cloudoll 4.4.0__tar.gz → 4.5.0__tar.gz

{cloudoll-4.4.0 → cloudoll-4.5.0}/CHANGELOG.md

@@ -1,5 +1,11 @@
 # Changelog
 
+## 4.5.0
+
+- Add native HTTPS support to `Application.run()` using `server.ssl_certfile` and `server.ssl_keyfile`, or a supplied `server.ssl_context`.
+- Add an async SQLite ORM engine with bounded pooling, transactions (including `IMMEDIATE` mode), streaming, URL parsing, SQLite SQL dialect, schema generation and model introspection.
+- Add SQLite upsert support and include `aiosqlite` as a runtime dependency.
+
 简体中文见 [CHANGELOG.zh-CN.md](CHANGELOG.zh-CN.md)。
 
 ## 4.4.0

{cloudoll-4.4.0 → cloudoll-4.5.0}/PKG-INFO

@@ -1,7 +1,7 @@
 Metadata-Version: 2.4
 Name: cloudoll
-Version: 4.4.0
-Summary: Assist in quickly creating distributable microservices based on aiohttp, with a focus on simplicity and ease of use.
+Version: 4.5.0
+Summary: A distributable microservice framework with async ORM support for MySQL, PostgreSQL and SQLite.
 Author-email: Qiu <smallerqiu@gmail.com>
 License-Expression: MIT
 Project-URL: Homepage, https://github.com/smallerqiu/cloudoll-py
@@ -28,6 +28,7 @@ Requires-Dist: psutil
 Requires-Dist: tabulate
 Requires-Dist: concurrent_log_handler
 Requires-Dist: portalocker
+Requires-Dist: aiosqlite<1,>=0.20
 Provides-Extra: cache
 Requires-Dist: redis[hiredis]; extra == "cache"
 Requires-Dist: aiomcache; extra == "cache"
@@ -54,8 +55,10 @@ Dynamic: license-file
 
 English | [简体中文](README.zh-CN.md)
 
-Cloudoll 4.4.0 is an aiohttp-based Python library for routing, middleware,
+Cloudoll 4.5.0 is an aiohttp-based Python library for routing, middleware,
 templates, sessions, JWT, database access and application scaffolding.
+The ORM supports SQLite out of the box and MySQL/PostgreSQL through optional drivers;
+`Application.run()` can serve HTTPS directly from PEM certificates.
 
 ## AI-assisted development
 
@@ -71,8 +74,8 @@ does not have `cloudoll ai --help`, consult its matching README and source first
 Python 3.9+ is required. Install only the optional drivers you use:
 
 ```sh
-python -m pip install 'cloudoll==4.4.0'
-python -m pip install 'cloudoll[mysql,postgres,cache]==4.4.0'
+python -m pip install 'cloudoll==4.5.0'
+python -m pip install 'cloudoll[mysql,postgres,cache]==4.5.0'
 ```
 
 The AWS extra requires Python 3.10+. Aurora remains untested; see
@@ -87,7 +90,7 @@ python -m pip install -r requirements.txt
 cloudoll start -n myapp
 ```
 
-The generated project targets 4.4.0 and receives random JWT and session secrets
+The generated project targets 4.5.0 and receives random JWT and session secrets
 in `.env`. Controllers
 and authentication middleware are examples: replace demo credentials and supply
 your own authorization and upload policies before deployment.
@@ -173,8 +176,8 @@ python -m ruff check cloudoll tests
 python -m mypy
 python -m pytest -q -m 'not integration' -k 'not aws'
 python -m build
-python tests/check_wheel.py dist/cloudoll-4.4.0-py3-none-any.whl
-python tests/check_release.py dist/cloudoll-4.4.0-py3-none-any.whl
+python tests/check_wheel.py dist/cloudoll-4.5.0-py3-none-any.whl
+python tests/check_release.py dist/cloudoll-4.5.0-py3-none-any.whl
 ```
 
 Native integration tests require explicitly configured disposable databases.

{cloudoll-4.4.0 → cloudoll-4.5.0}/README.md

@@ -2,8 +2,10 @@
 
 English | [简体中文](README.zh-CN.md)
 
-Cloudoll 4.4.0 is an aiohttp-based Python library for routing, middleware,
+Cloudoll 4.5.0 is an aiohttp-based Python library for routing, middleware,
 templates, sessions, JWT, database access and application scaffolding.
+The ORM supports SQLite out of the box and MySQL/PostgreSQL through optional drivers;
+`Application.run()` can serve HTTPS directly from PEM certificates.
 
