Skip to main content

zen-engine enrichment for custom node spec and cache ZenDecision objects for performance

Project description

zen-rule

zen-rule 是 zen-engine 加强版本:

  1. 提供多个 decision 的缓存.
  2. 提供自定义节点中多个函数调用表达式的定义, 解析和调用规范.

example

推荐线上使用 decision 缓存模式, 这样规则只需要加载,解析一次后重复使用,提高系统性能.
每次先判断zenRule实例中是否有规则的缓存,如果没有则去加载规则图; 如果有就直接通过规则键去调用规则对入参进行处理, 得到最后规则的输出.

from pathlib import Path
from zen_rule import ZenRule, udf


async def test_zenrule():
    """
        推荐线上生产环境使用此模式进行规则执行, 可以缓存决策对象, 提高性能.
    """
    zr = ZenRule({})
    basedir = Path(__file__).parent
    filename = basedir / "graph" / "custom_v3.json"
    key = filename

    if not zr.get_decision_cache(key):
        # 根据实际情况去加载规则图的内容.
        with open(filename, "r", encoding="utf8") as f:
            logger.warning(f"graph json: %s", filename)
            content =  f.read()
        zr.create_decision_with_cache_key(key, content)  # 将规则图缓存在键下, 这样可以只读取规则一次,解析一次,然后复用决策对象 decision
    result = await zr.async_evaluate(key, {"input": 7, "myvar": 15})
    print("zen rule custom_v3 result:", result)

如果提供 loader 函数, 就是如下调用示例.

from pathlib import Path
from zen_rule import ZenRule, udf


def loader(key):
    basedir = Path(__file__).parent
    filename = basedir / "graph" / key
    with open(filename, "r", encoding="utf8") as f:
        logger.warning(f"graph json: %s", filename)
        return f.read()


async def test_zenrule_with_loader():
    zr = ZenRule({"loader": loader})

    result = await zr.async_evaluate("custom_v3.json", {"input": 7, "myvar": 15})
    print("zen rule custom_v3 result:", result)

    result = await zr.async_evaluate("custom_v3.json", {"input": 7, "myvar": 15})
    print("zen rule custom_v3 result2:", result)

示例程序请参考 main.py, 运行 python main.py 即可运行测试示例.

包含自定义函数示例的规则:
custom_v3.json

自定义算子规范v3

在决策引擎中, 自定义算子主要承担外部数据查询, 状态交互, 以及一些自定义功能拓展. 考虑到界面中会对自定义算子进行分类, 这时候在某个自定义类别的节点中只能访问这个类别的函数, 这样有利于降低最后用户的使用难度.

alt text

格式一

如果支持不同自定义算子的嵌套, 那么这些节点分类就形同虚设. 考虑到自定义算子再未来的功能以及演化, 暂时决定自定义算子不支持嵌套调用和解析. 参考 zen-engine 的表达式测试用例. 为了简化参数的解析, 决定选用 ;; 作为函数的分隔符号.

foo;;myvar ;;max([5, 8, 2, 11, 7]);;rand(100);; 'fccd;;jny' ;;3+4

表示 foo 算子传入了五个参数:

  1. zen 表达式变量 myvar
  2. zen 表达式函数 max([5, 8, 2, 11, 7])
  3. zen 表达式 rand(100)
  4. zen 表达式 'fccd;;jny'
  5. zen 表达式 3+4

解析后得到如下结构, 解释执行即可:

["foo", "myvar", "max([5, 8, 2, 11, 7])", "rand(100)", "fccd;;jny", "3+4"]

解析逻辑如下:

def parse_oprator_expr_v3(expr):
    # 不能简单使用字符串分割, 因为字符串中可能会有分隔符的模式出现, 比如:
    # foo ;; myvar ;; bar(zoo('fccd;;jny',6, 3.14),'a');; a+string(xxx)
    # foo;;myvar;;max([5, 8, 2, 11, 7]);;rand(100);; 'fccd;;jny' ;;3+4
    # expr.split(";;")
    pattern = r""";;(?=(?:[^"'`]*["'`][^"'`]*["'`])*[^"'`]*$)"""
    # To split the string by these semicolon:
    _parts = re.split(pattern, expr)
    parts = [i.strip() for i in _parts]  # 去掉表达式前后的空格
    return parts

格式二

第二种就是保留技术更为熟悉的函数嵌套调用结构. 这种格式需要解析函数的嵌套调用, 还需要解析各种常量语法结构, 比如 数字, 布尔,字符串,数组,object对象, 以及各种数组和字符串下标索引访问和切片访问语法格式, 还有中缀表达式(算术运算和逻辑运算), 非空判定, 三元表达式.

