Simple flask app

[benbovy]

Closes #1.

I added a notebook in the examples folder as a demo of using the API from the client side.

• I directly addressed point 2 since sending GeoJSON formatted track coordinates in the query was pretty straightforward.
• I also considered point 3 when setting API routes. In addition to choosing the dataset, there is a (already implemented) query parameter for choosing a particular field (data variable).
• I added one more feature that allows providing from the client side some transform operations to be applied on the dataset before extracting the data (e.g., averaging over time).

TODO:

• [x] update Dockerfile (maybe use docker-compose to manage both the xpublish server in the backend and the flask application)
• [x] update README

[benbovy]

@alirezamdv you can have a look at an API example usage here.

I'm open to suggestions.

[willirath]

This is great!

[benbovy]

I slightly rearranged the code. Docker-compose is used to build two services running in two containers:

• the xpublish server for the demo dataset (internal use only)
• the main Flask application app that will be used from the front-end

I think it is enough for now. We'll need to figure out how to expand this architecture to support multiple datasets. I'm not sure running one container per dataset is a good solution. It would be nice if xpublish could support serving a collection of datasets.

[benbovy]

I was fixing the json response of ship tracks (not properly formatted) and also adding the possibility of setting multiple fieldnames to extract data from several data variables at once.

I will open a new PR :-)

[willirath]

I will open a new PR :-)

Sorry. :)