Logging function to be used in pd.DataFrame.pipe().

pandas_pipe_logger.py

def pipe_log(data, logger, message=None, level='DEBUG'):
    '''
    Logging function to be used in pd.DataFrame.pipe(). Insert anywhere 
    in a method chain to log head and shape of dataframe/series along 
    with a custom message with the set logging level, then returns the 
    dataframe unchanged.
    
    Params:
    
    - data: pandas.DataFrame or pandas.Series
    - logger: logging.Logger
    - message: str (optional, default=None)
    - level: str (optional, default='DEBUG')
    
    Example:
    
    # Logs message, head and shape before calling sum() on dataframe
    df.pipe(pipe_log, logger, message='Test', level='INFO').sum()
    '''
    
    lvl = getattr(logging, level, 10)
    
    if message is not None:
        logger.log(lvl, message)
    
    logger.log(lvl, data.head())
    logger.log(lvl, data.shape)

    return data

Could be further abstracted by accepting list of DataFrame properties and methods that should be logged, instead of hard coded .head() and .shape.