Describe the bug
When instantiating a Pydantic model with a custom field type that inherits from datetime and utilizes Arrow for datetime parsing, PyCharm's Pydantic plugin generates a type checking warning. The warning indicates a mismatch between the expected custom type and the actual Arrow type provided, even though the custom type is designed to accept Arrow objects or strings and convert them into datetime instances. The warning message is: "Expected type 'ArrowPydantic', got 'Arrow' instead."

To Reproduce
Steps to reproduce the behavior:

1. Define a custom Pydantic type ArrowPydantic that inherits from datetime and includes a validator to convert inputs into datetime instances, accepting either Arrow objects or strings.

# Based on https://stackoverflow.com/a/75949606
class ArrowPydantic(datetime):
    @classmethod
    def __get_validators__(cls):
        yield cls.validate

    @classmethod
    def validate(cls, value: Union[Arrow, str]) -> datetime:
        if type(value) is str:
            value: Arrow = arrow.get(value)
        return value._datetime

2. Define a Pydantic model AuditLogRecord that uses ArrowPydantic for one of its fields.

class AuditLogRecord(JsonResourceModel):
    """
    Record of an audit log, part of a project package.
    """
    actorEmailAddress: str = Field(
        description='Email address of the person/system performing the action.',
    )
    actorName: str = Field(
        description='Name of the person/system performing the action.',
    )
    action: str = Field(
        description='The action that was performed.',
    )
    performedAt: ArrowPydantic = Field(
        description='The time that the action was performed.',
    )

    class Config:
        arbitrary_types_allowed = True

3. Attempt to instantiate AuditLogRecord with an Arrow object for the ArrowPydantic field.

    # Okay, now create a record
    auditLogRecord = AuditLogRecord(
        actorEmailAddress=actorEmailAddress,
        actorName=actorName,
        action=action,
        performedAt=performedAt,
    )

4. Observe the type checking warning in PyCharm indicating a mismatch between the expected and actual types for the field.

Expected behavior
The expected behavior is that the PyCharm Pydantic plugin correctly recognizes the custom type's ability to handle both Arrow objects and strings, converting them into datetime instances, without generating a type checking warning. The instantiation of AuditLogRecord with an Arrow object should be considered valid, given the custom logic defined in ArrowPydantic.

Screenshots

Environments (please complete the following information):

• IDE: PyCharm Professional Edition 2023.3.4
• OS: Ubuntu 23.10
• Pydantic Version: 1.10.0
• Plugin version: 0.4.11

Additional context
None