Fix bug with loading data labeler from disk

[tonywu315]

When DataLabeler is passed in a dirpath parameter, it loads from the disk and calls _load_parameters in base_data_labeler.py. However, there is a bug when the code checks if the model, preprocessor, or postprocessor matches the one specified in the saved DataLabeler.

from dataprofiler.labelers import DataLabeler, CharPreprocessor, CharPostprocessor
from dataprofiler.labelers.character_level_cnn_model import CharacterLevelCnnModel

data_labeler = DataLabeler(labeler_type="unstructured")
data_labeler.save_to_disk("temp")

label_mapping = {"PAD": 0, "CITY": 1, "UNKNOWN": 1, "ADDRESS": 2}
load_options = {
    "model_class": CharacterLevelCnnModel(label_mapping=label_mapping),
    "preprocessor_class": CharPreprocessor(),
    "postprocessor_class": CharPostprocessor(),
}

new_data_labeler = DataLabeler(labeler_type='unstructured', dirpath="temp", load_options=load_options)

This code generates the error: 'CharacterLevelCNNModel' object has no attribute '__name__'. This is because instances do not have this attribute. The class does however, so all occurrences in the code should be replaced with __class__.__name__. After changing this, the example code works.

[taylorfturner]

LGTM ... great fix!