Skip to content

Predictor

Predictor #

Base class. Implements all low-level functionality.

Parameters:

  • path (str, default: None ) –

    Directory location to store all outputs. Defaults to None. If None, a new unique time-stamped directory is chosen.

  • name (str, default: None ) –

    Name of the subdirectory inside path where the model will be saved. The final model directory will be os.path.join(path, name). If None, defaults to the model's class name: self.__class__.__name__.

Source code in src/qtt/predictors/predictor.py 
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
class Predictor:
    """Base class. Implements all low-level functionality.

    Args:
        path (str, optional): Directory location to store all outputs. Defaults to None.
            If None, a new unique time-stamped directory is chosen.
        name (str, optional): Name of the subdirectory inside `path` where the model will be saved.
            The final model directory will be `os.path.join(path, name)`.
            If None, defaults to the model's class name: `self.__class__.__name__`.
    """

    model_file_name = "model.pkl"

    def __init__(
        self,
        name: str | None = None,
        path: str | None = None,
    ):
        if name is None:
            self.name = self.__class__.__name__
            logger.info(
                f"No name was specified for model, defaulting to class name: {self.name}",
            )
        else:
            self.name = name

        if path is None:
            self.path: str = setup_outputdir(path=self.name.lower())
            logger.info(
                f"No path was specified for predictor, defaulting to: {self.path}",
            )
        else:
            self.path = setup_outputdir(path)

        self.model = None

    def reset_path(self, path: str | None = None):
        """
        Reset the path of the model.

        Args:
            path (str, optional):
                Directory location to store all outputs. If None, a new unique time-stamped directory is chosen.
        """
        if path is None:
            path = setup_outputdir(path=self.name.lower())
        self.path = path

    @property
    def is_fit(self) -> bool:
        """Returns True if the model has been fit."""
        return self.model is not None

    def fit(self, X: pd.DataFrame, y: ArrayLike, **kwargs):
        """
        Fit model to predict values in y based on X.

        Models should not override the `fit` method, but instead override the `_fit` method which has the same arguments.

        Args:
            X (pd.DataFrame):
                The training data features.
            y (ArrayLike):
                The training data ground truth labels.
            **kwargs :
                Any additional fit arguments a model supports.
        """
        out = self._fit(X=X, y=y, **kwargs)
        if out is None:
            out = self
        return out

    def _fit(self, X: pd.DataFrame, y: ArrayLike, **kwargs):
        """
        Fit model to predict values in y based on X.

        Models should override this method with their custom model fit logic.
        X should not be assumed to be in a state ready for fitting to the inner model, and models may require special preprocessing in this method.
        It is very important that `X = self.preprocess(X)` is called within `_fit`, or else `predict` and `predict_proba` may not work as intended.
        It is also important that `_preprocess` is overwritten to properly clean the data.
        Examples of logic that should be handled by a model include missing value handling, rescaling of features (if neural network), etc.
        If implementing a new model, it is recommended to refer to existing model implementations and experiment using toy datasets.

        Refer to `fit` method for documentation.
        """
        raise NotImplementedError

    def _preprocess(self, **kwargs):
        """
        Data transformation logic should be added here.

        Input data should not be trusted to be in a clean and ideal form, while the output should be in an ideal form for training/inference.
        Examples of logic that should be added here include missing value handling, rescaling of features (if neural network), etc.
        If implementing a new model, it is recommended to refer to existing model implementations and experiment using toy datasets.

        In bagged ensembles, preprocessing code that lives in `_preprocess` will be executed on each child model once per inference call.
        If preprocessing code could produce different output depending on the child model that processes the input data, then it must live here.
        When in doubt, put preprocessing code here instead of in `_preprocess_nonadaptive`.
        """
        raise NotImplementedError

    def preprocess(self, **kwargs):
        """
        Preprocesses the input data into internal form ready for fitting or inference.
        """
        return self._preprocess(**kwargs)

    @classmethod
    def load(cls, path: str, reset_paths: bool = True, verbose: bool = True):
        """
        Loads the model from disk to memory.

