strava-oauth

Lightweight python server that implements Strava Oauth Web Flow

This repo comes with pre-configured Dockerfile and docker-compose.yml, which allows direct deployment by simply running:

make build

Must

Before using set environmental variables that represent your application:

export STRAVA_CLIENT_ID=<the-actual-id>
export STRAVA_CLIENT_SECRET=<the-actual-secret>

Quick start

git clone https://github.com/sladkovm/strava-oauth.git
cd strava-oauth
pipenv shell
pipenv install --pre

Run the server:

python api.py

In the web browser go to: http://127.0.0.1:5042/authorize

1. You will be redirected to the Strava authorization page
2. After authorization is granted, the browser will display a raw JSON with the authorization tokens
3. Take it from here and build on it

Example Response
{
  "token_type": "Bearer",
  "access_token": "987654321234567898765432123456789",
  "athlete": {
    #{summary athlete representation}
  },
  "refresh_token": "1234567898765432112345678987654321",
  "expires_at": 1531378346,
  "state": "https://github.com/sladkovm/strava-oauth"
}

Run as a dockerized application

Before you start, set the application url, where the Strava callbacks will be redirected:

export APP_URL=<http://myapp.com>

In this case the callbacks will be redirected to http://myapp.com:5042/authorization_successful

Some gotchas:

1. If you run docker directly from the terminal (e.g. not in the docker-machine), the docker containers will be spawned at the localhost. In this case you don't need to specify the APP_URL. The server will be run as in the Quick start case - at http://127.0.0.1:5042/authorize

2. If you run docker on the docker-machine, with a given IP, you will need to figure out how to redirect to this IP. The easiest way is to set export APP_URL=dev.myapp.com. Then add to the /etc/hosts the following line: <docker-machine IP> dev.myapp.com. This will send all Strava redirects to the docker-machine IP

Friendly warning

This server is built on python-responder, which is an awesome Python API Framework for Humans(TM), but is in active development phase ... that is why I use --pre and --skip-lock for example.