit is contained in:
Simon
2026-09-08 23:58:48 +08:00
parent a3fabf9ae9
commit 6c2f198261
8 changed files with 409 additions and 20 deletions
+25 -1
View File
@@ -10,6 +10,7 @@ import os
from dataclasses import dataclass
from functools import lru_cache
from pathlib import Path
from urllib.parse import quote_plus
import yaml
@@ -99,6 +100,26 @@ def _normalize_sqlite_url(url: str) -> str:
return f"{_SQLITE_PREFIX}{(PROJECT_ROOT / rest).resolve()}"
def _build_mysql_url(mysql: dict | None) -> str | None:
"""由 config.yaml database.mysql 段组装 mysql+pymysql URL。
规则:仅当 enabled 且 host/user/db 齐全时返回 URL;密码经 password_env
指定的环境变量读取(AGENT.md §33 密钥只放 .env),未设置则按空密码处理。
"""
if not mysql or not mysql.get("enabled"):
return None
host = mysql.get("host")
user = mysql.get("user")
db = mysql.get("db")
if not (host and user and db):
return None
port = mysql.get("port") or 3306
charset = mysql.get("charset") or "utf8mb4"
password = os.environ.get(mysql.get("password_env") or "MYSQL_PASSWORD", "")
auth = f"{quote_plus(user)}:{quote_plus(password)}" if password else quote_plus(user)
return f"mysql+pymysql://{auth}@{host}:{port}/{db}?charset={charset}"
@lru_cache
def get_settings() -> Settings:
"""加载并缓存 Settings。config / env 路径可通过参数覆盖以便测试。"""
@@ -115,7 +136,10 @@ def get_settings() -> Settings:
llm_model_env = agent_llm.get("model_env") or "LLM_MODEL"
default_url = "sqlite:///./data/quant.db"
database_url = os.environ.get(url_env) or default_url
# URL 优先级:环境变量(DATABASE_URL) > config.yaml database.mysql 段 > SQLite 兜底
database_url = os.environ.get(url_env) or _build_mysql_url(
_deep(cfg, "database.mysql") or {}
) or default_url
migrations_rel = _deep(cfg, "database.migrations_dir") or (
"app/infrastructure/persistence/migrations"
+1
View File
@@ -14,6 +14,7 @@ dependencies = [
# Qlib 研究引擎:PyPI 无 aarch64 wheel(见 docs/ROADMAP.md §2 备注),故从源码 git 安装并固定 commit。
# 网络下载困难时使用 HTTP 代理(见根目录 AGENT.md §0:192.168.1.160:3128)。
"pyqlib @ git+https://github.com/microsoft/qlib.git@79633dd",
"pymysql>=1.1",
]
[project.optional-dependencies]
+7 -1
View File
@@ -1,8 +1,10 @@
"""全局测试配置。
- 强制 SQLite 临时文件数据库:默认库已是 MySQL(config.yaml database.mysql),
测试绝不允许触碰开发 MySQL(qlib@192.168.1.10);该 env 在 app 模块
首次导入前生效,SessionLocal/engine 会按它构造。
- JOB_MODE=local:单测/CI 里 Job 在本进程执行(不 spawn 子进程、不依赖安装态)
- QLIB_SKIP_STALE_JOB_CLEANUP=1:TestClient 启动 lifespan 不清理残留 Job
(避免连接/写入开发库 data/quant.db)
必须在任何 app / config 模块首次导入前生效,故放在 conftest 模块顶层。
"""
@@ -11,6 +13,10 @@ from __future__ import annotations
import os
# 测试数据库:每进程唯一 /tmp sqlite 文件(隔离、不落项目目录、不触 MySQL)
_TEST_DB = f"/tmp/qlib-pytest-{os.getpid()}.db"
os.environ["DATABASE_URL"] = f"sqlite:///{_TEST_DB}"
# 强制(不是 setdefault):测试绝不 spawn 研究子进程 / 不触碰开发库
os.environ["JOB_MODE"] = "local"
os.environ["QLIB_SKIP_STALE_JOB_CLEANUP"] = "1"
+63 -14
View File
@@ -1,9 +1,10 @@
"""配置层测试:默认 SQLite 相对路径解析、config.yaml / .env 加载约定。"""
"""配置层测试:默认 MySQL URL 组装(config.yaml database.mysql)、.env 加载约定、
SQLite 兜底逻辑。conftest 已强制 DATABASE_URL=tmp sqlite,本文件内用
monkeypatch/cache_clear 单独验证「无 env 时走 mysql 段」的分支(只组装不连接)。"""
from __future__ import annotations
import os
from pathlib import Path
from app.core.config import PROJECT_ROOT, _load_env_file, get_settings
@@ -14,14 +15,62 @@ def test_project_root_points_to_repo_root() -> None:
assert (PROJECT_ROOT / "backend").is_dir()
def test_default_database_url_resolves_to_project_data_dir() -> None:
def test_test_runner_isolation_uses_tmp_sqlite() -> None:
"""conftest 强制每进程唯一 /tmp sqlite,测试绝不触碰 MySQL 开发库。"""
settings = get_settings()
assert settings.database_url.startswith("sqlite:///")
# 相对路径应解析到 <项目根>/data/quant.db
assert settings.database_url.endswith("/data/quant.db")
db_path = Path(settings.database_url.removeprefix("sqlite:///"))
assert db_path.is_absolute()
assert db_path == PROJECT_ROOT / "data" / "quant.db"
assert settings.database_url.startswith("sqlite:////tmp/qlib-pytest-")
def test_mysql_default_from_config_yaml(monkeypatch) -> None:
"""未设 DATABASE_URL 时,config.yaml database.mysql 段组装 MySQL URL(不连接)。"""
monkeypatch.delenv("DATABASE_URL", raising=False)
get_settings.cache_clear()
try:
url = get_settings().database_url
