django/django/core/validators.py

import ipaddress
import os
import re
from urllib.parse import urlsplit, urlunsplit

from django.core.exceptions import ValidationError
from django.utils.deconstruct import deconstructible
from django.utils.functional import SimpleLazyObject
from django.utils.ipv6 import is_valid_ipv6_address
from django.utils.translation import gettext_lazy as _, ngettext_lazy

# These values, if given to validate(), will trigger the self.required check.
EMPTY_VALUES = (None, '', [], (), {})


def _lazy_re_compile(regex, flags=0):
    """Lazily compile a regex with flags."""
    def _compile():
        # Compile the regex if it was not passed pre-compiled.
        if isinstance(regex, str):
            return re.compile(regex, flags)
        else:
            assert not flags, "flags must be empty if regex is passed pre-compiled"
            return regex
    return SimpleLazyObject(_compile)


@deconstructible
class RegexValidator:
    regex = ''
    message = _('Enter a valid value.')
    code = 'invalid'
    inverse_match = False
    flags = 0

    def __init__(self, regex=None, message=None, code=None, inverse_match=None, flags=None):
        if regex is not None:
            self.regex = regex
        if message is not None:
            self.message = message
        if code is not None:
            self.code = code
        if inverse_match is not None:
            self.inverse_match = inverse_match
        if flags is not None:
            self.flags = flags
        if self.flags and not isinstance(self.regex, str):
            raise TypeError("If the flags are set, regex must be a regular expression string.")

        self.regex = _lazy_re_compile(self.regex, self.flags)

    def __call__(self, value):
        """
        Validate that the input contains (or does *not* contain, if
        inverse_match is True) a match for the regular expression.
        """
        regex_matches = bool(self.regex.search(str(value)))
        invalid_input = regex_matches if self.inverse_match else not regex_matches
        if invalid_input:
            raise ValidationError(self.message, code=self.code)

    def __eq__(self, other):
        return (
            isinstance(other, RegexValidator) and
            self.regex.pattern == other.regex.pattern and
            self.regex.flags == other.regex.flags and
            (self.message == other.message) and
            (self.code == other.code) and
            (self.inverse_match == other.inverse_match)
        )


@deconstructible
class URLValidator(RegexValidator):
    ul = '\u00a1-\uffff'  # unicode letters range (must not be a raw string)

    # IP patterns
    ipv4_re = r'(?:25[0-5]|2[0-4]\d|[0-1]?\d?\d)(?:\.(?:25[0-5]|2[0-4]\d|[0-1]?\d?\d)){3}'
    ipv6_re = r'\[[0-9a-f:\.]+\]'  # (simple regex, validated later)

    # Host patterns
    hostname_re = r'[a-z' + ul + r'0-9](?:[a-z' + ul + r'0-9-]{0,61}[a-z' + ul + r'0-9])?'
    # Max length for domain name labels is 63 characters per RFC 1034 sec. 3.1
    domain_re = r'(?:\.(?!-)[a-z' + ul + r'0-9-]{1,63}(?<!-))*'
    tld_re = (
        r'\.'                                # dot
        r'(?!-)'                             # can't start with a dash
        r'(?:[a-z' + ul + '-]{2,63}'         # domain label
        r'|xn--[a-z0-9]{1,59})'              # or punycode label
        r'(?<!-)'                            # can't end with a dash
        r'\.?'                               # may have a trailing dot
    )
    host_re = '(' + hostname_re + domain_re + tld_re + '|localhost)'

    regex = _lazy_re_compile(
        r'^(?:[a-z0-9\.\-\+]*)://'  # scheme is validated separately
        r'(?:\S+(?::\S*)?@)?'  # user:pass authentication
        r'(?:' + ipv4_re + '|' + ipv6_re + '|' + host_re + ')'
        r'(?::\d{2,5})?'  # port
        r'(?:[/?#][^\s]*)?'  # resource path
        r'\Z', re.IGNORECASE)
    message = _('Enter a valid URL.')
    schemes = ['http', 'https', 'ftp', 'ftps']

    def __init__(self, schemes=None, **kwargs):
        super().__init__(**kwargs)
        if schemes is not None:
            self.schemes = schemes

    def __call__(self, value):
        # Check first if the scheme is valid
        scheme = value.split('://')[0].lower()
        if scheme not in self.schemes:
            raise ValidationError(self.message, code=self.code)

