Introduction

A simple rules engine with great flexibility capable of building binary decision trees of any depth.

Utilizes gjson for pathing for extra flexibility.

Inspired by grules

Simple Example (pass)

json := `
{
    "name": {"first": "anakin", "last": "skywalker"},
    "age": 22,
    "children": ["luke", "leia"],
    "order": "jedi",
    "friends": [
        {"first": "r2d2",  "last": "droid",      "order": "republic", "age": 13, "episodes": [1,2,3,4,5,6,7,8,9]},
        {"first": "ben",   "last": "kenobi",     "order": "jedi",     "age": 38, "episodes": [1,2,3,4,5,6]},
        {"first": "c3po",  "last": "droid",      "order": "republic", "age": 13, "episodes": [1,2,3,4,5,6,7,8,9]},
        {"first": "sheev", "last": "palpatine",  "order": "sith",     "age": 63, "episodes": [1,2,3,5,6,9]}
    ]
}
`

rule := `
{
    "comparer": "eq",
    "path": "name.first",
    "value": "anakin"
}
`

pass, failReason := grules.Evaluate(json, rule)
if !pass {
    fmt.Println("FAILED: ", failreason)
}

fmt.Println(pass)

Output: true

Simple Example (fail)

json := `
{
    "name": {"first": "anakin", "last": "skywalker"},
    "age": 22,
    "children": ["luke", "leia"],
    "order": "jedi",
    "friends": [
        {"first": "r2d2",  "last": "droid",      "order": "republic", "age": 13, "episodes": [1,2,3,4,5,6,7,8,9]},
        {"first": "ben",   "last": "kenobi",     "order": "jedi",     "age": 38, "episodes": [1,2,3,4,5,6]},
        {"first": "c3po",  "last": "droid",      "order": "republic", "age": 13, "episodes": [1,2,3,4,5,6,7,8,9]},
        {"first": "sheev", "last": "palpatine",  "order": "sith",     "age": 63, "episodes": [1,2,3,5,6,9]}
    ]
}
`

rule := `
{
    "comparer": "lt",
    "path": "age",
    "value": "20"
}
`

passed, failReason := grules.Evaluate(json, rule)
if !passed {
    fmt.Println("FAILED: ", failreason)
}

fmt.Println(passed)

Output: FAILED: value '22' at 'age' is not 'less than' rule value '20'

Complicated Example (pass)

json := `
{
    "name": {"first": "anakin", "last": "skywalker"},
    "age": 22,
    "children": ["luke", "leia"],
    "order": "jedi",
    "friends": [
        {"first": "r2d2",  "last": "droid",      "order": "republic", "age": 13, "episodes": [1,2,3,4,5,6,7,8,9]},
        {"first": "ben",   "last": "kenobi",     "order": "jedi",     "age": 38, "episodes": [1,2,3,4,5,6]},
        {"first": "c3po",  "last": "droid",      "order": "republic", "age": 13, "episodes": [1,2,3,4,5,6,7,8,9]},
        {"first": "sheev", "last": "palpatine",  "order": "sith",     "age": 63, "episodes": [1,2,3,5,6,9]}
    ]
}
`

rule := `
{
    "operator": "or",
    "rules": [
        {
            "operator": "and",
            "rules": [
                {
                    "path": "name.first",
                    "comparer": "eq",
                    "value": "darth"
                },
                {
                    "path": "name.last",
                    "comparer": "eq",
                    "value": "vader"
                }
            ]
        },
        {
            "operator": "or",
            "rules": [
                {
                    "path": "order",
                    "comparer": "eq",
                    "value": "first world order"
                },
                {
                    "operator": "or",
                    "path": "friends.#.order",
                    "comparer": "contains",
                    "value": "sith"
                }
            ]
        }
    ]
}
`

pass, failReason := grules.Evaluate(json, rule)
if !pass {
    fmt.Println("FAILED: ", failreason)
}

fmt.Println(pass)

Output: true

Comparators

• eq will return true if a == b
• neq will return true if a != b
• lt will return true if a < b
• lte will return true if a <= b
• gt will return true if a > b
• gte will return true if a >= b
• contains will return true if a contains b
• ncontains will return true if a does not contain b
• oneof will return true if a is one of b
• noneof will return true if a is not one of b
• regex will return true if a matches b

contains and ncontains work for substring comparisons as well as item-in-collection comparisons.

When used for item-in-collection comparisons, contains expects the first argument to be a slice. contains is different than oneof in that oneof expects the second argument to be a slice.

Benchmarks

Benchmark	N	Speed	Used	Allocs
BenchmarkEqual-12	650602549	5.52 ns/op	0 B/op	0 allocs/op
BenchmarkNotEqual-12	876894124	4.09 ns/op	0 B/op	0 allocs/op
BenchmarkLessThan-12	1000000000	2.84 ns/op	0 B/op	0 allocs/op
BenchmarkLessThanEqual-12	1000000000	2.57 ns/op	0 B/op	0 allocs/op
BenchmarkGreaterThan-12	1000000000	2.07 ns/op	0 B/op	0 allocs/op
BenchmarkGreaterThanEqual-12	1000000000	2.86 ns/op	0 B/op	0 allocs/op
BenchmarkRegex-12	4524237	793 ns/op	753 B/op	11 allocs/op
BenchmarkRegexPhone-12	1000000	3338 ns/op	3199 B/op	30 allocs/op
BenchmarkContains-12	499627219	7.16 ns/op	0 B/op	0 allocs/op
BenchmarkStringContains-12	405497102	8.87 ns/op	0 B/op	0 allocs/op
BenchmarkContainsLong50000-12	18992	184016 ns/op	0 B/op	0 allocs/op
BenchmarkNotContains-12	292932907	12.3 ns/op	0 B/op	0 allocs/op
BenchmarkStringNotContains-12	392618857	9.14 ns/op	0 B/op	0 allocs/op
BenchmarkNotContainsLong50000-12	19243	191787 ns/op	0 B/op	0 allocs/op
BenchmarkOneOf-12	1000000000	1.80 ns/op	0 B/op	0 allocs/op
BenchmarkNoneOf-12	1000000000	1.79 ns/op	0 B/op	0 allocs/op
BenchmarkPluckShallow-12	85997188	41.6 ns/op	16 B/op	1 allocs/op
BenchmarkPluckDeep-12	18789103	194 ns/op	112 B/op	1 allocs/op
BenchmarkRule_evaluate-12	69558996	51.1 ns/op	16 B/op	1 allocs/op
BenchmarkComposite_evaluate-12	59484760	55.7 ns/op	16 B/op	1 allocs/op
BenchmarkEngine_Evaluate-12	47892318	75.0 ns/op	16 B/op	1 allocs/op

To run benchmarks:

go test -run none -bench . -benchtime 3s -benchmem

All benchmarks were run on:

MacOS High Sierra 2.6Ghz Intel Core i7 16 GB 2400 MHz DDR4