Mielina can be described as either a framework or a foundation to build your chatbot upon. It runs on NodeJs and features an opinionated set of patterns and tools to build a multi-platform chatbot.
It implements abstractions to describe the system in a reasonably intuitive way by representing its functions as simulations of the functions performed by the human nervous system during a conversation coordinating the bot actions by transmitting signals to the relevant part of the super-system.
A super-system is just a large system integrated by many smaller systems or sub-systems. A modern chatbot is a super-system as it usually depends on one or more i/o platforms (messaging platforms in this case), one or more NLP systems, fulfillment sub-systems which in many cases are super-systems in their own merit.
Mielina is a solution to connect and coordinate all the components in a chatbot super-system.
Check the reference-implementation package for a simple (evolving) example of
how to get an NLP powered console bot running.
I am currently documenting that code but for now the most confusing part is the
env variables as they are mailny composed of api access credentials.
More docs on this coming soon but for now check the .env.example file for hints.
Writing a bot with my kids to help them get better at playing LoL I realized that the abstractions available for the task were too complicated for non developers to understand, so I challenged myself to write something that could be abstracted into concepts more relatable to our everyday communications experience. This is how bot-annah was born.
Mielina is an evolution of an older experiment called Bot-annah. She is fully operational, under light maintenance and open source.
MIT