        # Then check full URL
        try:
            super().__call__(value)
        except ValidationError as e:
            # Trivial case failed. Try for possible IDN domain
            if value:
                try:
                    scheme, netloc, path, query, fragment = urlsplit(value)
                except ValueError:  # for example, "Invalid IPv6 URL"
                    raise ValidationError(self.message, code=self.code)
                try:
                    netloc = netloc.encode('idna').decode('ascii')  # IDN -> ACE
                except UnicodeError:  # invalid domain part
                    raise e
                url = urlunsplit((scheme, netloc, path, query, fragment))
                super().__call__(url)
            else:
                raise
        else:
            # Now verify IPv6 in the netloc part
            host_match = re.search(r'^\[(.+)\](?::\d{2,5})?$', urlsplit(value).netloc)
            if host_match:
                potential_ip = host_match.groups()[0]
                try:
                    validate_ipv6_address(potential_ip)
                except ValidationError:
                    raise ValidationError(self.message, code=self.code)

        # The maximum length of a full host name is 253 characters per RFC 1034
        # section 3.1. It's defined to be 255 bytes or less, but this includes
        # one byte for the length of the name and one byte for the trailing dot
        # that's used to indicate absolute names in DNS.
        if len(urlsplit(value).netloc) > 253:
            raise ValidationError(self.message, code=self.code)


integer_validator = RegexValidator(
    _lazy_re_compile(r'^-?\d+\Z'),
    message=_('Enter a valid integer.'),
    code='invalid',
)


def validate_integer(value):
    return integer_validator(value)


@deconstructible
class EmailValidator:
    message = _('Enter a valid email address.')
    code = 'invalid'
    user_regex = _lazy_re_compile(
        r"(^[-!#$%&'*+/=?^_`{}|~0-9A-Z]+(\.[-!#$%&'*+/=?^_`{}|~0-9A-Z]+)*\Z"  # dot-atom
        r'|^"([\001-\010\013\014\016-\037!#-\[\]-\177]|\\[\001-\011\013\014\016-\177])*"\Z)',  # quoted-string
        re.IGNORECASE)
    domain_regex = _lazy_re_compile(
        # max length for domain name labels is 63 characters per RFC 1034
        r'((?:[A-Z0-9](?:[A-Z0-9-]{0,61}[A-Z0-9])?\.)+)(?:[A-Z0-9-]{2,63}(?<!-))\Z',
        re.IGNORECASE)
    literal_regex = _lazy_re_compile(
        # literal form, ipv4 or ipv6 address (SMTP 4.1.3)
        r'\[([A-f0-9:\.]+)\]\Z',
        re.IGNORECASE)
    domain_whitelist = ['localhost']

    def __init__(self, message=None, code=None, whitelist=None):
        if message is not None:
            self.message = message
        if code is not None:
            self.code = code
        if whitelist is not None:
            self.domain_whitelist = whitelist

    def __call__(self, value):
        if not value or '@' not in value:
            raise ValidationError(self.message, code=self.code)

        user_part, domain_part = value.rsplit('@', 1)

        if not self.user_regex.match(user_part):
            raise ValidationError(self.message, code=self.code)

        if (domain_part not in self.domain_whitelist and
                not self.validate_domain_part(domain_part)):
            # Try for possible IDN domain-part
            try:
                domain_part = domain_part.encode('idna').decode('ascii')
            except UnicodeError:
                pass
            else:
                if self.validate_domain_part(domain_part):
                    return
            raise ValidationError(self.message, code=self.code)

    def validate_domain_part(self, domain_part):
        if self.domain_regex.match(domain_part):
            return True

        literal_match = self.literal_regex.match(domain_part)
        if literal_match:
            ip_address = literal_match.group(1)
            try:
                validate_ipv46_address(ip_address)
                return True
            except ValidationError:
                pass
        return False

    def __eq__(self, other):
        return (
            isinstance(other, EmailValidator) and
            (self.domain_whitelist == other.domain_whitelist) and
            (self.message == other.message) and
            (self.code == other.code)
        )


validate_email = EmailValidator()

