Design a JSON Schema Validator

Problem Statement

In document processing and API integration platforms like DocuSign, incoming payloads must be strictly validated against predefined contracts before execution.

Implement a JSONSchemaValidator class that validates arbitrary JSON-like data against a simplified JSON schema specification. The validator must return whether the data is valid and a list of human-readable error descriptions if it is not.

Supported Schema Keywords

1. type: Matches target data type: "string", "number" (accepts float or int), "boolean", "object", or "array".
2. Numeric constraints (for type: "number"):
   • minimum: value must be >= minimum.
   • maximum: value must be <= maximum.
3. String constraints (for type: "string"):
   • pattern: a regular expression string that the value must match (via full or partial match per regex definition).
4. Object constraints (for type: "object"):
   • properties: a dictionary mapping property names to sub-schemas.
   • required: a list of property names that must be present.
   • additionalProperties: boolean. If False, no fields outside of properties are allowed (defaults to True).
5. Array constraints (for type: "array"):
   • items: a sub-schema that every element in the array must satisfy.

Class Interface

class JSONSchemaValidator:
    def validate(self, schema: dict, data: any) -> tuple[bool, list[str]]:
        """
        Validates `data` against `schema`.
        Returns (True, []) if valid, or (False, [error_messages]) if invalid.
        """
        pass

Example 1

schema = {
    "type": "object",
    "required": ["name", "age"],
    "additionalProperties": False,
    "properties": {
        "name": {"type": "string", "pattern": r"^[A-Za-z ]+$"},
        "age": {"type": "number", "minimum": 18, "maximum": 120}
    }
}

data = {"name": "Alice", "age": 25}
validator = JSONSchemaValidator()
is_valid, errors = validator.validate(schema, data)
# is_valid: True, errors: []

Example 2

data_invalid = {"name": "Alice", "age": 15, "extra": True}
is_valid, errors = validator.validate(schema, data_invalid)
# is_valid: False
# errors: ["age must be >= 18", "unexpected property: extra"]

Constraints

• The provided schema is guaranteed to be syntactically valid.
• Nesting depth of objects/arrays will not exceed 20 levels.
• data may contain None, bool, int, float, str, list, or dict.
• Note: in Python, isinstance(True, int) evaluates to True. Make sure boolean values are not mistakenly identified as numbers.