Ruby NLP C++ Grammar #BNF

At FriendlyData we had a Ruby-based pipeline for natural language processing. Our technology is centered around grammar-based natural language parsing, as well as various product features, and, as the core stack of the company historically is Ruby, the initial version of the pipeline was implemented in Ruby as well.

As we were entering the exponential growth phase, both technology- and product-wise, we looked into how could we speed up and extend the performance and flexibility of our [meta-]BNF-based parsing engine. Gradually, we built the pieces of the engine in C++.

Ultimately, the natural language parsing stack spans three universes and three software engineering paradigms: the declarative one, the functional one, and the imperative one. The imperative one was and remains implemented in Ruby, the functional one is implemented in a functional language (this part is under the NDA, while everything I am talking about here is part of the public talks we gave throughout 2017 and 2018), and the declarative part, which can loosely be thought of as being BNF-based, is now served by the C++ engine.

The C++ engine for the BNF part removed the immediate blockers, gave us 500x+ performance speedup, and enabled us to launch new product features, most notably query completions, suggestions, and spelling corrections.

Our first experience with .NET core was when we developed our OSS feature management platform - Tweek (https://github.com/soluto/tweek). We wanted to create a solution that is able to run anywhere (super important for OSS), has excellent performance characteristics and can fit in a multi-container architecture. We decided to implement our rule engine processor in F# , our main service was implemented in C# and other components were built using JavaScript / TypeScript and Go.

Visual Studio Code worked really well for us as well, it worked well with all our polyglot services and the .Net core integration had great cross-platform developer experience (to be fair, F# was a bit trickier) - actually, each of our team members used a different OS (Ubuntu, macos, windows). Our production deployment ran for a time on Docker Swarm until we've decided to adopt Kubernetes with almost seamless migration process.

After our positive experience of running .Net core workloads in containers and developing Tweek's .Net services on non-windows machines, C# had gained back some of its popularity (originally lost to Node.js), and other teams have been using it for developing microservices, k8s sidecars (like https://github.com/Soluto/airbag), cli tools, serverless functions and other projects...

I've used .NET for many years, but only in recent years, after Microsoft introduced .NET Core, I've found a new love and excitement for the technology again. The main driver for us using .NET Core is not that it is cross platform compatible, open source or blazingly fast (which it is!), but the fact that we can use (what we consider) the best programming languages (mainly F# and C#) to carry out our jobs without sacrificing the other benefits.

Today we run most of our web infrastructure on .NET Core in Docker containers, deployed into a Kubernetes cluster which spans across multiple time zones in the Google Cloud and we couldn't be happier. Due to the portability of the .NET Core platform we are even able to develop many new services as serverless functions with F# which has become an absolute game changer.