A micro conditional javascript engine used to parse the raw logical and comparison expressions, evaluate the expression in the given data context, and provide access to a text form of the given expressions.

Revision: March 22, 2022.

This project has been developed to provide a shared conditional logic between front-end and back-end code, stored in JSON or in any other data serialization format.

Code documentation could be found here: https://briza-insurance.github.io/illogical/index.html.

The library is being build as CommonJS module and ESM.

npm install -D @briza/illogical

yarn add @briza/illogical -D

Table of Content

• Basic Usage
  • Evaluate
  • Statement
  • Parse
    • Evaluable Function
  • Simplify
• Working with Expressions
  • Evaluation Data Context
  • Operand Types
    • Value
    • Reference
    • Collection
  • Comparison Expressions
    • Equal
    • Not Equal
    • Greater Than
    • Greater Than or Equal
    • Less Than
    • Less Than or Equal
    • In
    • Not In
    • Prefix
    • Suffix
    • Overlap
    • Undefined
    • Present
  • Logical Expressions
    • And
    • Or
    • Nor
    • Xor
    • Not
  • Arithmetic Expressions
    • Divide
    • Multiply
    • Subtract
    • Sum
• Engine Options
  • Parser Options
    • Reference Predicate
    • Reference Transform
    • Operator Mapping
• Breaking Changes
  • v1.2.0
  • v1.4.2
• Contributing
  • Pull Request Process
• License

// Import the illogical engine
import Engine from '@briza/illogical'

// Create a new instance of the engine
const engine = new Engine()

// Evaluate the raw expression
const result = engine.evaluate(['==', 5, 5])

For advanced usage, please Engine Options.

Evaluate comparison or logical expression as TRUE or FALSE result:

engine.evaluate(Comparison Expression or Logical Expression, Evaluation Data Context) => boolean

Data context is optional.

Example

// Comparison expression
engine.evaluate(['==', 5, 5])
engine.evaluate(['==', 'circle', 'circle'])
engine.evaluate(['==', true, true])
engine.evaluate(['==', '$name', 'peter'], { name: 'peter' })
engine.evaluate(['UNDEFINED', '$RefA'], {})

// Logical expression
engine.evaluate(['AND', ['==', 5, 5], ['==', 10, 10]])
engine.evaluate(['AND', ['==', 'circle', 'circle'], ['==', 10, 10]])
engine.evaluate(['OR', ['==', '$name', 'peter'], ['==', 5, 10]], {
  name: 'peter',
})

Get expression string representation:

engine.statement(Comparison Expression or Logical Expression) => string

Example

/* Comparison expression */

engine.statement(['==', 5, 5])
// (5 == 5)

engine.statement(['==', 'circle', 'circle'])
// ("circle" == "circle")

engine.statement(['==', true, true])
// (true == true)

engine.statement(['==', '$name', 'peter'], { name: 'peter' })
// ({name} == "peter")

engine.statement(['UNDEFINED', '$RefA'])
// ({RefA} is UNDEFINED)

/* Logical expression */

engine.statement(['AND', ['==', 5, 5], ['==', 10, 10]])
// ((5 == 5) AND (10 == 10))

engine.statement(['AND', ['==', 'circle', 'circle'], ['==', 10, 10]])
// (("circle" == "circle") AND (10 == 10))

engine.statement(['OR', ['==', '$name', 'peter'], ['==', 5, 10]], {
  name: 'peter',
})
// (({name} == "peter") OR (5 == 10))

Parse the expression into a evaluable object, i.e. it returns the parsed self-evaluable condition expression.

engine.parse(Comparison Expression or Logical Expression) => evaluable

• evaluable.evaluate(context) please see Evaluation Data Context.
• evaluable.toString() please see Statement.

Example

let evaluable = engine.parse(['==', '$name', 'peter'])

evaluable.evaluate({ name: 'peter' }) // true

evaluable.toString()
// ({name} == "peter")

Simplifies an expression with a given context. This is useful when you already have some of the properties of context and wants to try to evaluate the expression.