 ## AI-assisted development
 
@@ -19,8 +21,8 @@ does not have `cloudoll ai --help`, consult its matching README and source first
 Python 3.9+ is required. Install only the optional drivers you use:
 
 ```sh
-python -m pip install 'cloudoll==4.4.0'
-python -m pip install 'cloudoll[mysql,postgres,cache]==4.4.0'
+python -m pip install 'cloudoll==4.5.0'
+python -m pip install 'cloudoll[mysql,postgres,cache]==4.5.0'
 ```
 
 The AWS extra requires Python 3.10+. Aurora remains untested; see
@@ -35,7 +37,7 @@ python -m pip install -r requirements.txt
 cloudoll start -n myapp
 ```
 
-The generated project targets 4.4.0 and receives random JWT and session secrets
+The generated project targets 4.5.0 and receives random JWT and session secrets
 in `.env`. Controllers
 and authentication middleware are examples: replace demo credentials and supply
 your own authorization and upload policies before deployment.
@@ -121,8 +123,8 @@ python -m ruff check cloudoll tests
 python -m mypy
 python -m pytest -q -m 'not integration' -k 'not aws'
 python -m build
-python tests/check_wheel.py dist/cloudoll-4.4.0-py3-none-any.whl
-python tests/check_release.py dist/cloudoll-4.4.0-py3-none-any.whl
+python tests/check_wheel.py dist/cloudoll-4.5.0-py3-none-any.whl
+python tests/check_release.py dist/cloudoll-4.5.0-py3-none-any.whl
 ```
 
 Native integration tests require explicitly configured disposable databases.

{cloudoll-4.4.0 → cloudoll-4.5.0}/README.zh-CN.md

@@ -6,7 +6,9 @@
 
 Cloudoll 是一个基于 aiohttp 的 Python Web 开发库，提供路由自动注册、中间件、模板渲染、会话管理、JWT、数据库访问及命令行脚手架，帮助快速构建 Web 应用和微服务。
 
-本文适用于 Cloudoll 4.4.0。从 3.x 升级前请阅读下方迁移说明。
+ORM 内置 SQLite 支持，MySQL/PostgreSQL 按需安装驱动；`Application.run()` 可直接加载 PEM 证书启用 HTTPS。
+
+本文适用于 Cloudoll 4.5.0。从 3.x 升级前请阅读下方迁移说明。
 
 ## AI 辅助开发
 
@@ -25,10 +27,10 @@ Cloudoll 是一个基于 aiohttp 的 Python Web 开发库，提供路由自动
 
 ## 安装
 
-安装 4.4.0：
+安装 4.5.0：
 
 ```sh
-python -m pip install 'cloudoll==4.4.0'
+python -m pip install 'cloudoll==4.5.0'
 ```
 
 在当前仓库目录安装源码：
@@ -41,16 +43,16 @@ python -m pip install -e .
 
 ```sh
 # MySQL
-python -m pip install 'cloudoll[mysql]==4.4.0'
+python -m pip install 'cloudoll[mysql]==4.5.0'
 
 # PostgreSQL
-python -m pip install 'cloudoll[postgres]==4.4.0'
+python -m pip install 'cloudoll[postgres]==4.5.0'
 
 # AWS 数据库驱动
-python -m pip install 'cloudoll[aws]==4.4.0'
+python -m pip install 'cloudoll[aws]==4.5.0'
 
 # Redis / Memcached 会话或 Redis 数据库
-python -m pip install 'cloudoll[cache]==4.4.0'
+python -m pip install 'cloudoll[cache]==4.5.0'
 ```
 
