strava-mongo-lambda

In particular, I like to engage in virtual cycling on the Zwift platform. Which then saves my indoor cycling activities to Strava as VirtualRide activities.

This project exposes a FastAPI tiny webserver API behind an AWS lambda funtion URL which is protected with OAuth2 via Cognito.

The API allows me to securely trigger a sync of my Strava data to a Google Sheet for further analysis.

As an intermediate, this sync will save a copy of the Strava activities to a free tier of Mongo Atlas, giving me the option to pivot away from Google Sheets later.

Local Development

invoke dev

Deployment

Setup

You will need to define credential details in .env and also log into ECR registry. You will also need Docker installed and active to build lambda container image

invoke ecr-login

Build Artifact

invoke build-lambda-container deploy-lambda-container

Go Live

This part is still clickops to make the lambda point at the latest uploaded image. Room for automation improvement of course.

Maintenance

Strava SDK

[OPTIONAL] Generate Strava SDK using Swagger Codegen

Step 1 - Generate SDK

# Using OAPIv3 codegen
brew install swagger-codegen
swagger-codegen generate -i https://developers.strava.com/swagger/swagger.json -l python -o strava

This will create:

strava
├── README.md
├── docs
│   ├── ActivitiesApi.md
│   ├── ...
│   └── Zones.md
├── git_push.sh
├── requirements.txt
├── setup.py
├── swagger_client
│   ├── __init__.py
│   ├── api
│   │   ├── __init__.py
│   │   ├── activities_api.py
│   │   ├── .....
│   │   └── uploads_api.py
│   ├── api_client.py
│   ├── configuration.py
│   ├── models
│   │   ├── __init__.py
│   │   ├── activities_body.py
│   │   ├── ...
│   │   └── zones.py
│   └── rest.py
├── test
│   ├── __init__.py
│   ├── test_activities_api.py
│   ├── ...
│   └── test_zones.py
├── test-requirements.txt
└── tox.ini

Step 2 - Update app/core/strava.py

For what I needed I decided that using the swagger generated code was overkill.

app/core/strava.py is a httpx implementation of the raw API for just the requests I needed.

In particular I only needed the /athlete/activities endpoint.

So just check the filtered attributes that are grabbed are still the subset we want and there are no typos, renames, missing attributes or new attributes we wish to include.