template for model.py file for model deployment

model.py

class Model():
    """This defines the necessary attributes for the model object
    """
    
    def __init__(self):
    # specify your data for fitting here
        self.X_fit = []
        self.Y_fit = []
        self.model_store = "really.awesome.fast.model.store"
        self.feature_store = "really.awesome.fast.feature.store"
        self.pre_processing = True
        self.FEATURE_COLUMNS = ['']

    def fit(self):
        """Fit the model based on input data
        Args:
            X (list of dict): list of dictionaries containing training data information
            y (list of int): list of relevance scores for each input feature vector
        """
        print("fitting model ...")
        print("model fit success")

    def get_features(self,*args):
        """Extract the required features and manipulate as necessary
        Best practice is to move the manipulation to a preprocessing job done in the ETL tool.
        Use this to manipulate as a last resort
        Args:
            X (list of dict): list of dictionaries containing training data information
            y (list of int): list of relevance scores for each input feature vector
        """
        print("extracting features ... from {}".format(self.feature_store))
        print("feature extraction success")
        return({})

    def pre_process(self, payload, features=None):
        """tertiary processing prior to predictions
        Args:
            payload from request
        Return:
            prediction payload
        """
        processed = payload

        return processed

    def predict(self, X):
        """Predict relevance score for given search results
        Args:
            X (list of dict): feature vectors for results in a search session
        Return:
            list of float: relevance score for given content
        """

        # Just use heuristic for now
        return X

    def save(self):
        """Save trained model to HDFS
        This method will be run after a full training build on the remote instance to save the
            completed model to a place where it can be easily retrieved
        Args:
            model_store (str): Directory in HDFS within which the model should be saved
        """
        print("saving model to {}...".format(self.model_store))
        print("model save success")

    def load(self):
        """Load trained model from HDFS
        This method will be run upon each worker being spun up to retrieve the latest trained model
        Args:
            model_store (str): Directory in HDFS within which the model should be saved
        """
        print("loading model from {}...".format(self.model_store))
        print("model load success")