 源码开发时可用 `python -m pip install -e '.[mysql,postgres,cache,dev]'` 安装。
@@ -546,8 +548,8 @@ cloudoll restart -n myapp
 python -m pip install -e '.[mysql,postgres,cache,dev]'
 python -m pytest -q -m 'not integration' -k 'not aws'
 python -m build
-python tests/check_wheel.py dist/cloudoll-4.4.0-py3-none-any.whl
-python tests/check_release.py dist/cloudoll-4.4.0-py3-none-any.whl
+python tests/check_wheel.py dist/cloudoll-4.5.0-py3-none-any.whl
+python tests/check_release.py dist/cloudoll-4.5.0-py3-none-any.whl
 ```
 
 最后一条命令用于验证构建出的 wheel，版本号变化后请调整文件名。测试使用模拟数据库连接及本机 HTTP 服务；真实数据库连接和 AWS 故障切换需要单独的集成环境。

cloudoll-4.5.0/cloudoll/__init__.py

@@ -0,0 +1 @@
+__version__ = "4.5.0"

{cloudoll-4.4.0 → cloudoll-4.5.0}/cloudoll/ai/docs/changelog.md

@@ -6,10 +6,16 @@ icon: logs
 
 # 更新日志
 
-按版本倒序记录 Cloudoll 的主要变化。当前使用方法请以本站 4.4.0 文档为准，历史功能不代表当前仍保留相同 API。
+按版本倒序记录 Cloudoll 的主要变化。当前使用方法请以本站 4.5.0 文档为准，历史功能不代表当前仍保留相同 API。
 
 历史版本日期为对应源码提交日期。各版本源码可通过表格中的链接查看。
 
+## 4.5.0
+
+- `Application.run()` 支持通过证书路径或自定义 `ssl.SSLContext` 启用 HTTPS。
+- ORM 新增基于 `aiosqlite` 的 SQLite 引擎、事务、流式查询、upsert 和建表/模型反向生成支持。
+- SQLite 事务可选择 `DEFERRED`、`IMMEDIATE` 或 `EXCLUSIVE` 模式；SQLite 不支持行级 `FOR UPDATE`。
+
 ## 4.4.0
 
 缺陷修复版本，向后兼容；Python 3.9+ 仍然支持。

{cloudoll-4.4.0 → cloudoll-4.5.0}/cloudoll/ai/docs/config.md

@@ -113,7 +113,9 @@ database:
     url: $POSTGRES_URL
 ```
 
-URL 格式分别是 `mysql://用户:密码@主机:3306/库名` 和 `postgres://用户:密码@主机:5432/库名`。账号、密码含特殊字符时需 URL 编码；分字段配置更容易避免编码错误。显式关键字参数优先于 URL 中的同名选项。
+URL 格式分别是 `mysql://用户:密码@主机:3306/库名`、`postgres://用户:密码@主机:5432/库名` 和 `sqlite:///相对或绝对路径.db`。账号、密码含特殊字符时需 URL 编码；分字段配置更容易避免编码错误。显式关键字参数优先于 URL 中的同名选项。
+
+SQLite 无需额外驱动；内存数据库可写为 `sqlite:///:memory:`。服务器可在 `server` 下配置 `ssl_certfile` 与 `ssl_keyfile` 启用 HTTPS，也可通过 Python API 传入自定义 `ssl_context`。
 
 在请求中通过 `request.app.db.mysql` / `request.app.db.postgres` 取连接。详细的超时、TLS、事务和 SQL 日志见[数据库](/database)及[日志](/logs)。
 

{cloudoll-4.4.0 → cloudoll-4.5.0}/cloudoll/ai/docs/database.md

@@ -6,7 +6,7 @@ icon: database
 
 # 数据库与 ORM
 
-本页介绍 Cloudoll 4.4.0 的原生 MySQL/PostgreSQL 接口，包括保存点、流式查询与 SQL 日志。从 3.x 升级请先阅读[升级说明](/migration)。
+本页介绍 Cloudoll 4.5.0 的原生 MySQL、PostgreSQL 与 SQLite 接口，包括保存点、流式查询与 SQL 日志。从 3.x 升级请先阅读[升级说明](/migration)。
 
 ## 应用代码默认使用实体
 
@@ -26,7 +26,7 @@ Python API 的准确拼写是 `order_by` 和大写 `As`，不是 `orderby` / `as
 ## 安装与连接
 
 ```sh
-python -m pip install 'cloudoll[mysql,postgres]==4.4.0'
+python -m pip install 'cloudoll[mysql,postgres]==4.5.0'
 ```
 
