Logics is a user-friendly formula language with a subset of Python''s expression syntax
Project description
About
Logics is a formula language with the aim of flexibly implementing programmable logics such as conditions, calculations or validations, so that developers as well as administrators and experienced users can use such expressions within a framework specified by the system to influence, calculate or otherwise dynamically adapt certain processes of a system.
The language can be compared to Excel's formula language, where expressions can be defined based on variables and calculated dynamically. But Logics has a Python-like syntax at this point, but provides its own semantics, so that numbers and strings, for example, can be directly linked, None-values are assumed as 0, and Exceptions are generally not raised during execution.
In addition, Logics provides a sandbox environment, which does not allow potential attackers to break out of the provided environment or slow down the system. There are appropriate security mechanisms integrated into the language for this purpose.
Logics has been developed at Mausbrand Informationssysteme GmbH as part of the ViUR ecosystem, but it is a standalone project that can also be used outside the ViUR context.
Previous uses of Logics included
- event-based state evaluation
- dynamic rule systems
- programmable conditions for filters
- template engine generating text modules
- questionnaires with queries that depend on previous answers
Usage
Since Logics is used on both the client and server side, the language has been implemented in two separate implementations:
- logics-js is a pure (vanilla) JavaScript implementation of Logics provided as npm-package.
- logics-py is a pure Python 3.10 implementation of Logics provided as PyPI-package, with no other dependencies.
Both packages are under recent development and not stable right now. They are maintained in separate version numbers, which is planned to be changed soon, when they become almost feature-complete.
Using Logics in JavaScript:
// npm install logics-js
import Logics from "logics-js";
let logics = new Logics("a + 2 * 3 + b");
console.log(logics.run({a: 1, b: "-Logics"}).toString()); // "7-Logics"
Using Logics in Python:
# pip install logics-py
from logics import Logics
logics = Logics("a + 2 * 3 + b")
print(logics.run({"a": 1, "b": "-Logics"})) # "7-Logics"
Features
- Secure, native, running in a sandboxed environment apart from the host language
- Disallows variable assignment, except in comprehensions
- logics-js: Implementation in JavaScript
- logics-py: Implementation in Python
- Python-inspired syntax and semantics
- Make use of all standard operators
- Unary
+
,-
,~
,not
- Binary
+
,*
,-
,/
,//
,**
- Comparison
==
,!=
,<>
,<
,<=
,>
,>=
,in
,not in
- Logical
and
,or
- Conditions
y if x else z
- Comprehensions
[x for x in y if z]
- Unary
- Slices
x[:]
- Attribute access
x[y]
# comments
in separate lines- Dedicated Value object abstraction of native types for
None
True
,False
int
,float
,str
list
for lists (similar to arrays)dict
for dicts (similar to structured objects)
- Make use of all standard operators
- Provides a set of functions that can be used in expressions
- Extendable to custom functions
- Embeddable into other languages
Logics
vs. Python
Logics does look like Python, but it isn't Python!
It differs both in syntax and semantics:
- Expressions can be used with arbitrary whitespace and line-breaks
- There are no methods on objects, but functions that work on values, e.g.
- Python's
dict.keys()
iskeys(dict)
in Logics - Python's
";".join(["a", "b", "c"])
isjoin(["a", "b", "c"], ";")
in Logics
- Python's
- No exceptions during runtime, but a defined behavior
- Access to e.g. invalid index or key just returns
None
- Access of undefined variables returns
None
- Division by zero returns string
"#ERR: Division by zero"
- Access to e.g. invalid index or key just returns
- Dynamic and automatic value conversion
- e.g. the content of strings is automatically converted when used in calculations,
so
"42" ** 3
produces 74088, and not a TypeError.
- e.g. the content of strings is automatically converted when used in calculations,
so
Building & Packaging
Logics is built using the UniCC LALR(1) Parser Generator, which supports generating parsers in multiple target languages. UniCC should be compiled from source, as the latest version 1.8 is required.
Whenever something is changed on the syntax, ensure unicc
is installed properly and run make
, which regenerates the parser modules.
Shared test cases
Logics-related test cases are placed as *.lgx
-files into the tests/
folder. They are shared by both implementations.
# In lgx-files, variables can be set using `#set:` as shown below.
# The part behind the second colon may contain any logics expression, including variables.
#set:x:6 * 7
#
# To verify if an expression produces an expected result, provide the expression followed by an `#expect:`.
#
x
#expect:42
x * x
#expect:1764
logics-js
Formatting
To format the source code, use prettier as follows:
cd logics-js
npm run fmt
Tests
Tests are implemented using Mocha:
cd logics-js
npm test
Packaging
Publish on npmjs.com:
cd logics-js
npm publish
logics-py
Formatting
To format the source code, use Black as follows:
cd logics-py
pipenv run fmt
Tests
Tests are implemented using pytest:
cd logics-py
pipenv run test
Packaging
Publish on PyPI:
cd logics-py
pipenv install --dev
pipenv run build
pipenv run publish
pipenv run clean
pipenv --rm
License
Copyright © 2023 by Mausbrand Informationssysteme GmbH.
Mausbrand and ViUR are registered trademarks of Mausbrand Informationssysteme GmbH.
Logics is free software under the MIT license.
Please see the LICENSE file for details.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distributions
Built Distribution
File details
Details for the file logics_py-0.0.11-py3-none-any.whl
.
File metadata
- Download URL: logics_py-0.0.11-py3-none-any.whl
- Upload date:
- Size: 17.1 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/4.0.2 CPython/3.10.11
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | f00258640697ecc777e4c085a439950ac8c641add046448647b28265f6cd7a0d |
|
MD5 | ef872ee18032272dd09baeb5168994fb |
|
BLAKE2b-256 | 545efe1b2a3c6fdf159a6179d97d40a1968f4cf8c3ed3ed2b71b426f55a92b15 |