Example

engine.simplify(['AND', ['==', '$a', 10], ['==', '$b', 20]], { a: 10 }) // ['==', '$b', 20]

engine.simplify(['AND', ['==', '$a', 10], ['==', '$b', 20]], { a: 20 }) // false

Values not found in the context will cause the parent operand not to be evaluated and returned as part of the simplified expression.

In some situations we might want to evaluate the expression even if referred value is not present. You can provide a list of keys that will be strictly evaluated even if they are not present in the context.

Example

engine.simplify(
  ['AND', ['==', '$a', 10], ['==', '$b', 20]],
  { a: 10 },
  ['b'] // '$b' will be evaluated to undefined.
) // false

Alternatively we might want to do the opposite and strictly evaluate the expression for all referred values not present in the context except for a specified list of optional keys.

Example

engine.simplify(
  ['OR', ['==', '$a', 10], ['==', '$b', 20], ['==', '$c', 20]],
  { c: 10 },
  undefined,
  ['b'] // except for '$b' everything not in context will be evaluated to undefined.
) // ['==', '$b', 20]

The evaluation data context is used to provide the expression with variable references, i.e. this allows for the dynamic expressions. The data context is object with properties used as the references keys, and its values as reference values.

Valid reference values: object, string, number, boolean, string[], number[].

To reference the nested reference, please use "." delimiter, e.g.: $address.city

If the key of the nested reference includes the "." delimiter, please wrap the whole key with backticks `, e.g.: $address.`city.code` can reference the object

{
  address: {
    'city.code': 'TOR'
  }
}

$address.`city.code`[0] can reference the object

{
  address: {
    'city.code': ['TOR']
  }
}

when the value of the nested reference is an array.

$options[1]

$options[{index}]

• The index reference is resolved within the data context as an array index.

$address.{segment}

• The segment reference is resolved within the data context as a property key.

$shape{shapeType}

• The shapeType reference is resolved within the data context, and inserted into the outer reference key.
• E.g. shapeType is resolved as "B" and would compose the $shapeB outer reference.
• This resolution could be n-nested.

$payment.amount.(Type)

Cast the given data context into the desired data type before being used as an operand in the evaluation.

Note: If the conversion is invalid, then a warning message is being logged.

Supported data type conversions:

• .(String): cast a given reference to String.
• .(Number): cast a given reference to Number.

Example

// Data context
const ctx = {
  name: 'peter',
  country: 'canada',
  age: 21,
  options: [1, 2, 3],
  address: {
    city: 'Toronto',
    country: 'Canada',
  },
  index: 2,
  segment: 'city',
  shapeA: 'box',
  shapeB: 'circle',
  shapeType: 'B',
}

// Evaluate an expression in the given data context
engine.evaluate(['>', '$age', 20], ctx) // true

// Evaluate an expression in the given data context
engine.evaluate(['==', '$address.city', 'Toronto'], ctx) // true

// Accessing Array Element
engine.evaluate(['==', '$options[1]', 2], ctx) // true

// Accessing Array Element via Reference
engine.evaluate(['==', '$options[{index}]', 3], ctx) // true

// Nested Referencing
engine.evaluate(['==', '$address.{segment}', 'Toronto'], ctx) // true

// Composite Reference Key
engine.evaluate(['==', '$shape{shapeType}', 'circle'], ctx) // true

// Data Type Casting
engine.evaluate(['==', '$age.(String)', '21'], ctx) // true

The Comparison Expression expect operands to be one of the below:

Simple value types: string, number, boolean.

Example

;['==', 5, 5][('==', 'circle', 'circle')][('==', true, true)]

The reference operand value is resolved from the Evaluation Data Context, where the the operands name is used as key in the context.

The reference operand must be prefixed with $ symbol, e.g.: $name. This might be customized via Reference Predicate Parser Option.

Example

The operand could be an array mixed from Value and Reference.

Example

Expression format: ["==", Left Operand, Right Operand].

