Raku for Prediction
summary
A system that parses constrained natural sentences into runnable code for R, Python and Mathematica using Raku grammars and DSLs, running a Cro web service on DigitalOcean with ZeroMQ plumbing and a Raku Ripple sandbox for safe evaluation. It includes multi-language DSLs, a lingua numeric word forms utility for spoken numbers, and leverages monarch pipelines to model data-wrangling flows. The translator emits code but does not itself execute full workflows, uses randomized tests for validation, and claims it covers roughly 60–80% of typical data‑science tasks.
AI-written summary · about
chapters (8)
- 0:00Introduction to Raku for prediction
- 10:15Why natural language and grammars
- 20:07System architecture and design
- 26:48Data wrangling workflow example
- 31:01Monarch pipelines and testing
- 35:15Unix philosophy and alternatives
- 38:06Stochastic methods and future plans
- 43:53Live demo and Q&A
related talks
29mHow to automate the deployment of ...
17mR & Python playing nice, in Production
33mSyntax Isn't Everything: NLP for Rubyists
26mAdvanced Document Parsing for AI Workflows
59mMaking the most of artificial and human intelligence for data science
16mEmpower your organization with production-ready RAG apps on Azure
27mBuilding a Modular RAG Pipeline in Python
42mScala: The Unpredicted Lingua Franca for Data Science