Open Source Surf Report

Developer Setup (Backend)

Disclaimer: You're likely going to need some env variables to get things running correctly. Feel free to contact me at [email protected] and I'd be more than happy to help you get setup.

1. Fork the repo.
2. Clone the forked repo: git clone [email protected]:<your_github_username>/surf-detector.git
3. Enter correct directory: cd surf-detector
4. Choose what to work on: cd data OR cd models OR cd server
5. Follow instructions below depending on what section your working on.

Data

6. Create a virtual environment: python -m venv env
7. Install Requirements: pip install -r requirements.txt
8. cd into source directory and run: cd src AND python main.py

Models

6. No developer setup needed yet.

Server

6. Create a virtual environment: python -m venv env
7. Install Requirements: pip install -r requirements.txt
8. Run locally: flask run

Developer Setup (Frontend)

Disclaimer: You're likely going to need some env variables to get things running correctly. Feel free to contact me at [email protected] and I'd be more than happy to help you get setup.

1. Fork the repo.
2. Clone the forked repo: git clone [email protected]:<your_github_username>/surf-detector.git
3. Enter correct directory: cd surf-detector AND cd frontend AND cd app
4. Install Requirements: npm install
5. Run locally: npm run dev

TODO (Backend)

• Implement object detection model that can detect how many surfers are in water.
• Configure cloud db (CockroachDB, Supabase)
• Fetch swell and weather data from NOAA and relative databases, given specific location.
• Define db methods.
• Add NOAA buoy data (SCRIPPS) to the db.
• Deploy scheduled tasks using AWS ECS and ECR.
• Grab historical NOAA and SL data to train prediction models more quickly.
• Create AI forecast using GPT API (3 day forecast and 14 day forecast).
• Configure data collection from bouys other than SCRIPPS.
• Train crowd prediction model.
• Train height prediction model. (given actual buoy reading, predict wave height)
• Train buoy prediction model (given gfs reading, predict actual buoy reading)
• Update root README for developer setup.
• Add cam URL dictionary to .env.
• Adding data for every spot appears to not work on AWS ECS. Figure out a different way.
• Configure a local setup that works without the need for env variables.

TODO (Frontend)

• Add chart to represent wave height.
• Chart to represent tides and current crowd.
• Add point to tide chart showing current time.
• Style tide chart to better represent mouse hover position.
• Add Icons to navbar.
• Update root README for developer setup.
• Handle Mobile Web Version for spot forecast.
• Landing Page
• Supabase confirmation link page.
• Fix bug where swell chart renders multiple times. Usually occurs when spot is spam clicked.
• Find a way to create custom scroll bar for day forecasts.
• Set minimum of tide chart to -1.5 OR to the calculated minimum of all the possible datapoints (As was the height of swell chart).
• Make mobile Compatible
• Box that shows current tide is not centering text vertically.
• Change style of nabvbar on mobile screens (design is on figma).
• Migrate vanilla css to Tailwind (low priority).
• Configure a local setup that works without the need for env variables.
• Make Home page footer work
• Home page login on hover should be pointer