JSON Schema Expression Validator

About

---

Expression Validator is an example application which demonstrates validation of condition expressions. It is a process which scans and validates expressions against JSON Schema in three phases.

• First phase is a lexical analysis - scanning through the list of characters and group them together into the smallest sequences that represent something. We called them a lexemes.

• In second phase we try to handle expression which can nest arbitrarily deeply. For that job to be done we must use Context-Free grammar which has its own set of rules. In particular, we're defining an abstract syntax tree (AST). In a parse tree, every single grammar production becomes a node in the tree. We parse a string - a series of tokens - then map those tokens to terminals in the grammar to figure out which rules could have generated that string.

  Used grammar:

  expression     → equality ;
  equality       → comparison ( ( "!=" | "==" ) comparison )* ;
  comparison     → primary ( ( ">" | ">=" | "<" | "<=" ) primary )* ;
  primary        → NUMBER | STRING | BOOLEAN | "nil" | "(" expression ")" ;
  

• The third phase is evaluating expression - we recursively interpret each node from previous parsed tree and return boolean value if expression is valid or not.

---

Notification pattern

Instead of throwing exceptions, validation errors are collected using Notification Pattern. Of course, if validation failed, you are able to fetch validation results using validator.getValidationResult() method which is a part of the expression validator.

---

Interpreter

Interpreter evaluates expression using the Visitor Pattern. More details here.

Lexer

Lexer performs the Lexical Analysis of expression.

Parser

Parser parses series of tokens - we map those tokens to terminals in the grammar to figure out could have generated that string.

---

JSON Schemas

Supported JSON Schemas:

• ObjectSchema
• IntegerSchema
• NumberSchema
• BooleanSchema
• ArraySchema
• StringSchema

---

Example JSON Schema

{
  "definition": {
    "properties": {
      "name": {
        "validation": {
          "required": true
        },
        "type": "STRING"
      },
      "age": {
        "validation": {
          "required": true
        },
        "type": "INTEGER"
      },
      "employed": {
        "type": "BOOLEAN",
        "validation": {
          "required": true
        }
      }
    },
    "type": "OBJECT"
  }
}

Example JSON schema can be found here.

Example expressions:

($.age >= 30) || ($.age < 40)
$.name != null
$.name != \"null\"
$.name == \"someCoolName\"
$.employed == false
(($.employed == false) && ($.employed != true))
$.employed != false && $.age > 60
($.age != 30) || ($.age < 40) && ($.employed == true)

More details in Expression Validator Test.

---

Installation

Run maven install command with mvn clean install

Run maven tests mvn test