Find best 20-minute segment since latest FTP test

pel.py

import numpy as np
from datetime import datetime
from tqdm.auto import tqdm
from pytz import timezone
from peloton import PelotonWorkout

my_tz = timezone('US/Eastern')

print('Getting workout list...')
workouts = PelotonWorkout.list()
print('Finding latest FTP test...')
ftp = [workout for workout in workouts
       if 'ftp' in workout.ride.title.lower().split() and
       'warm' not in workout.ride.title.lower()]
assert len(ftp) >= 1, 'No FTP test found'
ftp = sorted(ftp, key=lambda x: x.start_time)[-1]  # newest one
factor = 0.95
ftp_watts = np.array(ftp.metrics.output.values).mean()
print(f'{ftp.ride.instructor.first_name}\'s {ftp.ride.title}: '
      f'{ftp.start_time.astimezone(my_tz)}\n'
      f'    {ftp_watts * factor:0.1f} watts (from {ftp_watts:0.1f})')
workouts = [
    workout for workout in tqdm(workouts, desc='Fetching metrics')
    if 'ride' in workout.ride.title.lower().split() and  # is a ride
    'stretch' not in workout.ride.title.lower() and  # not a stretch
    workout.start_time > ftp.start_time and  # strictly after latest FTP
    hasattr(workout.metrics, 'output') and  # has meaningful metrics
    len(workout.metrics.output.values)
]
# figure out the best 20-minute for each ride
print(f'Finding best non-FTP ride (of {len(workouts)} eligible)...')
mean_20 = np.ones(20 * 60)  # over 20 min
mean_20 /= len(mean_20)
cont_20s = [
    np.convolve(mean_20, workout.metrics.output.values, mode='same').max()
    for workout in workouts]
idx = np.argmax(cont_20s)
workout = workouts[idx]
print(f'{workout.ride.instructor.first_name}\'s {workout.ride.title}: '
      f'{workout.start_time.astimezone(my_tz)}\n'
      f'    {cont_20s[idx] * factor:0.1f} watts '
      f'(from {cont_20s[idx]:0.1f})')