这部分需要使用上下无关文法定义解析或者peg语法解析.

foo(myvar,max([5, 8, 2, 11, 7]),rand(100), 'fccd;;jny', 3+4)

解析后格式如下:

[["foo", "myvar", "max([5, 8, 2, 11, 7])", "rand(100)", "'fccd;;jny'", "3+4"]]

此实现需要不断的去补充关于 zen 表达式语法结构的解析.
详细参考 tests/test_zen_expr_parser.py 中的实现.

$ python '/home/ryefccd/workspace/zen-rule/tests/test_zen_expr_parser.py'
foo(myvar,max([5, 8, 2, 11, 7]),rand(100), 'fccd;;jny', 3+4) --> [['foo', 'myvar', 'max([5, 8, 2, 11, 7])', 'rand(100)', "'fccd;;jny'", '3+4']]
...

目前使用 pyparsing 来解析, 将来考虑使用lark 来做语法解析.

解析后格式

{
    "id": "138b3b11-ff46-450f-9704-3f3c712067b2",
    "type": "customNode",
    "position": {
    "x": 470,
    "y": 240
    },
    "name": "customNode1",
    "content": {
    "kind": "sum",
    "config": {
        "version": "v2",
        "meta": {
        "user": "wanghao@geetest.com",
        "proj": "proj_id"
        },
        "prop1": "{{ a + 10 }}",
        "passThrough": true,
        "inputField": null,
        "outputPath": null,
        "expressions": [/*expressions 是约定的前端格式*/
        {
            "id": "52d41e3d-067d-4930-89bd-832b038cd08f",
            "key": "result",
            "value": "foo;;myvar ;;max([5, 8, 2, 11, 7]);;rand(100);; 'fccd;;jny' ;;3+4"
        }
        ],
        "expr_asts": [/*expr_asts 是后端解析动态产生的*/
        {
            "id": "52d41e3d-067d-4930-89bd-832b038cd08f",
            "key": "result",
            "value": ["foo", "myvar", "max([5, 8, 2, 11, 7])", "rand(100)", "\\'fccd;;jny\\'", "3+4"]
        }
        ]
    }
    }
}

develop

使用如下命令构建开发环境,安装依赖库

uv pip install -e .

使用此命令将当前包以编辑模式(--editable, -e)安装在当前的虚拟环境中, 这样可以使用 python main.py 运行程序即可.

(zen-rule) ryefccd@republic:~/workspace/zen-rule$ uv pip list
Package                  Version Editable project location
------------------------ ------- --------------------------------
...                      ...
zen-engine               0.49.1
zen-rule                 0.10.1  /home/ryefccd/workspace/zen-rule

包构建

先把 dist 包中的文件删除, 然后 uv build 执行构建.

rm -r dist  
uv build

包发布

需要在 pypi 上创建一个账号, 在完成 uv build 之后, 使用 uv publish 进行包的上传.

uv publish

To set your API token for PyPI, you can create a $HOME/.pypirc similar to:

[pypi]
username = __token__
password = <PyPI token>

Using a PyPI token
Building and publishing a package

unit tests

增加多个python版本的测试.

logging

Configuring Logging for a Library
What are the uses of Null handler in python logging?

reference

Setting up testing with pytest and uv

pypi src:

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

zen_rule-0.12.0.tar.gz (327.1 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

zen_rule-0.12.0-py3-none-any.whl (15.6 kB view details)

Uploaded Python 3

File details

Details for the file zen_rule-0.12.0.tar.gz.

File metadata

  • Download URL: zen_rule-0.12.0.tar.gz
  • Upload date:
  • Size: 327.1 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: uv/0.7.21

File hashes

Hashes for zen_rule-0.12.0.tar.gz
Algorithm Hash digest
SHA256 29bf674a5e6a378547a2887b7eb9c7f12d2efde1839c326f741f94169d1468d8
MD5 fc139fb56f7255b6b4fbef62c3981efe
BLAKE2b-256 890a68899052841738574ae6cc0ac2ed6e250128a4cbf250bc5578e8195a9444

See more details on using hashes here.

File details

Details for the file zen_rule-0.12.0-py3-none-any.whl.

File metadata

  • Download URL: zen_rule-0.12.0-py3-none-any.whl
  • Upload date:
  • Size: 15.6 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: uv/0.7.21

File hashes

Hashes for zen_rule-0.12.0-py3-none-any.whl
Algorithm Hash digest
SHA256 320d23091a8b64768fdaf83ee04722a21037302d3abe2afdfc85b7726195ed93
MD5 4884587ff07de7e96787e4b2d3c88925
BLAKE2b-256 00e29e9d20a4972df91bae040bff7b1783bfc6c197126372309cbfff9df37e4f

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page