 Web 应用在 [database 配置](/config)中声明连接，由框架建立、关闭，通过 `request.app.db.别名` 获取。不要每个请求创建一个池，也不要在请求结束时关闭共享池。
@@ -50,7 +50,9 @@ if __name__ == "__main__":
     asyncio.run(main())
 ```
 
-DATABASE_URL 为 mysql://... 或 postgres://...。`create_engine` 是异步函数；同一个引擎在所属事件循环中复用。关闭前先退出所有事务和流式查询上下文。
+DATABASE_URL 可以是 `mysql://...`、`postgres://...` 或 `sqlite:///path/to/app.db`。SQLite 内存库使用 `sqlite:///:memory:`，引擎会将池大小限制为 1，确保连接间共享同一个内存数据库。`create_engine` 是异步函数；同一个引擎在所属事件循环中复用。关闭前先退出所有事务和流式查询上下文。
+
+SQLite 基于 `aiosqlite`，支持 ORM 查询、绑定参数、批量写入、事务、嵌套保存点、流式读取和模型建表。`async with engine.transaction(mode="IMMEDIATE")` 可在并发写前取得 SQLite 写锁；mode 支持 `DEFERRED`、`IMMEDIATE`、`EXCLUSIVE`。SQLite 不支持行级 `FOR UPDATE`。其余高级 SQL 函数和 DDL 特性仍按 SQLite 自身能力限制。
 
 ## 模型定义
 

{cloudoll-4.4.0 → cloudoll-4.5.0}/cloudoll/ai/docs/deployment.md

@@ -6,6 +6,20 @@ icon: deployment
 
 # 应用部署
 
+## 直接启用 HTTPS
+
+Cloudoll 可直接使用 PEM 证书和私钥：
+
+```yaml
+server:
+  host: 0.0.0.0
+  port: 443
+  ssl_certfile: /etc/ssl/cloudoll/fullchain.pem
+  ssl_keyfile: /etc/ssl/cloudoll/privkey.pem
+```
+
+`Application.run()` 会创建 TLS server context 并加载证书。也可在 Python 中通过 `server.ssl_context` 传入自定义 `ssl.SSLContext`，但不能与证书路径同时设置。证书续期后需重启进程以加载新证书。反向代理部署仍可由代理终止 TLS。
+
 ## 生产启动与 CLI 边界
 
 在应用根目录创建虚拟环境、安装经过验证并锁定版本的 requirements.txt：

{cloudoll-4.4.0 → cloudoll-4.5.0}/cloudoll/ai/docs/start.md

@@ -9,7 +9,7 @@ path: /
 
 Cloudoll 是基于 aiohttp 的异步 Python Web 库，提供路由、模板、Session、ORM 和 CLI。
 
-> 本文档按 Cloudoll 4.4.0 源码更新，实际发布状态以 PyPI 为准。从 3.x 升级前，请先阅读[升级说明](/migration)，确认配置和 API 的兼容性变化。
+> 本文档按 Cloudoll 4.5.0 源码更新，实际发布状态以 PyPI 为准。从 3.x 升级前，请先阅读[升级说明](/migration)，确认配置和 API 的兼容性变化。
 
 ## 环境与安装
 
@@ -18,8 +18,8 @@ Cloudoll 是基于 aiohttp 的异步 Python Web 库，提供路由、模板、Se
 ```sh
 python3 -m venv .venv
 source .venv/bin/activate
-python -m pip install 'cloudoll==4.4.0'
-python -m pip install 'cloudoll[mysql,postgres,cache]==4.4.0'
+python -m pip install 'cloudoll==4.5.0'
+python -m pip install 'cloudoll[mysql,postgres,cache]==4.5.0'
 ```
 
 Windows PowerShell 激活命令为 `.venv\Scripts\Activate.ps1`。只使用 Web 功能时不需要安装数据库 extras。

{cloudoll-4.4.0 → cloudoll-4.5.0}/cloudoll/ai/manifest.json

@@ -1,5 +1,5 @@
 {
-  "version": "4.4.0",
+  "version": "4.5.0",
   "documents": {
     "ai": {
       "source": "src/views/ai.md",
@@ -14,12 +14,12 @@
     "changelog": {
       "source": "src/views/changelog.md",
       "url": "https://cloudoll.chuchur.com/changelog",
-      "sha256": "54ac62d5384e70db5911d882df93e5a3b8fef5d7983363cb342c4a0d905d2fc8"
+      "sha256": "848012a2052408b97d5c1f5e2c23fe3a109e5787f513ff9fe8593aa8df5a9cd0"
     },
     "config": {
       "source": "src/views/config.md",
       "url": "https://cloudoll.chuchur.com/config",
-      "sha256": "2e4f7c9dfbca9e71e179512e67ae188f608fdf4b84bdcb64a424fccfeb7603d0"
+      "sha256": "4a3437732d80bf9df315d7e67b05fc7582857082fd9f4ba646cead3f66f1fdde"
     },
     "cookie-and-session": {
       "source": "src/views/cookie-and-session.md",
@@ -29,12 +29,12 @@
     "database": {
       "source": "src/views/database.md",
       "url": "https://cloudoll.chuchur.com/database",
-      "sha256": "aed56825418dd4f2aa1f587cee45f0a429905aef03428b4f2e6b5900ee8e9f39"
+      "sha256": "1f48429ffa99df95d554cc70429bb30098e3bc4c6601e97d0ab0dbd0b5cffd76"
     },
     "deployment": {
       "source": "src/views/deployment.md",
       "url": "https://cloudoll.chuchur.com/deployment",
-      "sha256": "1ba77f6764d38ec0c510f15432de9b33608b5225a89fd6320c96f449d36f0d82"
+      "sha256": "5e4774561bf1cd91229314bb33659c8c2f36d3e7ab6889133489f22476eeac0e"
     },
     "event-source": {
       "source": "src/views/event-source.md",
@@ -89,7 +89,7 @@
     "start": {
       "source": "src/views/start.md",
       "url": "https://cloudoll.chuchur.com/",
-      "sha256": "d9c42da379ce7913f8c058f2fdede3251dd4ae61473fce1af1cb023abea04f27"
+      "sha256": "4541305a1d3414cf4f62aa09c874d47c78dbafe3d03262d6bf1bd693f904b329"
     },
     "structure": {
       "source": "src/views/structure.md",

