def test_activity_power_profile_max_duration_too_large():
# test that there is no segmentation fault when max_duration is set too
# large and that we fall back to the largest possible interval.
activity = bikeread(load_fit()[0])
power_profile = activity_power_profile(activity, max_duration=1000000)
assert power_profile.shape == (13536, )
assert power_profile.iloc[-1] == pytest.approx(8.2117765957446736)
def test_load_fit(returned_type, set_data, expected_filenames):
filenames = load_fit(returned_type=returned_type,
set_data=set_data)
if not isinstance(filenames, list):
filenames = [filenames]
for f, e in zip(filenames, expected_filenames):
assert e in f
def test_activity_power_profile_max_duration_too_large():
# test that there is no segmentation fault when max_duration is set too
# large and that we fall back to the largest possible interval.
activity = bikeread(load_fit()[0])
power_profile = activity_power_profile(activity, max_duration=1000000)
assert power_profile.shape == (13536,)
assert power_profile.iloc[-1] == pytest.approx(8.2117765957446736)
def test_load_power_normal_file():
filenames = load_fit(set_data='corrupted')
pattern = '2013-04-24-22-22-25.fit'
for f in filenames:
if pattern in f:
filename = f
df = load_power_from_fit(filename)
assert_allclose(df['power'], ride)
assert df.index[0].date() == date(2013, 4, 24)
def test_load_power_if_no_record():
filenames = load_fit(set_data='corrupted')
pattern = '2015-11-27-18-54-57.fit'
for f in filenames:
if pattern in f:
filename = f
msg = "does not contain any data."
with pytest.raises(IOError, message=msg):
load_power_from_fit(filename)
def test_load_power_normal_file():
filenames = load_fit(set_data='corrupted')
pattern = '2013-04-24-22-22-25.fit'
for f in filenames:
if pattern in f:
filename = f
df = load_power_from_fit(filename)
assert_allclose(df['power'], ride)
assert df.index[0].date() == date(2013, 4, 24)
def test_load_power_if_no_record():
filenames = load_fit(set_data='corrupted')
pattern = '2015-11-27-18-54-57.fit'
for f in filenames:
if pattern in f:
filename = f
msg = "does not contain any data."
with pytest.raises(IOError, message=msg):
load_power_from_fit(filename)
"""
print(__doc__)
# Authors: Guillaume Lemaitre <*****@*****.**>
# License: BSD 3 clause
###############################################################################
# We will use the :class:`skcycling.Rider` class to compute power-profile for
# the toy data sets.
from skcycling.datasets import load_fit
from skcycling import Rider
rider = Rider()
rider.add_activities(load_fit())
print('The computed activities are:\n {}'.format(rider.power_profile_))
###############################################################################
# We can store and load the information using the `to_csv` and `from_csv`
# methods.
filename_rider = 'rider.csv'
rider.to_csv(filename_rider)
rider_reloaded = Rider.from_csv(filename_rider)
###############################################################################
# Clean the temporary csv file
import os
cyclist during a cycling ride.
"""
print(__doc__)
# Authors: Guillaume Lemaitre <*****@*****.**>
# License: BSD 3 clause
###############################################################################
# We can first read a toy ride from the library
from skcycling.datasets import load_fit
from skcycling.io import bikeread
ride = bikeread(load_fit()[0])
###############################################################################
# Different scores are available in ``scikit-cycling``. We can first compute
# the normalized power® which corresponds to an average power during the ride
# with an additional smoothing. To compute this score, we need to know the
# maximum power aerobic (or the functional threshold power).
from skcycling.metrics import normalized_power_score
mpa = 400
np_score = normalized_power_score(ride['power'], mpa)
print('Normalized power {:.2f}'.format(np_score))
###############################################################################
# The intensity factor® normalize the normalized power using the functional
to a cyclist to infer power produced by a cyclist. """ print(__doc__) # Authors: Guillaume Lemaitre <*****@*****.**> # License: BSD 3 clause ############################################################################### # We can first grab a file and read all data available from skcycling.datasets import load_fit from skcycling.io import bikeread ride = bikeread(load_fit()[0], drop_nan='columns') ############################################################################### # We can use a physical model to infer the power. from skcycling.model import strava_power_model power = strava_power_model(ride, cyclist_weight=72) ############################################################################### # We can plot the measured and estimated power to observe the difference. We # can also compute the median absolute error to asses the quality of the # estimation. To ease the interpretation, we will first resample the data. import matplotlib.pyplot as plt from sklearn.metrics import median_absolute_error
def test_check_filename_fit():
filename = load_fit()[0]
my_filename = check_filename_fit(filename)
assert my_filename == filename
def test_activity_power_profile(max_duration, power_profile_shape,
first_element):
activity = bikeread(load_fit()[0])
power_profile = activity_power_profile(activity, max_duration=max_duration)
assert power_profile.shape == power_profile_shape
assert power_profile.iloc[-1] == pytest.approx(first_element)
"""
# Authors: Guillaume Lemaitre <*****@*****.**>
# License: BSD 3 clause
print(__doc__)
###############################################################################
# First, we will load an activity from the toy data set available in
# scikit-cycling.
from skcycling.datasets import load_fit
from skcycling.io import bikeread
ride = bikeread(load_fit()[0], drop_nan='columns')
###############################################################################
# We will only select some of interesting information
columns_selected = ['power', 'speed', 'cadence']
ride = ride[columns_selected]
###############################################################################
# The power-profile is extracted from the ride. By default, the maximum
# duration corresponds to the duration of the ride. However, to limit the
# processing, we limit the extraction to 8 minutes.
from skcycling.extraction import activity_power_profile
power_profile = activity_power_profile(ride, '00:08:00')
print('The power-profile is:\n {}'.format(power_profile))
# Authors: Guillaume Lemaitre <*****@*****.**>
# Cedric Lemaitre
# License: BSD 3 clause
from os.path import dirname, join
import pytest
from skcycling.datasets import load_fit
from skcycling.utils import validate_filenames
filenames = load_fit()
@pytest.mark.parametrize("filenames, expected_filenames",
[(filenames, filenames),
(join(dirname(filenames[0]), '*.fit'), filenames)])
def test_validate_filenames(filenames, expected_filenames):
assert list(validate_filenames(filenames)) == expected_filenames
:func:`skcycling.io.bikeread` using pandas.
"""
# Authors: Guillaume Lemaitre <*****@*****.**>
# License: BSD 3 clause
print(__doc__)
###############################################################################
# `scikit-cycling` has couple of `fit` files stored which can be used as toy
# data.
from skcycling.datasets import load_fit
filename_fit = load_fit()[0] # catch the first toy file
print('The fit file which will be used is stored at: \n {}'
.format(filename_fit))
###############################################################################
# The function :func:`skcycling.io.bikeread` allows to read the file without
# any extra information regarding the format.
from skcycling.io import bikeread
ride = bikeread(filename_fit, drop_nan='columns')
print('The ride is the following:\n {}'.format(ride.head()))
###############################################################################
# First, we can list the type of data available in the DataFrame
def test_check_filename_fit():
filename = load_fit()[0]
my_filename = check_filename_fit(filename)
assert my_filename == filename