Copyright (c) 2021. Sergei Sazonov. All Rights Reserved

HARDIO is a free Performance Management tool for cyclists (and potentially runners in future) with emphasis on analysing cardiovascular and metabolic effects of training with power.

Besides the Software Development Learning purpose (more on that below) in my cycling training journey I needed a simple way to analyse and visualize progress of my ability to recover after various training efforts. I'm talking mainly about post-interval Heart Rate (HR) recovery and HR drift spanning across weeks and month of training. More ideas to come.

HARDIO is deployed on a free Heroku Dyno and is available for use (may take a while to load if Dyno has fallen asleep). You can either create your account and authorize application to get your cycling rides data from Strava or play with test dataset based on my own rides.

Main emphasis of the project is a "close-to-real-world" software development experience as part of my Open Source Society University journey.

Purpose: Gain experience of building a web application by doing along with learning and exercise different concepts, techniques and frameworks. Here is the ever-growing list of things:

• Python
  • OOP
  • Abstract Classes / Interfaces
  • Dataclasses
  • Pydantic
  • Type hinting
  • Pandas basics
  • Package and requirements management
  • requests lib
• Design Patterns in general, including following implemented in the application:
  • Abstract Factory
  • Strategy
  • Observer (tbd)
• git basics
• Unit and Functional testing (Pytest)
• Flask
  • SQLAlchemy basics
  • Flask migrate
  • Basics of html, css, jinja2, Bootstrap
• Plotly/Dash
• SwaggerClient
• Decorators
• Web Hooks (tbd)
• Packaging and deployment (Heroku)

Next Steps:

• Design/implement automatic interval detection
• Implement post-interval recovery detection
• Implement long-term analytics of post-interval recovery rate, HR, decoupling
• Implement in-memory database (likely redis) to improve speed of stateless Plotly/Dash
• Implement async activities retrieval from Strava (most likely rabbitMQ brocker + async fastapi workers)
• Reconsider storage strategy: store activity json objects on file-system instead of Postgress
• Implement CTL and CTL Chart

1. Set environment variables (heroku config):

   export SECRET_KEY='your long key'

   export SQLALCHEMY_DATABASE_URI='sqlite:///site.db' or 'potrgresql://...'

   export EMAIL_USER="your@email"

   export EMAIL_PASS="your_pass"

   export STRAVA_AUTH_URL="https://www.strava.com/oauth/authorize"

   export STRAVA_APP_CLIENT_ID="your_strava_app_id"

   export STRAVA_APP_CLIENT_SECRET='your_strava_app_client_secret'

   export STRAVA_APP_REDIRECT_URI="http:<HOSTNAME>/exchange_token"

   export FLASK_APP="hardio:create_app()"

2. git push heroku master