{cloudoll-4.4.0 → cloudoll-4.5.0}/cloudoll/clitool/m2d.py

@@ -133,6 +133,16 @@ async def create_model(pool: DatabaseEngine, table_name: str) -> str:
 
 async def get_table_cols(pool: DatabaseEngine, table_name: str) -> list[Any]:
     driver = driver_name(pool.driver)
+    if driver == "sqlite":
+        rows = await pool.all(
+            "PRAGMA table_info(" + dialect_for(driver).identifier(table_name) + ")", None
+        )
+        return [
+            {"field": row["name"], "default": row["dflt_value"],
+             "column_type": row["type"], "null": "NO" if row["notnull"] else "YES",
+             "Key": "PRI" if row["pk"] else "", "comment": None}
+            for row in rows
+        ]
     if driver == "mysql":
         return await pool.all(
             "SHOW FULL COLUMNS FROM " + dialect_for(driver).identifier(table_name), None
@@ -163,6 +173,10 @@ async def get_table_cols(pool: DatabaseEngine, table_name: str) -> list[Any]:
 
 async def get_all_tables(pool: DatabaseEngine) -> list[Any]:
     driver = driver_name(pool.driver)
+    if driver == "sqlite":
+        return await pool.all(
+            "SELECT name AS table_name FROM sqlite_master WHERE type = 'table' AND name NOT LIKE 'sqlite_%' ORDER BY name", None
+        )
     if driver == "mysql":
         return await pool.all(
             "SELECT TABLE_NAME AS table_name FROM information_schema.tables WHERE TABLE_SCHEMA = DATABASE() AND TABLE_TYPE = 'BASE TABLE' ORDER BY TABLE_NAME",
@@ -299,7 +313,17 @@ def _default(value: Any, kind: str, postgres: bool) -> Any:
 
 
 def get_col(field: dict[str, Any], driver: str = "mysql") -> dict[str, Any]:
-    postgres = driver_name(driver) == "postgres"
+    normalized_driver = driver_name(driver)
+    if normalized_driver == "sqlite":
+        kind = field.get("column_type", "text").lower().split("(", 1)[0]
+        type_map = {"int": "integer", "bigint": "integer", "real": "float", "numeric": "numeric", "blob": "text"}
+        return dict(name=field["field"], column_type=type_map.get(kind, kind),
+                    primary_key=field.get("Key") == "PRI", default=_default(field.get("default"), kind, False),
+                    charset=None, max_length=None, scale_length=None,
+                    auto_increment=bool(field.get("auto_increment", False)),
+                    NOT_NULL=field.get("null") == "NO", update_generated=False,
+                    unsigned=False, comment=None)
+    postgres = normalized_driver == "postgres"
     if not postgres:
         match = re.fullmatch(
             r"([a-z]+)(?:\((\d+)(?:,(\d+))?\))?( unsigned)?", field["Type"].lower()

{cloudoll-4.4.0 → cloudoll-4.5.0}/cloudoll/clitool/schema.py

@@ -42,11 +42,13 @@ TYPES = (
 
 
 def driver_name(driver: str) -> str:
+    if driver == "sqlite":
+        return driver
     if driver == "mysql":
         return driver
     if driver in {"postgres", "postgresql", "postgressql"}:
         return "postgres"
-    raise ValueError("Schema generation supports native MySQL/PostgreSQL only")
+    raise ValueError("Schema generation supports native MySQL/PostgreSQL/SQLite only")
 
 
 def current_timestamp(value: Any) -> Optional[str]:
@@ -91,6 +93,7 @@ def compile_column(
 ) -> CompiledQuery:
     driver = driver_name(driver)
     pg = driver == "postgres"
+    sqlite = driver == "sqlite"
     dialect = dialect_for(driver)
     kind = field.column_type.lower()
     if kind not in TYPES:
@@ -105,6 +108,8 @@ def compile_column(
         "timestamp without time zone",
     }:
         raise ValueError(f"MySQL cannot represent {kind}")
+    if sqlite and (field.unsigned or field.charset or field.update_generated or field.comment is not None):
+        raise ValueError("SQLite schema generation does not support unsigned, charset, update-generated or comment metadata")
     mapped = (
         {
             "tinyint": "smallint",
@@ -121,6 +126,15 @@ def compile_column(
         if pg
         else {"integer": "int", "real": "float", "double precision": "double"}
