pydantic validation to raise ValidationError instead of ValueError about pandera HOT 2 OPEN

Digging deeper, i found that the problem is in the function "pydantic_validate" (pandera/typing/pandas) line: 240

try:
    valid_data = schema.validate(data)
except SchemaError as exc:
    raise ValueError(str(exc)) from exc

The issue is that it raises a simplified Exception with no context, pydantic raises a ValidationError.
I found that this little change in the code fixes de issue:

try:
    valid_data = schema.validate(data)
except SchemaError as exc:
    errors_ = ErrorWrapper(exc=exc, loc=exc.schema.name)
    raise ValidationError([errors_], DataFrame[T])

I'm not sure about the "DataFrame[T]" part, but its working.

now i'm getting a much more detailed error:

{'loc': ('body', 'customer_alias'), 'msg': ..., 'type': 'value_error.schema', 'ctx': {'schema': <Schema Column(name=customer_alias, type=DataType(str))>, 'data': ..., 'check': "dtype('str')", 'check_index': None, 'check_output': None, 'reason_code': <SchemaErrorReason.WRONG_DATATYPE: 'wrong_dtype'>}}

from pandera.

the solution i gave may not be correct for pydantic v2
my fastapi project is running pydantic==1.10.12, as pandera already uses pydantic==2.3.0

i think it needs some backwards compatibility.

i'm a bit confused, can someone help?

from pandera.