Valid operand types: string, number, boolean.

["==", 5, 5]

engine.evaluate(['==', 5, 5]) // true

Expression format: ["!=", Left Operand, Right Operand].

Valid operand types: string, number, boolean.

["!=", "circle", "square"]

engine.evaluate(['!=', 'circle', 'square']) // true

Expression format: [">", Left Operand, Right Operand].

Valid operand types: number, string.

• String comparison only supports ISO-8601 formatted dates.

[">", 10, 5]
[">", "2023-01-01", "2022-12-31"]

engine.evaluate(['>', 10, 5]) // true
engine.evaluate(['>', '2023-01-01', '2022-12-31']) // true

Expression format: [">=", Left Operand, Right Operand].

Valid operand types: number, string.

• String comparison only supports ISO-8601 formatted dates.

[">=", 5, 5]
[">=", "2023-01-01",  "2023-01-01"]

engine.evaluate(['>=', 5, 5]) // true
engine.evaluate(['>=', '2023-01-01', '2023-01-01']) // true

Expression format: ["<", Left Operand, Right Operand].

Valid operand types: number, string.

• String comparison only supports ISO-8601 formatted dates.

["<", 5, 10]
["<", "2022-12-31",  "2023-01-01"]

engine.evaluate(['<', 5, 10]) // true
engine.evaluate(['<', '2022-12-31', '2023-01-01']) // true

Expression format: ["<=", Left Operand, Right Operand].

Valid operand types: number, string.

• String comparison only supports ISO-8601 formatted dates.

["<=", 5, 5]
["<=", "2023-01-01",  "2023-01-01"]

engine.evaluate(['<=', 5, 5]) // true
engine.evaluate(['<=', '2023-01-01', '2023-01-01']) // true

Expression format: ["IN", Left Operand, Right Operand].

Valid operand types: number and number[] or string and string[].

["IN", 5, [1,2,3,4,5]]
["IN", ["circle", "square", "triangle"], "square"]

engine.evaluate(['IN', 5, [1, 2, 3, 4, 5]]) // true
engine.evaluate(['IN', ['circle', 'square', 'triangle'], 'square']) // true

Expression format: ["NOT IN", Left Operand, Right Operand].

Valid operand types: number and number[] or string and string[].

["IN", 10, [1,2,3,4,5]]
["IN", ["circle", "square", "triangle"], "oval"]

engine.evaluate(['NOT IN', 10, [1, 2, 3, 4, 5]]) // true
engine.evaluate(['NOT IN', ['circle', 'square', 'triangle'], 'oval']) // true

Expression format: ["PREFIX", Left Operand, Right Operand].

Valid operand types: string.

• Left operand is the PREFIX term.
• Right operand is the tested word.

["PREFIX", "hemi", "hemisphere"]

engine.evaluate(['PREFIX', 'hemi', 'hemisphere']) // true
engine.evaluate(['PREFIX', 'hemi', 'sphere']) // false

Expression format: ["SUFFIX", Left Operand, Right Operand].

Valid operand types: string.

• Left operand is the tested word.
• Right operand is the SUFFIX term.

["SUFFIX", "establishment", "ment"]

engine.evaluate(['SUFFIX', 'establishment', 'ment']) // true
engine.evaluate(['SUFFIX', 'establish', 'ment']) // false

Expression format: ["OVERLAP", Left Operand, Right Operand].

Valid operand types number[] or string[].

["OVERLAP", [1, 2], [1, 2, 3, 4, 5]]
["OVERLAP", ["circle", "square", "triangle"], ["square"]]

engine.evaluate(['OVERLAP', [1, 2, 6], [1, 2, 3, 4, 5]]) // true
engine.evaluate([
  'OVERLAP',
  ['circle', 'square', 'triangle'],
  ['square', 'oval'],
]) // true

Expression format: ["UNDEFINED", Reference Operand].

