ksc-sdk-python/kscore/validate.py

"""User input parameter validation.

This module handles user input parameter validation
against a provided input model.

Note that the objects in this module do *not* mutate any
arguments.  No type version happens here.  It is up to another
layer to properly convert arguments to any required types.

Validation Errors
-----------------


"""

from kscore.compat import six
import decimal
from datetime import datetime

from kscore.utils import parse_to_aware_datetime
from kscore.exceptions import ParamValidationError


def validate_parameters(params, shape):
    """Validates input parameters against a schema.

    This is a convenience function that validates parameters against a schema.
    You can also instantiate and use the ParamValidator class directly if you
    want more control.

    If there are any validation errors then a ParamValidationError
    will be raised.  If there are no validation errors than no exception
    is raised and a value of None is returned.

    :param params: The user provided input parameters.

    :type shape: kscore.model.Shape
    :param shape: The schema which the input parameters should
        adhere to.

    :raise: ParamValidationError

    """
    validator = ParamValidator()
    report = validator.validate(params, shape)
    if report.has_errors():
        raise ParamValidationError(report=report.generate_report())


def type_check(valid_types):
    def _create_type_check_guard(func):
        def _on_passes_type_check(self, param, shape, errors, name):
            if _type_check(param, errors, name):
                return func(self, param, shape, errors, name)

        def _type_check(param, errors, name):
            if not isinstance(param, valid_types):
                valid_type_names = [six.text_type(t) for t in valid_types]
                errors.report(name, 'invalid type', param=param,
                              valid_types=valid_type_names)
                return False
            return True

        return _on_passes_type_check
    return _create_type_check_guard


def range_check(name, value, shape, error_type, errors):
    failed = False
    min_allowed = float('-inf')
    max_allowed = float('inf')
    if 'min' in shape.metadata:
        min_allowed = shape.metadata['min']
        if value < min_allowed:
            failed = True
    if failed:
        errors.report(name, error_type, param=value,
                      valid_range=[min_allowed, max_allowed])



class ValidationErrors(object):
    def __init__(self):
        self._errors = []

    def has_errors(self):
        if self._errors:
            return True
        return False

    def generate_report(self):
        error_messages = []
        for error in self._errors:
            error_messages.append(self._format_error(error))
        return '\n'.join(error_messages)

    def _format_error(self, error):
        error_type, name, additional = error
        name = self._get_name(name)
        if error_type == 'missing required field':
            return 'Missing required parameter in %s: "%s"' % (
                name, additional['required_name'])
        elif error_type == 'unknown field':
            return 'Unknown parameter in %s: "%s", must be one of: %s' % (
                name, additional['unknown_param'], ', '.join(additional['valid_names']))
        elif error_type == 'invalid type':
            return 'Invalid type for parameter %s, value: %s, type: %s, valid types: %s' % (
                name, additional['param'],
                str(type(additional['param'])),
                ', '.join(additional['valid_types']))
        elif error_type == 'invalid range':
            min_allowed = additional['valid_range'][0]
            max_allowed = additional['valid_range'][1]
            return ('Invalid range for parameter %s, value: %s, valid range: '
                    '%s-%s' % (name, additional['param'],
                               min_allowed, max_allowed))
        elif error_type == 'invalid length':
            min_allowed = additional['valid_range'][0]
            max_allowed = additional['valid_range'][1]
            return ('Invalid length for parameter %s, value: %s, valid range: '
                    '%s-%s' % (name, additional['param'],
                               min_allowed, max_allowed))

    def _get_name(self, name):
        if not name:
            return 'input'
        elif name.startswith('.'):
            return name[1:]
        else:
            return name

    def report(self, name, reason, **kwargs):
        self._errors.append((reason, name, kwargs))


class ParamValidator(object):
    """Validates parameters against a shape model."""

    def validate(self, params, shape):
        """Validate parameters against a shape model.

        This method will validate the parameters against a provided shape model.
        All errors will be collected before returning to the caller.  This means
        that this method will not stop at the first error, it will return all
        possible errors.

        :param params: User provided dict of parameters
        :param shape: A shape model describing the expected input.

        :return: A list of errors.

