Is there a way to validate file size using python formencode?

Question

I'd like to validate uploaded file's size in my Pyramid application using formencode. As far as I understand, I need to create a class inherited from formencode.validators.FormValidator) and put it to chained_validators. But I can't figure out a way to check the uploaded file's size in the validate_python method. Is it even possible?

Thanks in advance, Ivan.

Answer 1

Another way to do it:

class CheckFileSize(formencode.validators.FormValidator):
    __unpackargs__ = ('upload_field', 'max_file_size')

    def validate_python(self, value_dict, state):
        log.info('test')
        if value_dict.get(self.upload_field) is None:
            return value_dict
        fileobj = getattr(value_dict.get(self.upload_field), 'file', None)
        fileobj.seek(0, os.SEEK_END)
        if int(fileobj.tell()) > int(self.max_file_size):
            raise formencode.Invalid(
                _('File too big'),
                value_dict, state,
                error_dict={self.upload_field:
                    formencode.Invalid(_('File too big'), value_dict, state)})
        return value_dict

class CreateNewCaseForm(formencode.Schema):
    ...
    chained_validators = [
        CheckFileSize('file', max_upload_size),
    ]

Answer 2

You can use len() (in the validator) on the file object itself to check the file size, since its counting the bytes.

size = len(fileObj)

Answer 3

Sure there is - although I did that with turbogears, it should work with pyramid as well:

class MyFileValidator(FancyValidator):
    def _to_python(self, value, state):
        max_size = 10*1024*1024

        payload = value.file.read(max_size+1)

        # rewind so that the application can access the content
        value.file.seek(0)

        if len(payload) == max_size:
            raise Invalid(u"The file is too big (>10MB)", value, state)

        return value


class MySchema(Schema):
    my_file = MyFileValidator(not_empty=True)

Note that read() ing the whole data should not be necessary (see other answer) - I did that for further content validation.