     ).get(kind, kind)
+    if sqlite:
+        mapped = {
+            "tinyint": "INTEGER", "smallint": "INTEGER", "mediumint": "INTEGER",
+            "int": "INTEGER", "integer": "INTEGER", "bigint": "INTEGER",
+            "boolean": "INTEGER", "float": "REAL", "double": "REAL",
+            "double precision": "REAL", "decimal": "NUMERIC", "numeric": "NUMERIC",
+            "varchar": "TEXT", "char": "TEXT", "mediumtext": "TEXT", "longtext": "TEXT",
+            "datetime": "TEXT", "timestamp": "TEXT", "date": "TEXT", "json": "TEXT",
+        }.get(kind, kind.upper())
     size, scale = field.max_length, field.scale_length
     if size is not None:
         size = _integer(size, "max_length", 0 if kind in TEMPORAL else 1)
@@ -158,13 +172,15 @@ def compile_column(
     if field.auto_increment:
         if kind not in INTEGER or field.default is not None:
             raise ValueError("Identity requires an integer without a separate default")
-        if not pg and not field.primary_key:
+        if not pg and not sqlite and not field.primary_key:
             raise ValueError(
                 "MySQL AUTO_INCREMENT requires a primary key in generated schemas"
             )
-        sql += " GENERATED BY DEFAULT AS IDENTITY" if pg else " AUTO_INCREMENT"
+        sql += " GENERATED BY DEFAULT AS IDENTITY" if pg else (" AUTOINCREMENT" if sqlite else " AUTO_INCREMENT")
     if field.primary_key:
         sql += " PRIMARY KEY"
+    if sqlite and field.auto_increment and field.primary_key:
+        sql = dialect.identifier(field.name) + " INTEGER PRIMARY KEY AUTOINCREMENT"
     if field.NOT_NULL:
         sql += " NOT NULL"
     params: list[Any] = []
@@ -186,7 +202,7 @@ def compile_column(
         expression = current_timestamp(field.default) if kind in TEMPORAL else None
         default_sql = expression or parameter(field.default)
         # MySQL text/JSON defaults require expression parentheses (8.0.13+).
-        if not pg and kind in {"text", "mediumtext", "longtext", "json"}:
+        if not pg and not sqlite and kind in {"text", "mediumtext", "longtext", "json"}:
             default_sql = "(" + default_sql + ")"
         sql += " DEFAULT " + default_sql
     if field.update_generated:
@@ -195,7 +211,7 @@ def compile_column(
         sql += " ON UPDATE CURRENT_TIMESTAMP" + (
             f"({size})" if size is not None else ""
         )
-    if field.comment is not None and not pg:
+    if field.comment is not None and not pg and not sqlite:
         sql += " COMMENT " + parameter(field.comment)
     return CompiledQuery(sql, params)
 

{cloudoll-4.4.0 → cloudoll-4.5.0}/cloudoll/orm/__init__.py

@@ -52,6 +52,10 @@ async def create_engine(**kw: Any) -> Any:
         from .postgres import Postgres
 
         return await Postgres().create_engine(**{**configs, **query})
+    elif driver == "sqlite":
+        from .sqlite import SQLite
+
+        return await SQLite().create_engine(**{**configs, **query})
     elif driver in ["redis", "rediss"]:
         from redis import asyncio as aioredis
 
@@ -77,6 +81,10 @@ def __getattr__(name: str) -> Any:
         from .postgres import Postgres
 
         return Postgres
+    if name == "SQLite":
+        from .sqlite import SQLite
+
+        return SQLite
     if name == "Query":
         from .query import Query
 

{cloudoll-4.4.0 → cloudoll-4.5.0}/cloudoll/orm/aws_engine.py

@@ -312,7 +312,9 @@ class AwsEngine(SavepointMixin, MeteBase):
             raise
 
     @asynccontextmanager
-    async def transaction(self) -> AsyncIterator[AwsEngine]:
+    async def transaction(self, *, mode: Optional[str] = None) -> AsyncIterator[AwsEngine]:
+        if mode is not None:
+            raise ValueError("Transaction mode is only supported by SQLite")
         if self._state() is not None:
             async with self.savepoint():
                 yield self

{cloudoll-4.4.0 → cloudoll-4.5.0}/cloudoll/orm/base.py

@@ -27,7 +27,7 @@ class QueryTypes(Enum):
