FOWD

🌊 Processing framework for FOWD, a free ocean wave dataset, ready for your ML application.

Installation

After downloading the repository, you can install FOWD and all dependencies (preferably into a virtual environment) via:

$ pip install -r requirements.txt
$ pip install .

Usage

After installing the Python code, you can use the command line tool fowd to create a FOWD dataset from a raw source.

CDIP

The best supported source is CDIP buoy data:

$ fowd process-cdip 433p1 -o fowd-cdip-out

will process all CDIP data located in the 433p1 folder.

Datawell Waverider data

Raw data from the ubiquitous Datawell Waverider MK III (.raw and .spt files) can be processed into yearly netcdf files.

$ fowd prepare-buoy-data indir -o outdir

The output is then compatible with the main FOWD processing tool using:

$ fowd process-buoy-data indir -o outdir

Channel Coastal Observatory (CCO) download tool

One source of raw wave buoy data is the UK's Channel Coastal Observatory (CCO) real time data webpage (https://coastalmonitoring.org), this data can be programmatically downloaded using:

$ fowd download-cco out_folder station_metadata_path

Generic inputs

Use

$ fowd process-generic infile.nc -o outdir

Generic inputs must be netCDF files with the following structure:

Variables:
- time
- displacement

Attributes:
- sampling_rate
- water_depth
- longitude
- latitude

QC plots

All data processing writes QC information in JSON format. You can visualize records in that QC file by using

$ fowd plot-qc qcfile.json

Testing

Run tests and sanity checks via

$ fowd run-tests

Test results are checked automatically, but sanity checks have to be inspected manually.

Credits

This project is based on the original FOWD toolkit developed by Dion Häfner: https://github.com/dionhaefner/FOWD

This fork extends the toolkit with additional processing and data acquisition tools developed by Edward Roome.