        Args:
            path (str):
                Path to the saved model, minus the file name.
                This should generally be a directory path ending with a '/' character (or appropriate path separator value depending on OS).
                The model file is typically located in os.path.join(path, cls.model_file_name).
            reset_paths (bool):
                Whether to reset the self.path value of the loaded model to be equal to path.
                It is highly recommended to keep this value as True unless accessing the original self.path value is important.
                If False, the actual valid path and self.path may differ, leading to strange behaviour and potential exceptions if the model needs to load any other files at a later time.
            verbose (bool):
                Whether to log the location of the loaded file.

        Returns:
            model (Predictor): Loaded model object.
        """
        file_path = os.path.join(path, cls.model_file_name)
        with open(file_path, "rb") as f:
            model = pickle.load(f)
        if reset_paths:
            model.path = path
        if verbose:
            logger.info(f"Model loaded from: {file_path}")
        return model

    def save(self, path: str | None = None, verbose: bool = True) -> str:
        """
        Saves the model to disk.

        Args:
            path (str): Path to the saved model, minus the file name.
                This should generally be a directory path ending with a '/' character (or appropriate path separator value depending on OS).
                If None, self.path is used.
                The final model file is typically saved to os.path.join(path, self.model_file_name).
            verbose (bool): Whether to log the location of the saved file.

        Returns:
            path: Path to the saved model, minus the file name. Use this value to load the model from disk via cls.load(path), cls being the class of the model object, such as ```model = PerfPredictor.load(path)```
        """
        if path is None:
            path = self.path
        path = setup_outputdir(path, create_dir=True, warn_if_exist=True)
        file_path = os.path.join(path, self.model_file_name)
        with open(file_path, "wb") as f:
            pickle.dump(self, f, protocol=pickle.HIGHEST_PROTOCOL)
        if verbose:
            logger.info(f"Model saved to: {file_path}")
        return path

    def predict(self, **kwargs) -> np.ndarray | Tuple[np.ndarray, ...]:
        """
        Predicts the output for the given input data.

        Models should not override the `predict` method, but instead override the `_predict` method
        which has the same arguments.
        """
        return self._predict(**kwargs)

    def _predict(self, **kwargs):
        """
        Predicts the output for the given input data.

        New predictors should override this method with their custom prediction logic.
        """
        raise NotImplementedError

is_fit: bool property #

Returns True if the model has been fit.

fit(X, y, **kwargs) #

Fit model to predict values in y based on X.

Models should not override the fit method, but instead override the _fit method which has the same arguments.

Parameters:

  • X (DataFrame) –

    The training data features.

  • y (ArrayLike) –

    The training data ground truth labels.

  • **kwargs –

    Any additional fit arguments a model supports.

Source code in src/qtt/predictors/predictor.py 
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
def fit(self, X: pd.DataFrame, y: ArrayLike, **kwargs):
    """
    Fit model to predict values in y based on X.

    Models should not override the `fit` method, but instead override the `_fit` method which has the same arguments.

    Args:
        X (pd.DataFrame):
            The training data features.
        y (ArrayLike):
            The training data ground truth labels.
        **kwargs :
            Any additional fit arguments a model supports.
    """
    out = self._fit(X=X, y=y, **kwargs)
    if out is None:
        out = self
    return out

load(path, reset_paths=True, verbose=True) classmethod #

Loads the model from disk to memory.

Parameters:

  • path (str) –

    Path to the saved model, minus the file name. This should generally be a directory path ending with a '/' character (or appropriate path separator value depending on OS). The model file is typically located in os.path.join(path, cls.model_file_name).

  • reset_paths (bool, default: True ) –

    Whether to reset the self.path value of the loaded model to be equal to path. It is highly recommended to keep this value as True unless accessing the original self.path value is important. If False, the actual valid path and self.path may differ, leading to strange behaviour and potential exceptions if the model needs to load any other files at a later time.

  • verbose (bool, default: True ) –

    Whether to log the location of the loaded file.

Returns:

  • model ( Predictor ) –

    Loaded model object.