 class MeteBase:
     driver: str
 
-    def transaction(self) -> AbstractAsyncContextManager[Any]:
+    def transaction(self, *, mode: Optional[str] = None) -> AbstractAsyncContextManager[Any]:
         raise NotImplementedError("This engine does not implement transactions")
 
     async def __aenter__(self: E) -> E:

{cloudoll-4.4.0 → cloudoll-4.5.0}/cloudoll/orm/compiler.py

@@ -125,17 +125,28 @@ class SQLCompiler:
             if isinstance(amount, bool) or not isinstance(amount, int) or amount < 0:
                 raise ValueError("Interval amount must be a non-negative integer")
             sign = ">=" if op.startswith("LASTED_") else "<"
+            if getattr(self.dialect, "is_sqlite", False):
+                modifiers = {
+                    "DAY": "days", "MONTH": "months", "YEAR": "years",
+                    "HOUR": "hours", "MINUTE": "minutes", "SECOND": "seconds",
+                }
+                return f"datetime({column}) {sign} datetime('now', '-{amount} {modifiers[unit]}')"
             return f"{column} {sign} CURRENT_TIMESTAMP - {self.dialect.interval(amount, unit)}"
         if op in {"IS_TODAY", "THIS_WEEK", "THIS_MONTH", "THIS_YEAR"}:
             if op == "IS_TODAY":
                 return f"DATE({column}) = CURRENT_DATE"
             unit = op.split("_")[1].lower()
+            if getattr(self.dialect, "is_sqlite", False):
+                formats = {"week": "%Y-%W", "month": "%Y-%m", "year": "%Y"}
+                return f"strftime('{formats[unit]}', {column}) = strftime('{formats[unit]}', 'now')"
             if self.dialect.is_postgres:
                 return f"DATE_TRUNC('{unit}', {column}) = DATE_TRUNC('{unit}', CURRENT_DATE)"
             patterns = {"week": "%x-%v", "month": "%Y-%m", "year": "%Y"}
             pattern = patterns[unit]
             return f"DATE_FORMAT({column}, '{pattern}') = DATE_FORMAT(CURRENT_DATE, '{pattern}')"
         if op in {"JSON_CONTAINS_ARRAY", "JSON_CONTAINS_OBJECT"}:
+            if getattr(self.dialect, "is_sqlite", False):
+                raise NotImplementedError("JSON containment expressions are not supported by the SQLite ORM")
             values = node.rpt if isinstance(node.rpt, (tuple, list)) else (node.rpt,)
             args = ",".join(self.expression(v, params) for v in values)
             kind = "array" if op.endswith("ARRAY") else "object"
@@ -149,6 +160,8 @@ class SQLCompiler:
                 return f"STRING_AGG(CAST({column} AS TEXT), ',')"
             return f"GROUP_CONCAT({column})"
         if op == "DATE_FORMAT":
+            if getattr(self.dialect, "is_sqlite", False):
+                raise NotImplementedError("DATE_FORMAT uses MySQL format strings and is not supported by SQLite")
             if self.dialect.is_postgres:
                 raise NotImplementedError(
                     "DATE_FORMAT uses MySQL format strings; use PostgreSQL SQL explicitly"
@@ -244,6 +257,8 @@ class SQLCompiler:
         if state.offset is not None:
             sql += f" OFFSET {state.offset}"
         if state.lock:
+            if getattr(self.dialect, "is_sqlite", False):
+                raise ValueError("SQLite does not support row-level FOR UPDATE locks; use an IMMEDIATE transaction")
             if state.lock not in {
                 "FOR UPDATE",
                 "FOR UPDATE NOWAIT",
@@ -328,7 +343,7 @@ class SQLCompiler:
         validate(update_fields, "update_fields")
         if model.__primary_key__ in update_fields:
             raise ValueError("upsert cannot update the primary key")
-        if not self.dialect.is_postgres and conflict_fields is not None:
+        if not self.dialect.is_postgres and not getattr(self.dialect, "is_sqlite", False) and conflict_fields is not None:
             raise ValueError(
                 "MySQL cannot select conflict_fields; any unique key can conflict"
             )
@@ -348,13 +363,21 @@ class SQLCompiler:
                 + ",".join(self.dialect.identifier(key) for key in target)
