1. import pandas as pd
  2. import numpy as np
  4. def flag_cookie_anomalies(df, threshold=3, dry_run=True):
  5. """
  6. Flags anomalous cookies in a DataFrame.
  8. Args:
  9. df (pd.DataFrame): DataFrame containing cookie data. Must have columns like 'cookie_name', 'value', 'timestamp'.
  10. threshold (float): Number of standard deviations from the mean to consider a cookie anomalous.
  11. dry_run (bool): If True, only prints flagged cookies; doesn't modify the DataFrame.
  13. Returns:
  14. pd.DataFrame: DataFrame with an added 'is_anomaly' column indicating whether a cookie is anomalous.
  15. """
  17. if not isinstance(df, pd.DataFrame):
  18. raise TypeError("df must be a pandas DataFrame")
  20. required_columns = ['cookie_name', 'value', 'timestamp']
  21. for col in required_columns:
  22. if col not in df.columns:
  23. raise ValueError(f"DataFrame must contain column '{col}'")
  26. # Calculate statistics for cookie values
  27. cookie_values = df['value'].values
  28. mean = np.mean(cookie_values)
  29. std = np.std(cookie_values)
  31. # Identify anomalous cookies
  32. df['is_anomaly'] = np.isclose(cookie_values, mean, atol=threshold * std)
  34. if not dry_run:
  35. # Filter and print anomalous cookies
  36. anomalous_cookies = df[df['is_anomaly']]
  37. print("Anomalous Cookies:")
  38. print(anomalous_cookies)
  39. else:
  40. print("Dry run mode: No changes made to the DataFrame.")
  42. return df

Add your comment