slug_re = _lazy_re_compile(r'^[-a-zA-Z0-9_]+\Z')
validate_slug = RegexValidator(
    slug_re,
    # Translators: "letters" means latin letters: a-z and A-Z.
    _("Enter a valid 'slug' consisting of letters, numbers, underscores or hyphens."),
    'invalid'
)

slug_unicode_re = _lazy_re_compile(r'^[-\w]+\Z')
validate_unicode_slug = RegexValidator(
    slug_unicode_re,
    _("Enter a valid 'slug' consisting of Unicode letters, numbers, underscores, or hyphens."),
    'invalid'
)


def validate_ipv4_address(value):
    try:
        ipaddress.IPv4Address(value)
    except ValueError:
        raise ValidationError(_('Enter a valid IPv4 address.'), code='invalid')


def validate_ipv6_address(value):
    if not is_valid_ipv6_address(value):
        raise ValidationError(_('Enter a valid IPv6 address.'), code='invalid')


def validate_ipv46_address(value):
    try:
        validate_ipv4_address(value)
    except ValidationError:
        try:
            validate_ipv6_address(value)
        except ValidationError:
            raise ValidationError(_('Enter a valid IPv4 or IPv6 address.'), code='invalid')


ip_address_validator_map = {
    'both': ([validate_ipv46_address], _('Enter a valid IPv4 or IPv6 address.')),
    'ipv4': ([validate_ipv4_address], _('Enter a valid IPv4 address.')),
    'ipv6': ([validate_ipv6_address], _('Enter a valid IPv6 address.')),
}


def ip_address_validators(protocol, unpack_ipv4):
    """
    Depending on the given parameters, return the appropriate validators for
    the GenericIPAddressField.
    """
    if protocol != 'both' and unpack_ipv4:
        raise ValueError(
            "You can only use `unpack_ipv4` if `protocol` is set to 'both'")
    try:
        return ip_address_validator_map[protocol.lower()]
    except KeyError:
        raise ValueError("The protocol '%s' is unknown. Supported: %s"
                         % (protocol, list(ip_address_validator_map)))


def int_list_validator(sep=',', message=None, code='invalid', allow_negative=False):
    regexp = _lazy_re_compile(r'^%(neg)s\d+(?:%(sep)s%(neg)s\d+)*\Z' % {
        'neg': '(-)?' if allow_negative else '',
        'sep': re.escape(sep),
    })
    return RegexValidator(regexp, message=message, code=code)


validate_comma_separated_integer_list = int_list_validator(
    message=_('Enter only digits separated by commas.'),
)


@deconstructible
class BaseValidator:
    message = _('Ensure this value is %(limit_value)s (it is %(show_value)s).')
    code = 'limit_value'

    def __init__(self, limit_value, message=None):
        self.limit_value = limit_value
        if message:
            self.message = message

    def __call__(self, value):
        cleaned = self.clean(value)
        params = {'limit_value': self.limit_value, 'show_value': cleaned, 'value': value}
        if self.compare(cleaned, self.limit_value):
            raise ValidationError(self.message, code=self.code, params=params)

    def __eq__(self, other):
        return (
            isinstance(other, self.__class__) and
            self.limit_value == other.limit_value and
            self.message == other.message and
            self.code == other.code
        )

    def compare(self, a, b):
        return a is not b

    def clean(self, x):
        return x


@deconstructible
class MaxValueValidator(BaseValidator):
    message = _('Ensure this value is less than or equal to %(limit_value)s.')
    code = 'max_value'

    def compare(self, a, b):
        return a > b


@deconstructible
class MinValueValidator(BaseValidator):
    message = _('Ensure this value is greater than or equal to %(limit_value)s.')
    code = 'min_value'

    def compare(self, a, b):
        return a < b


@deconstructible
class MinLengthValidator(BaseValidator):
    message = ngettext_lazy(
        'Ensure this value has at least %(limit_value)d character (it has %(show_value)d).',
        'Ensure this value has at least %(limit_value)d characters (it has %(show_value)d).',
        'limit_value')
    code = 'min_length'

    def compare(self, a, b):
        return a < b

    def clean(self, x):
        return len(x)


@deconstructible
class MaxLengthValidator(BaseValidator):
    message = ngettext_lazy(
        'Ensure this value has at most %(limit_value)d character (it has %(show_value)d).',
        'Ensure this value has at most %(limit_value)d characters (it has %(show_value)d).',
        'limit_value')
    code = 'max_length'

    def compare(self, a, b):
        return a > b

    def clean(self, x):
        return len(x)