Source code in src/qtt/predictors/predictor.py 
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
@classmethod
def load(cls, path: str, reset_paths: bool = True, verbose: bool = True):
    """
    Loads the model from disk to memory.

    Args:
        path (str):
            Path to the saved model, minus the file name.
            This should generally be a directory path ending with a '/' character (or appropriate path separator value depending on OS).
            The model file is typically located in os.path.join(path, cls.model_file_name).
        reset_paths (bool):
            Whether to reset the self.path value of the loaded model to be equal to path.
            It is highly recommended to keep this value as True unless accessing the original self.path value is important.
            If False, the actual valid path and self.path may differ, leading to strange behaviour and potential exceptions if the model needs to load any other files at a later time.
        verbose (bool):
            Whether to log the location of the loaded file.

    Returns:
        model (Predictor): Loaded model object.
    """
    file_path = os.path.join(path, cls.model_file_name)
    with open(file_path, "rb") as f:
        model = pickle.load(f)
    if reset_paths:
        model.path = path
    if verbose:
        logger.info(f"Model loaded from: {file_path}")
    return model

predict(**kwargs) #

Predicts the output for the given input data.

Models should not override the predict method, but instead override the _predict method which has the same arguments.

Source code in src/qtt/predictors/predictor.py 
175
176
177
178
179
180
181
182
def predict(self, **kwargs) -> np.ndarray | Tuple[np.ndarray, ...]:
    """
    Predicts the output for the given input data.

    Models should not override the `predict` method, but instead override the `_predict` method
    which has the same arguments.
    """
    return self._predict(**kwargs)

preprocess(**kwargs) #

Preprocesses the input data into internal form ready for fitting or inference.

Source code in src/qtt/predictors/predictor.py 
116
117
118
119
120
def preprocess(self, **kwargs):
    """
    Preprocesses the input data into internal form ready for fitting or inference.
    """
    return self._preprocess(**kwargs)

reset_path(path=None) #

Reset the path of the model.

Parameters:

  • path (str, default: None ) –

    Directory location to store all outputs. If None, a new unique time-stamped directory is chosen.

Source code in src/qtt/predictors/predictor.py 
51
52
53
54
55
56
57
58
59
60
61
def reset_path(self, path: str | None = None):
    """
    Reset the path of the model.

    Args:
        path (str, optional):
            Directory location to store all outputs. If None, a new unique time-stamped directory is chosen.
    """
    if path is None:
        path = setup_outputdir(path=self.name.lower())
    self.path = path

save(path=None, verbose=True) #

Saves the model to disk.

Parameters:

  • path (str, default: None ) –

    Path to the saved model, minus the file name. This should generally be a directory path ending with a '/' character (or appropriate path separator value depending on OS). If None, self.path is used. The final model file is typically saved to os.path.join(path, self.model_file_name).

  • verbose (bool, default: True ) –

    Whether to log the location of the saved file.

Returns:

  • path ( str ) –

    Path to the saved model, minus the file name. Use this value to load the model from disk via cls.load(path), cls being the class of the model object, such as model = PerfPredictor.load(path)

Source code in src/qtt/predictors/predictor.py 
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
def save(self, path: str | None = None, verbose: bool = True) -> str:
    """
    Saves the model to disk.

    Args:
        path (str): Path to the saved model, minus the file name.
            This should generally be a directory path ending with a '/' character (or appropriate path separator value depending on OS).
            If None, self.path is used.
            The final model file is typically saved to os.path.join(path, self.model_file_name).
        verbose (bool): Whether to log the location of the saved file.

    Returns:
        path: Path to the saved model, minus the file name. Use this value to load the model from disk via cls.load(path), cls being the class of the model object, such as ```model = PerfPredictor.load(path)```
    """
    if path is None:
        path = self.path
    path = setup_outputdir(path, create_dir=True, warn_if_exist=True)
    file_path = os.path.join(path, self.model_file_name)
    with open(file_path, "wb") as f:
        pickle.dump(self, f, protocol=pickle.HIGHEST_PROTOCOL)
    if verbose:
        logger.info(f"Model saved to: {file_path}")
    return path