Script to parse the Webtime Tracker CSV file and plot it

parse_webtime.py

#!/usr/bin/env python

import argparse
from datetime import datetime

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt


def smoothed(t: np.ndarray, smoothing: int) -> np.ndarray:
    if smoothing <= 0:
        return t
    left, right = smoothing // 2, (smoothing - 1) // 2
    window = np.hanning(smoothing)
    window /= window.sum()
    smooth_t = np.convolve(t, window)[left:-right]
    return smooth_t


def main(args: argparse.Namespace) -> None:
    data = pd.read_csv(args.csv)
    largest = data.loc[data.sum(1).nlargest(args.num_sites).index.values]
    domains = list(largest["Domain"])
    times = largest.to_numpy()[:, 1:]
    dates = [
        datetime.strptime(date_string, '%Y-%m-%d')
        for date_string in data.columns[1:]
    ]

    fig = plt.figure(figsize=(10, 5))
    for domain, time in zip(domains, times):
        time = (np.array(time, dtype=np.int32) + 30) // 60
        plt.plot(dates, smoothed(time, args.smoothing), label=domain)
    plt.legend(title="Sites", bbox_to_anchor=(1.05, 1), loc="upper left")
    fig.autofmt_xdate()
    plt.xlabel("Date")
    plt.ylabel("Minutes")
    plt.tight_layout()
    plt.savefig("time.svg")


if __name__ == "__main__":
    parser = argparse.ArgumentParser(description="Parses Webtime tracker data")
    parser.add_argument("csv", help="Path to the Webtime CSV")
    parser.add_argument("-n", "--num-sites", type=int, default=10,
                        help="Number of sites to look at")
    parser.add_argument("-s", "--smoothing", type=int, default=20,
                        help="Smoothing window length")
    args = parser.parse_args()

    main(args)