@deconstructible
class DecimalValidator:
    """
    Validate that the input does not exceed the maximum number of digits
    expected, otherwise raise ValidationError.
    """
    messages = {
        'max_digits': ngettext_lazy(
            'Ensure that there are no more than %(max)s digit in total.',
            'Ensure that there are no more than %(max)s digits in total.',
            'max'
        ),
        'max_decimal_places': ngettext_lazy(
            'Ensure that there are no more than %(max)s decimal place.',
            'Ensure that there are no more than %(max)s decimal places.',
            'max'
        ),
        'max_whole_digits': ngettext_lazy(
            'Ensure that there are no more than %(max)s digit before the decimal point.',
            'Ensure that there are no more than %(max)s digits before the decimal point.',
            'max'
        ),
    }

    def __init__(self, max_digits, decimal_places):
        self.max_digits = max_digits
        self.decimal_places = decimal_places

    def __call__(self, value):
        digit_tuple, exponent = value.as_tuple()[1:]
        decimals = abs(exponent)
        # digit_tuple doesn't include any leading zeros.
        digits = len(digit_tuple)
        if decimals > digits:
            # We have leading zeros up to or past the decimal point. Count
            # everything past the decimal point as a digit. We do not count
            # 0 before the decimal point as a digit since that would mean
            # we would not allow max_digits = decimal_places.
            digits = decimals
        whole_digits = digits - decimals

        if self.max_digits is not None and digits > self.max_digits:
            raise ValidationError(
                self.messages['max_digits'],
                code='max_digits',
                params={'max': self.max_digits},
            )
        if self.decimal_places is not None and decimals > self.decimal_places:
            raise ValidationError(
                self.messages['max_decimal_places'],
                code='max_decimal_places',
                params={'max': self.decimal_places},
            )
        if (self.max_digits is not None and self.decimal_places is not None and
                whole_digits > (self.max_digits - self.decimal_places)):
            raise ValidationError(
                self.messages['max_whole_digits'],
                code='max_whole_digits',
                params={'max': (self.max_digits - self.decimal_places)},
            )

    def __eq__(self, other):
        return (
            isinstance(other, self.__class__) and
            self.max_digits == other.max_digits and
            self.decimal_places == other.decimal_places
        )


@deconstructible
class FileExtensionValidator:
    message = _(
        "File extension '%(extension)s' is not allowed. "
        "Allowed extensions are: '%(allowed_extensions)s'."
    )
    code = 'invalid_extension'

    def __init__(self, allowed_extensions=None, message=None, code=None):
        if allowed_extensions is not None:
            allowed_extensions = [allowed_extension.lower() for allowed_extension in allowed_extensions]
        self.allowed_extensions = allowed_extensions
        if message is not None:
            self.message = message
        if code is not None:
            self.code = code

    def __call__(self, value):
        extension = os.path.splitext(value.name)[1][1:].lower()
        if self.allowed_extensions is not None and extension not in self.allowed_extensions:
            raise ValidationError(
                self.message,
                code=self.code,
                params={
                    'extension': extension,
                    'allowed_extensions': ', '.join(self.allowed_extensions)
                }
            )

    def __eq__(self, other):
        return (
            isinstance(other, self.__class__) and
            self.allowed_extensions == other.allowed_extensions and
            self.message == other.message and
            self.code == other.code
        )


def get_available_image_extensions():
    try:
        from PIL import Image
    except ImportError:
        return []
    else:
        Image.init()
        return [ext.lower()[1:] for ext in Image.EXTENSION]


validate_image_file_extension = FileExtensionValidator(
    allowed_extensions=get_available_image_extensions(),
)


@deconstructible
class ProhibitNullCharactersValidator:
    """Validate that the string doesn't contain the null character."""
    message = _('Null characters are not allowed.')
    code = 'null_characters_not_allowed'

    def __init__(self, message=None, code=None):
        if message is not None:
            self.message = message
        if code is not None:
            self.code = code

    def __call__(self, value):
        if '\x00' in str(value):
            raise ValidationError(self.message, code=self.code)

    def __eq__(self, other):
        return (
            isinstance(other, self.__class__) and
            self.message == other.message and
            self.code == other.code
        )