["UNDEFINED", "$RefA"]

engine.evaluate(['UNDEFINED', 'RefA'], {}) // true
engine.evaluate(['UNDEFINED', 'RefA'], { RefA: undefined }) // true
engine.evaluate(['UNDEFINED', 'RefA'], { RefA: 10 }) // false

Evaluates as FALSE when the operand is UNDEFINED or NULL.

Expression format: ["PRESENT", Reference Operand].

["PRESENT", "$RefA"]

engine.evaluate(['PRESENT', 'RefA'], {}) // false
engine.evaluate(['PRESENT', 'RefA'], { RefA: undefined }) // false
engine.evaluate(['PRESENT', 'RefA'], { RefA: null }) // false
engine.evaluate(['PRESENT', 'RefA'], { RefA: 10 }) // true
engine.evaluate(['PRESENT', 'RefA'], { RefA: false }) // true
engine.evaluate(['PRESENT', 'RefA'], { RefA: 0 }) // true

The logical AND operator (&&) returns the boolean value TRUE if both operands are TRUE and returns FALSE otherwise.

Expression format: ["AND", Left Operand 1, Right Operand 2, ... , Right Operand N].

Valid operand types: Comparison Expression or Nested Logical Expression.

["AND", ["==", 5, 5], ["==", 10, 10]]

engine.evaluate(['AND', ['==', 5, 5], ['==', 10, 10]]) // true

The logical OR operator (||) returns the boolean value TRUE if either or both operands is TRUE and returns FALSE otherwise.

Expression format: ["OR", Left Operand 1, Right Operand 2, ... , Right Operand N].

Valid operand types: Comparison Expression or Nested Logical Expression.

["OR", ["==", 5, 5], ["==", 10, 5]]

engine.evaluate(['OR', ['==', 5, 5], ['==', 10, 5]]) // true

The logical NOR operator returns the boolean value TRUE if both operands are FALSE and returns FALSE otherwise.

Expression format: ["NOR", Left Operand 1, Right Operand 2, ... , Right Operand N]

Valid operand types: Comparison Expression or Nested Logical Expression.

["NOR", ["==", 5, 1], ["==", 10, 5]]

engine.evaluate(['NOR', ['==', 5, 1], ['==', 10, 5]]) // true

The logical NOR operator returns the boolean value TRUE if both operands are FALSE and returns FALSE otherwise.

Expression format: ["XOR", Left Operand 1, Right Operand 2, ... , Right Operand N]

Valid operand types: Comparison Expression or Nested Logical Expression.

["XOR", ["==", 5, 5], ["==", 10, 5]]

engine.evaluate(['XOR', ['==', 5, 5], ['==', 10, 5]]) // true

["XOR", ["==", 5, 5], ["==", 10, 10]]

engine.evaluate(['XOR', ['==', 5, 5], ['==', 10, 10]]) // false

The logical NOT operator returns the boolean value TRUE if the operand is FALSE, TRUE otherwise.

Expression format: ["NOT", Operand]

Valid operand types: Comparison Expression or Nested Logical Expression.

["NOT", ["==", 5, 5]]

engine.evaluate(['NOT', ['==', 5, 5]]) // true

Arithmetic Expressions are not supported as root level expressions since they must evaluate to a boolean. But it can be used nested within Comparisson Expressions.

The arithmetical operator for division produces the quotient of its operands where the left-most operand is the dividend and the subsequent one is the divisor, done from left to right.

Expression format: ["/", First Operand, Second Operand, ... , Nth Operand].

Valid operand types: Arithmetic Expressions or Operands.

["==", ["/", 100, 10], 10]

engine.evaluate(["==", ["/", 100, 10], 10]) // true

The arithmetical operator for multiplication produces the product of the operands.

Expression format: ["*", First Operand, Second Operand, ... , Nth Operand].

Valid operand types: Arithmetic Expressions or Operands.

["==", ["*", 100, 10], 10]

engine.evaluate(["==", ["*", 10, 10], 100]) // true

The arithmetical operator for subtraction subtracts the operands, producing their difference.