        """
        errors = ValidationErrors()
        self._validate(params, shape, errors, name='')
        return errors

    def _validate(self, params, shape, errors, name):
        getattr(self, '_validate_%s' % shape.type_name)(params, shape, errors, name)

    @type_check(valid_types=(dict,))
    def _validate_structure(self, params, shape, errors, name):
        # Validate required fields.
        for required_member in shape.metadata.get('required', []):
            if required_member not in params:
                errors.report(name, 'missing required field',
                              required_name=required_member, user_params=params)
        members = shape.members
        known_params = []
        # Validate known params.
        for param in params:
            if param not in members:
                errors.report(name, 'unknown field', unknown_param=param,
                              valid_names=list(members))
            else:
                known_params.append(param)
        # Validate structure members.
        for param in known_params:
            self._validate(params[param], shape.members[param],
                           errors, '%s.%s' % (name, param))

    @type_check(valid_types=six.string_types)
    def _validate_string(self, param, shape, errors, name):
        # Validate range.  For a string, the min/max contraints
        # are of the string length.
        # Looks like:
        # "WorkflowId":{
        #   "type":"string",
        #   "min":1,
        #   "max":256
        #  }
        range_check(name, len(param), shape, 'invalid length', errors)

    @type_check(valid_types=(list, tuple))
    def _validate_list(self, param, shape, errors, name):
        member_shape = shape.member
        range_check(name, len(param), shape, 'invalid length', errors)
        for i, item in enumerate(param):
            self._validate(item, member_shape, errors, '%s[%s]' % (name, i))

    @type_check(valid_types=(dict,))
    def _validate_map(self, param, shape, errors, name):
        key_shape = shape.key
        value_shape = shape.value
        for key, value in param.items():
            self._validate(key, key_shape, errors, "%s (key: %s)"
                           % (name, key))
            self._validate(value, value_shape, errors, '%s.%s' % (name, key))

    @type_check(valid_types=six.integer_types)
    def _validate_integer(self, param, shape, errors, name):
        range_check(name, param, shape, 'invalid range', errors)

    def _validate_blob(self, param, shape, errors, name):
        if isinstance(param, (bytes, bytearray, six.text_type)):
            return
        elif hasattr(param, 'read'):
            # File like objects are also allowed for blob types.
            return
        else:
            errors.report(name, 'invalid type', param=param,
                         valid_types=[str(bytes), str(bytearray),
                                      'file-like object'])

    @type_check(valid_types=(bool,))
    def _validate_boolean(self, param, shape, errors, name):
        pass

    @type_check(valid_types=(float, decimal.Decimal) + six.integer_types)
    def _validate_double(self, param, shape, errors, name):
        range_check(name, param, shape, 'invalid range', errors)

    _validate_float = _validate_double

    @type_check(valid_types=six.integer_types)
    def _validate_long(self, param, shape, errors, name):
        range_check(name, param, shape, 'invalid range', errors)

    def _validate_timestamp(self, param, shape, errors, name):
        # We don't use @type_check because datetimes are a bit
        # more flexible.  You can either provide a datetime
        # object, or a string that parses to a datetime.
        is_valid_type = self._type_check_datetime(param)
        if not is_valid_type:
            valid_type_names = [six.text_type(datetime), 'timestamp-string']
            errors.report(name, 'invalid type', param=param,
                            valid_types=valid_type_names)

    def _type_check_datetime(self, value):
        try:
            parse_to_aware_datetime(value)
            return True
        except (TypeError, ValueError, AttributeError):
            # Yes, dateutil can sometimes raise an AttributeError
            # when parsing timestamps.
            return False


class ParamValidationDecorator(object):
    def __init__(self, param_validator, serializer):
        self._param_validator = param_validator
        self._serializer = serializer

    def serialize_to_request(self, parameters, operation_model):
        input_shape = operation_model.input_shape
        if input_shape is not None:
            report = self._param_validator.validate(parameters,
                                                    operation_model.input_shape)
            if report.has_errors():
                raise ParamValidationError(report=report.generate_report())
        return self._serializer.serialize_to_request(parameters,
                                                     operation_model)