vmpy's Introduction

VMPY - Performance Velo Metrics Toolbox

VMPY is a toolbox for evaluating typical Cycling Performance Metrics from the ride data e.g. power, heart-rate, velocity, gradient, cadence streams.

All functions within the package do follow the convention, where input/output formats are either traditional Python built-in data structures or are the nd-arrays. This design choice favors easy integration into other projects, e.g. velometria.com

To help you getting started VMPY also provides a very thin wrapper around the Strava API.

Installation

Official release:

pip install vmpy

The bleeding edge work in progress:

pip install git+git://github.com/sladkovm/vmpy.git@development

Scope

The VMPY package provides the following functionality:

streams.py: Streams shape preserving calculations e.g. masking, filtering, zone conversions

metrics.py: Cycling Performance Metrics

strava.py: Python wrapper around the Strava API v3 for fetching Athletes, Activities and Stream data

Usage

>>> from vmpy import strava
>>> stream = strava.retrieve_streams(activity_id=1282167861, access_token=STRAVA_ACCESS_TOKEN)

>>> from vmpy import streams
>>> power_zones = streams.compute_zones(stream['watts'], ftp=270)
>>> hr_zones = streams.compute_zones(stream['heartrate'], lthr=160)
>>> watts_3sec = streams.rolling_mean(stream['watts'], window=3, mask=stream['moving'])
>>> gradient_wo_outliers = streams.median_filter(stream['grade_smooth'], window=31, threshold=1)

>>> from vmpy import metrics
>>> normalizes_power = metrics.normalized_power(stream['watts'])
>>> time_in_power_zones = metrics.time_in_zones(stream['watts'], ftp=260)

Quick Start

Register Strava App

In order to be able to use Strava API the user App must be registered at the linkhttp://www.strava.com/developers:

Application Name: ex.: My Awesome App

Website: ex.: myawesomeapp.com (can be anything, even your FB or Strava page will do)

Application Description ex.: Just fooling around with some Strava data

Authorization Callback Domain 127.0.0.1 (unless you are building a serious App)

Access Token

The Access Token will be found at the users profile page It will look like this: 83ebeabdec09f6670863766f792ead24d61fe3f9

Access Token must be passed explicitly as an argument to the functions found in strava.py module

Contribution guidelines

This project is absolutely open for contributions. No strong guidelines yet, except for:

Don't push on master branch
Test
Write docstrings in NumPy style

Useful links

Strava API v3 documentation - https://strava.github.io/api/

vmpy's People

Contributors

Stargazers

Watchers

vmpy's Issues

Push changes to the repo

Tidy app activity Strava plot method

Top panel:

Elevation profile - shaded
Speed - line
cadence - bar chart

Middle panel:

Elevation profile
Power profile

Lowest chart:

Elevation profile
Heartbeat profile

Provide choice of X axis - distance/time

Set up ranges for axes:
X axis is tight

Reduce number of ticks on the Y axis

Add some totals next the the Y axis:

Speed - max, median, average
Power - max, NP, average
Heartbeat - max, median, average

Add before and after parameters to the athlete.retrieve_list_athlete_activities_json

Add logging to the code

Update README.md

Move athletic_pandas algorithms to the vmpy

Move algos from https://github.com/AartGoossens/athletic_pandas to vmpy
Add vmpy as a dependency for athletic pandas

Automate streams shift estimate

Algorithms to implement:

Cross-correlation analysis - shift would correspond to maximum correlation value
Use DTW - the mean shift in the warp path is a good measure of optimal shift

use on power - 1D
use on latlng - 2D
Try python version of imregcorr - essentially the same as cross-correlation

sladkovm / vmpy Goto Github PK