finally:
get_settings.cache_clear()
assert url.startswith("mysql+pymysql://qlib:")
assert "192.168.1.10:3306/qlib?charset=utf8mb4" in url
def test_build_mysql_url(monkeypatch) -> None:
from app.core.config import _build_mysql_url
monkeypatch.setenv("MYSQL_PASSWORD", "p@ss:word")
cfg = {
"enabled": True,
"host": "10.0.0.2",
"port": 3307,
"db": "q",
"user": "u",
"password_env": "MYSQL_PASSWORD",
}
url = _build_mysql_url(cfg)
assert url == "mysql+pymysql://u:p%40ss%3Aword@10.0.0.2:3307/q?charset=utf8mb4"
# 未启用 / 缺字段 → None(回退 sqlite 兜底)
assert _build_mysql_url({**cfg, "enabled": False}) is None
assert _build_mysql_url({**cfg, "host": None}) is None
assert _build_mysql_url(None) is None
def test_mysql_url_password_optional(monkeypatch) -> None:
"""密码留空(未设 env)也可组装 URL —— 方便仅内网/免密场景。"""
from app.core.config import _build_mysql_url
monkeypatch.delenv("MYSQL_PASSWORD", raising=False)
cfg = {"enabled": True, "host": "h", "db": "d", "user": "u"}
assert _build_mysql_url(cfg) == "mysql+pymysql://u@h:3306/d?charset=utf8mb4"
def test_sqlite_url_normalization_and_fallback() -> None:
"""仅 sqlite 相对路径被解析为项目根绝对路径;其它 URL 原样透传。"""
from app.core.config import _normalize_sqlite_url
assert _normalize_sqlite_url("mysql+pymysql://u:p@h/d") == "mysql+pymysql://u:p@h/d"
abs_url = _normalize_sqlite_url("sqlite:///./data/quant.db")
assert abs_url.startswith("sqlite:///")
# 绝对路径 sqlite 不再重复解析
assert _normalize_sqlite_url("sqlite:////abs/x.db") == "sqlite:////abs/x.db"
def test_settings_loaded_from_config_yaml() -> None:
@@ -32,7 +81,7 @@ def test_settings_loaded_from_config_yaml() -> None:
assert settings.data_source_fallback == "sina"
def test_env_file_loading(tmp_path: Path, monkeypatch) -> None:
def test_env_file_loading(tmp_path, monkeypatch) -> None:
monkeypatch.delenv("TUSHARE_TOKEN", raising=False)
env_file = tmp_path / ".env"
env_file.write_text(
@@ -57,14 +106,14 @@ def test_llm_from_config_yaml() -> None:
def test_llm_env_overrides_yaml(monkeypatch) -> None:
from app.core.config import get_settings
monkeypatch.setenv("LLM_MODEL", "env-model")
monkeypatch.setenv("LLM_BASE_URL", "https://example.com/v1")
monkeypatch.setenv("LLM_API_KEY", "sk-test")
get_settings.cache_clear() # Settings 有 lru_cache,刷新以读取新环境变量
s = get_settings()
try:
s = get_settings()
finally:
get_settings.cache_clear()
assert s.llm_model == "env-model"
assert s.llm_base_url == "https://example.com/v1"
assert s.llm_api_key == "sk-test"
get_settings.cache_clear()
+11
View File
@@ -3262,6 +3262,15 @@ wheels = [
{ url = "https://files.pythonhosted.org/packages/64/02/b2606ddf52fa615d4ff3edb2aa05fb064337fa871f99ebb184a39e051cc8/pymongo-4.18.0-cp314-cp314t-win_arm64.whl", hash = "sha256:4a81d166a43e8af1e5152b6854a263ba0a8831f7dd2ca1badc716f219f4f1bc0", size = 817605, upload-time = "2026-09-03T16:00:41.547Z" },
]
[[package]]
name = "pymysql"
version = "1.2.0"
source = { registry = "https://pypi.org/simple" }
sdist = { url = "https://files.pythonhosted.org/packages/c9/bc/1c6a92f385940f727daeecf3bacaf186e03875dff57197801046c583bcf0/pymysql-1.2.0.tar.gz", hash = "sha256:6c7b17ca686988104d7426c27895b455cdeea3e9d3ceb1270f0c3704fead8c33", size = 49021, upload-time = "2026-05-19T08:26:22.302Z" }
wheels = [
{ url = "https://files.pythonhosted.org/packages/c4/bd/2534e130295c8cfd4f0a2e31623baab7502278f1e97bcfe61db75656a77f/pymysql-1.2.0-py3-none-any.whl", hash = "sha256:62169ce6d5510f08e140c5e7990ee884a9764024e4a9a27b2cc11f1099322ae0", size = 45716, upload-time = "2026-05-19T08:26:20.974Z" },
]
[[package]]
name = "pyparsing"
version = "3.3.2"
@@ -3540,6 +3549,7 @@ dependencies = [
{ name = "alembic" },
{ name = "fastapi" },
{ name = "httpx" },
{ name = "pymysql" },
{ name = "pyqlib" },
{ name = "pyyaml" },
{ name = "sqlalchemy" },
@@ -3563,6 +3573,7 @@ requires-dist = [
{ name = "alembic", specifier = ">=1.13" },
{ name = "fastapi", specifier = ">=0.115" },
{ name = "httpx", specifier = ">=0.28.1" },
{ name = "pymysql", specifier = ">=1.1" },
{ name = "pyqlib", git = "https://github.com/microsoft/qlib.git?rev=79633dd" },
{ name = "pyyaml", specifier = ">=6.0" },
{ name = "sqlalchemy", specifier = ">=2.0" },