                 + ") DO UPDATE SET "
             )
+        elif getattr(self.dialect, "is_sqlite", False):
+            target = conflict_fields or ([model.__primary_key__] if model.__primary_key__ else [])
+            validate(target, "conflict_fields")
+            if set(target) & set(update_fields):
+                raise ValueError("upsert cannot update conflict target fields")
+            compiled_sql = (
+                compiled.sql + " ON CONFLICT ("
+                + ",".join(self.dialect.identifier(key) for key in target)
+                + ") DO UPDATE SET "
+            )
         else:
             compiled_sql = compiled.sql + " ON DUPLICATE KEY UPDATE "
         incoming = dict(zip(keys, values))
         # Bind incoming values again: avoids MySQL's deprecated VALUES(column).
-        compiled_sql += ",".join(
-            self.dialect.identifier(key) + "=?" for key in update_fields
-        )
+        compiled_sql += ",".join(self.dialect.identifier(key) + "=?" for key in update_fields)
         return CompiledQuery(
             compiled_sql, compiled.params + [incoming[key] for key in update_fields]
         )

{cloudoll-4.4.0 → cloudoll-4.5.0}/cloudoll/orm/dialects.py

@@ -80,7 +80,22 @@ class PostgreSQLDialect(MySQLDialect):
         return f"INTERVAL '{amount} {unit.lower()}'"
 
 
+class SQLiteDialect(MySQLDialect):
+    is_postgres = False
+    is_sqlite = True
+    quote = '"'
+
+    def _code(self, text: str) -> str:
+        text = re.sub(r"\bCURDATE\(\)", "date('now')", text, flags=re.I)
+        return re.sub(r"\bNOW\(\)", "datetime('now')", text, flags=re.I)
+
+    def prepare(self, sql: str, *, escape_percent: bool = False) -> str:
+        return self._code(sql)
+
+
 def dialect_for(driver: str) -> MySQLDialect:
+    if driver == "sqlite":
+        return SQLiteDialect()
     if driver in {
         "postgres",
         "postgresql",

{cloudoll-4.4.0 → cloudoll-4.5.0}/cloudoll/orm/engine.py

@@ -239,13 +239,19 @@ class AsyncEngine(SavepointMixin, MeteBase):
             raise
 
     @asynccontextmanager
-    async def transaction(self: AE) -> AsyncIterator[AE]:
+    async def transaction(self: AE, *, mode: Optional[str] = None) -> AsyncIterator[AE]:
         """Pin a connection to the current task; commit or roll back on exit.
 
         Nested scopes use savepoints. DDL/explicit transaction-control SQL is
         unsupported. Use one transaction per task; no automatic write retries.
         """
+        if mode is not None and getattr(self, "driver", None) != "sqlite":
+            raise ValueError("Transaction mode is only supported by SQLite")
+        if mode is not None and mode not in {"DEFERRED", "IMMEDIATE", "EXCLUSIVE"}:
+            raise ValueError("SQLite transaction mode must be DEFERRED, IMMEDIATE or EXCLUSIVE")
         if self._state() is not None:
+            if mode is not None:
+                raise TransactionError("Cannot change the mode of a nested transaction")
             async with self.savepoint():
                 yield self
             return
@@ -262,7 +268,7 @@ class AsyncEngine(SavepointMixin, MeteBase):
         try:
             try:
                 await asyncio.wait_for(
-                    self._control(connection, "BEGIN"), self.query_timeout
+                    self._control(connection, "BEGIN" + (f" {mode}" if mode else "")), self.query_timeout
                 )
             except (asyncio.CancelledError, asyncio.TimeoutError):
                 connection.close()
@@ -396,7 +402,7 @@ class AsyncEngine(SavepointMixin, MeteBase):
         rows = None
         try:
             await asyncio.wait_for(
-                self._control(connection, "START TRANSACTION READ ONLY"),
+                self._control(connection, "BEGIN" if self.driver == "sqlite" else "START TRANSACTION READ ONLY"),
                 self.query_timeout,
             )
             self._log_sql("STREAM", sql, params)