调用KickstarterModel类方法时触发NotImplementedError，求排查方案

Let's break down the issues in your code that are causing the error and fix them step by step:

1. The Root Cause of NotImplementedError #

Looking at your error traceback, it seems your preprocess_training_data method contains an invalid line that triggers the NotImplementedError. In the snippet from your error log:

def preprocess_training_data(self, df):
    NotImplementedError:

This line is leftover from a placeholder pattern (common when sketching out class methods initially) but was never removed once you added the actual preprocessing logic. Just writing NotImplementedError: like this is invalid syntax and will throw the error immediately when the method runs.

2. Additional Bugs in Your Preprocessing Logic #

Even after fixing the error above, your code has two critical issues that will break functionality:

• Undefined variable y: The final line y_train = y references a variable y that doesn't exist. You meant to use the y_train variable you defined earlier in the method.
• Wasted processed data: You're storing the cleaned features and target in local variables X_train and y_train instead of the class instance attributes self.X_train and self.y_train. This means the data won't be accessible to other methods (like a train method) later on.

Fixed Code #

Here's the corrected version of your KickstarterModel class:

from sklearn.tree import DecisionTreeClassifier

class KickstarterModel:
    def __init__(self):
        self.model = DecisionTreeClassifier()
        self.X_train = None
        self.y_train = None
    
    def preprocess_training_data(self, df):
        # Define target variable
        y_train = df['state'].apply(lambda x: 0 if x == 'failed' else 1)
        
        # Process numerical features
        numerical_columns = df.describe().columns
        X = df[numerical_columns].copy()
        X.set_index('id', inplace=True)
        X["dif_launched_created"] = X.launched_at - X.created_at
        X["dif_deadline_created"] = X.deadline - X.created_at
        X["dif_deadline_launched"] = X.deadline - X.launched_at
        X['goal_in_usd'] = X.goal * X.static_usd_rate
        X.drop(['deadline', 'created_at', 'launched_at', 'goal', 'static_usd_rate'], axis=1, inplace=True)
        
        # Save processed data to instance attributes
        self.X_train = X
        self.y_train = y_train

Key Changes Made: #

• Removed the invalid NotImplementedError: line from the method body
• Fixed the undefined y variable by using the existing y_train variable
• Assigned processed features and target to self.X_train and self.y_train so the data persists in the class instance for later use (like training the classifier)

After making these changes, calling model.preprocess_training_data(df) should run without errors, and your cleaned training data will be ready to use with your decision tree model.

内容的提问来源于stack exchange，提问作者Tom Smith