firefly is an open source micro framework to deploy Python functions as web services. firefly was created with the aim of simplifying the deployment Machine Learning models as RESTful API. But as fate would have it, it became our favorite tool for building microservices. firefly takes care of processing the HTTP requests, forwarding the data to the python functions and encoding the result back to an HTTP request. It also has data validation, authentication support, transferring any type of file. You can also define a configuration file specifying the URL resource structure resulting in an elegant RESTful API.
It comes with a client library that makes calling remote functions calling as easy as calling functions present locally. It is a WSGI application and can be deployed and scaled through any WSGI server like gunicorn. There are many other features in the pipeline like multiple authentication modes through the plugin system.
This talk will focus on introducing firefly, it’s notable features like plugin support, building microservices efficiently with various examples.
Share this session: