re.error not wrapped in ValidationError when invalid regex is provided for Pattern field

[Gordon Shankman]

I have a data model that accepts a user-provided regular expression with the model defined as:

@dataclass
class Regex:
    pattern: Pattern = field(metadata=schema(min_len=1))
    group: int = field(default=0, metadata=schema(min=0))

I have several validation tests to confirm expected behavior and, when I try to verify that a pattern of "(unclosed" results in a ValidationError, my tests fail because the re.error thrown by re.compile() is being raised instead.

I tried defining a custom deserializer for Pattern, but it isn't getting executed

@deserializer
def to_pattern(pattern: str) -> Pattern:
    try:
        return re.compile(pattern)
    except re.error as err:
        raise ValidationError(f"invalid regex pattern '{pattern}': {err}")

Not sure if I'm doing something wrong with the configuration or if the default deserializer for re.Pattern needs to be updated.

[Gordon Shankman]

This works as a workaround, but was hoping to have the more semantically correct typing.

@dataclass
class Regex:
    pattern: str = field(metadata=schema(min_len=1))
    group: int = field(default=0, metadata=schema(min=0))

    @validator
    def check_regex_pattern(self):
        try:
            re.compile(self.pattern)
        except re.error as err:
            raise ValidationError(f"invalid regex pattern '{self.pattern}': {err}")