Analyze

A predicate expression evaluator designed to embed logical expressions in YAML or JSON. Think of being able to apply a SQL where clause on a set of data that you supply. For example, let say we have an array of objects.

[
  {
    "firstName": "John", 
    "lastName": "Smith"
  },
  {
    "firstName": "Mary",
    "lastName": "Jones"
  }
]

You could then apply the following expression to see if there were someone with the first name John in your collection of objects.

@len( @select(data, "$elt.firstName == 'John'")) > 0

This becomes useful when you want to build tools that perform analysis defined in a configuration file, so you might collect data on resources in kubernetes and check to see if a particular value is defined in a ConfigMap. The Go code to perform this looks like this. In this case we'd put the results of our collection in a map[string]interface{} rooted at 'resources' and 'config_maps' would refer to an array of config map resources.

ctx, err := context.New(data)
if err != nil {
	log.Fatal(err)
}
result, err := expression.EvaluateContext(ctx, `@len( @select( $resources.config_maps, "$elt.firstName == 'John'" ) ) > 0`)
if result {
	fmt.Println("found John!")
}

Expressions are composed of operators needed to compose a predicate expression and a way to declare variables that will reference a supplied data set as well as a set of builtin functions. If the existing functionality doesn't support your use case it's easy to define your own functions.

func MyFunction(args  []interface{})(interface{}, error) {
	// your implementation 
}

ctx, _ := context.New(data, context.Func("@myfunc", MyFunction))
result, _ := expression.EvaluateContext(ctx, `@myfunc($foo) == "hello"`)

Examples

Programs illustrating the usage of Analyze can be found in the examples directory. Also see the unit tests in the analyzer/expression package for more examples of expressions and how they are used.