pd-window-decorators

pd-window-decorators

Easily apply windowing to functions that mutate Pandas DataFrames. Useful for data science projects where you want to apply a function for a DataFrame using smaller chunks of the DataFrame automatically. A moving window can be easily applied to the function using Python decorators.

Sliding Window

To apply a sliding window, import df_sliding_window from pd_window_decorators and apply it to your function. The decorator takes a timedelta object as a required argument to define the slice size. By default, the decorator will look for a Pandas DataFrame named df in the function arguments. The DataFrame must also have a time column. The column name is ds by default. The decorator also expects the function to return a DataFrame.

Example

Using the decorator with default arguments:

@df_sliding_window(window_size=timedelta(days=2))
def sum_all(df):
    df.loc[:, 'sum'] = df['y'].sum()
    return df

Using the decorator with custom arguments:

@df_sliding_window(window_size=timedelta(days=2), df_name='my_df', time_column='my_time')
def sum_all(my_df):
    df.loc[:, 'sum'] = df['y'].sum()
    return df

Note that in the second example, the DataFrame is named my_df and the df_name argument is set to my_df, so they match.

Arguments

Argument	Type	Optional	Description
window_size	timedelta	True	The size of the window to apply to the function.
df_name	str	False	The name of the DataFrame variable to pass to the function as an argument. Defaults to df.
time_column	str	False	The name of the column in the DataFrame that contains the time information. Defaults to ds.