Python script for plotting my weight over time

plot_my_weight.py

""" plot_my_weight.py

This is a simple scatter-and-line plot of my body weight at various times. The challenge is that
the data is very irregular. This program reads in a CSV of this form:

datetime,weight(lbs),Steps,Meals,Calories,Run (km),anything,else...
2010-02-18,165
2010-02-25,156.6
2010-03-11,154.4
2022-06-22,159.1,10000,2,1500,5,...
"""
from datetime import datetime, timedelta
import matplotlib as mpl
import numpy as np
import sys

from matplotlib import style
style.use("ggplot")

# set the matplotlib backend
gui_env = ['TKAgg','GTKAgg','Qt4Agg','WXAgg']
for gui in gui_env:
    try:
        mpl.use(gui, warn=False, force=True)
        from matplotlib import pyplot as plt
        break
    except:
        continue
print(f"Using: {mpl.get_backend()}")
from matplotlib import pyplot as plt

# CONFIG
START_DATE = datetime(1981, 6, 22)
END_DATE = datetime(2200, 1, 1)
WEIGHT_FILE_PATH = "/home/john/Recordings/Personal/Code/ops/my_weight.csv"


# CONSTANTS
DAY_FORMAT = "%Y-%m-%d"
HOUR_FORMAT = "%Y-%m-%d-%H"
MIN_FORMAT = "%Y-%m-%d-%H-%M"


def main():
    """ A simple command-line parser to make this script easier to use. """
    # Constants
    DAY_FORMAT = "%Y-%m-%d"
    start_date = START_DATE
    end_date = END_DATE
    weight_file = WEIGHT_FILE_PATH

    try:
        if len(sys.argv) > 1:
            start_date = datetime.strptime(sys.argv[1], DAY_FORMAT)
        if len(sys.argv) > 2:
            end_date = datetime.strptime(sys.argv[2], DAY_FORMAT)
        if len(sys.argv) > 3:
            weight_file = sys.argv[3]
    except ValueError:
        usage()
    if len(sys.argv) > 4:
        usage()

    plot_my_weight(start_date, end_date, weight_file)


def usage():
    """ This is just a simple print-to-screen if the users throw nonsense at the command line. """
    print("\nUsage:\n\npython weights.py\n\nOR\n")
    print("python plot_my_weight.py 2014-01-01")
    print("python plot_my_weight.py 2014-01-01 2015-12-31")
    print("python plot_my_weight.py 2014-01-01 2015-12-31 /path/to/my_weight.csv\n")
    exit()


def plot_my_weight(start_date, end_date, weight_file):
    """This is the meat of the script.
    Create a quick matplotlib plot of weights and display it.
    """
    # read weight file
    f = open(weight_file, "r")
    _ = f.readline()
    lines = f.readlines()
    f.close()

    # ingest weight data
    dates = []
    weights = []

    # loop through all lines in file
    for line in lines:
        # split line into columns
        bits = line.strip().split(",")
        date_str = bits[0].strip()
        weight = bits[1].strip()

        if not weight:
            continue

        # parse dates
        if len(date_str) == 10:
            date = datetime.strptime(date_str, DAY_FORMAT)
        elif len(date_str) == 13:
            date = datetime.strptime(date_str, HOUR_FORMAT)
        elif len(date_str) == 16:
            date = datetime.strptime(date_str, MIN_FORMAT)

        # make sure date is in range
        if date >= start_date and date <= end_date:
            dates.append(date)
        else:
            continue

        # parse weight
        weights.append(float(weight))

    # convert data to NumPy array
    x = np.array(dates)
    y = np.array(weights)

    # build better weekly average
    # 1. find the smallest value in each day
    sdays = {}
    for i, d in enumerate(dates):
        yr = d.year
        jday = d.timetuple().tm_yday
        sday = (yr, jday)
        if sday not in sdays or weights[i] < sdays[sday]:
            sdays[sday] = weights[i]

    # 2. build the weekly data we need to average a week
    weekly = {}
    for sday, min_val in sdays.items():
        yr, jday = sday
        d = datetime.strptime(str(yr) + str(jday), "%Y%j").date()
        monday = d - timedelta(days=d.weekday())
        if monday not in weekly:
            weekly[monday] = []
        weekly[monday].append(min_val)

    # 3. do the weekly averages
    weekly_dates = []
    weekly_weights_min = []
    weekly_weights_avg = []
    weekly_weights_max = []
    for monday in sorted(weekly.keys()):
        weekly_dates.append(monday + timedelta(days=3))
        avg = sum(weekly[monday]) / len(weekly[monday])
        weekly_weights_avg.append(avg)
        weekly_weights_min.append(min(weekly[monday]))
        weekly_weights_max.append(max(weekly[monday]))

    # 4. create plot of raw data and weekly averages
    plt.plot(weekly_dates, weekly_weights_avg, color="0.4", zorder=5, linewidth=6)
    plt.scatter(x, y, c=y/max(y), s=25, zorder=3, linewidth=0, cmap="rainbow")
    #plt.plot(weekly_dates, weekly_weights_avg, color="0.8", zorder=1, linewidth=24)
    plt.plot(weekly_dates, weekly_weights_min, color="0.8", zorder=2, linewidth=2)
    plt.plot(weekly_dates, weekly_weights_max, color="0.8", zorder=2, linewidth=2)
    plt.ylim(ymin=min(y) - 5)
    min_yr = min([d.year for d in dates])
    max_yr = max([d.year for d in dates])
    plt.xticks([datetime(yr, 1, 1) for yr in range(min_yr, max_yr + 2)], rotation=45)

    # 5. add vertical lines for when I lived where
    plt.axvline(datetime(2018, 4, 1), color="gray", label="Boston")
    plt.axvline(datetime(2021, 7, 4), color="gray", label="Seattle")
    plt.axvline(datetime(2022, 7, 15), color="gray", label="Spokane")

    plt.show()


if __name__ == "__main__":
    main()