Expression format: ["-", First Operand, Second Operand, ... , Nth Operand].

Valid operand types: Arithmetic Expressions or Operands.

["==", ["-", 20, 10], 10]

engine.evaluate(["==", ["-", 20, 10], 10]) // true

The arithmetical operator for addition produces the sum of the operands.

Expression format: ["+", First Operand, Second Operand, ... , Nth Operand].

Valid operand types: Arithmetic Expressions or Operands.

["==", ["+", 5, 5], 10]

engine.evaluate(["==", ["+", 5, 5], 10]) // true

Below described, are individual options object properties which could be used individually. Any missing options will be substituted with the default options.

Usage

// Import the illogical engine
import Engine from '@briza/illogical'

// Create a new instance of the engine
const opts = {
  referencePredicate: (operand) => operand.startsWith('$'),
}
const engine = new Engine(opts)

A function used to determine if the operand is a reference type, otherwise evaluated as a static value.

referencePredicate: (operand: string) => boolean

Return value:

• true = reference type
• false = value type

Default reference predicate:

The $ symbol at the begging of the operand is used to predicate the reference type., E.g. $State, $Country.

A function used to transform the operand into the reference annotation stripped form. I.e. remove any annotation used to detect the reference type. E.g. "$Reference" => "Reference".

referenceTransform: (operand: string) => string

Default reference transform: It removes the $ symbol at the begging of the operand name.

Mapping of the operators. The key is unique operator key, and the value is the key used to represent the given operator in the raw expression.

operatorMapping: Map<symbol, string>

Default operator mapping:

  // Comparison
  [OPERATOR_EQ, '=='],
  [OPERATOR_NE, '!='],
  [OPERATOR_GT, '>'],
  [OPERATOR_GE, '>='],
  [OPERATOR_LT, '<'],
  [OPERATOR_LE, '<='],
  [OPERATOR_IN, 'IN'],
  [OPERATOR_NOT_IN, 'NOT IN'],
  [OPERATOR_PREFIX, 'PREFIX'],
  [OPERATOR_SUFFIX, 'SUFFIX'],
  [OPERATOR_OVERLAP, 'OVERLAP'],
  [OPERATOR_UNDEFINED, 'UNDEFINED'],
  [OPERATOR_PRESENT, 'PRESENT'],
  // Logical
  [OPERATOR_AND, 'AND'],
  [OPERATOR_OR, 'OR'],
  [OPERATOR_NOR, 'NOR'],
  [OPERATOR_XOR, 'XOR'],
  [OPERATOR_NOT, 'NOT'],
  // Arithmetic
  [OPERATOR_SUM, '+'],
  [OPERATOR_SUBTRACT, '-'],
  [OPERATOR_MULTIPLY, '*'],
  [OPERATOR_DIVIDE, '/'],

The operator keys are unique symbols which could be imported from the engine package:

import {
  OPERATOR_EQ,
  OPERATOR_NE,
  OPERATOR_GT,
  OPERATOR_GE,
  OPERATOR_LT,
  OPERATOR_LE,
  OPERATOR_IN,
  OPERATOR_NOT_IN,
  OPERATOR_PREFIX,
  OPERATOR_SUFFIX,
  OPERATOR_OVERLAP,
  OPERATOR_UNDEFINED,
  OPERATOR_PRESENT,
  OPERATOR_AND,
  OPERATOR_OR,
  OPERATOR_NOR,
  OPERATOR_XOR,
  OPERATOR_NOT,
  OPERATOR_DIVIDE,
  OPERATOR_MULTIPLY,
  OPERATOR_SUBTRACT,
  OPERATOR_SUM,
} from '@briza/illogical'

• Change on @babel/env preset to target > 1%, node 12 this will remove some polyfills that were causing performance problems in some projects.

• Removed strict mode from the Engine constructor options. const engine = new Engine(strictMode, opts); -> const engine = new Engine(opts);

See contributing.md.

Illogical is released under the MIT license. See license.txt.