feat(recipes): add Turing Bayesian probabilistic programming library

[jordanrule]

Overview

This PR adds the Turing library to the Modular community channel. Turing is a focused Mojo library for Bayesian probabilistic programming with HKT-style type safety and practical inference engine paths.

Package Details

• Name: Turing
• Version: 0.1.0
• Repository: https://github.com/TuringCore/turing
• License: BSD-3-Clause
• Maintainer: @jordanrule

Key Features

• HKT-inspired type design: Prevents invalid signal composition at compile time through semantic type tagging
• Multiple inference paths: MAP, Metropolis-Hastings, Sequential Monte Carlo, and Variational Inference
• Mojo compile-system integration: Target-agnostic model semantics with explicit CPU/GPU lowering behavior
• Practical applications: Finance (risk scoring), Healthcare (triage), Operations (forecasting), and Recommendations (feed ranking)

What's Included

• Core library: src/hkt_probprog.mojo with HKT-inspired inference engine
• Examples: Social recommendation demo showing library usage with Bluesky
• Tests: Comprehensive tests for type safety and algorithm correctness

Checklist

• recipe.yaml specifies MAX/Mojo version compatibility (compatible with mojo-compiler, no version constraint)
• License file is packaged (LICENSE file present in repository)
• Build number set to 0 (first submission)
• Package structure follows Modular community standards

Related Links

• Library Documentation: https://github.com/TuringCore/turing#readme
• Example Application: https://github.com/TuringCore/bluesky (Bluesky social media recommendation demo)
• Recipe Reference: /recipes/turing/recipe.yaml

Testing

The recipe includes automated tests that verify:

1. HKT type system correctness (test_hkt_demo.mojo)
2. Inference engine path behavior across all algorithm variants (test_inference_paths.mojo)

Logo

Logo included as image.jpeg - purple infinity symbol representing probabilistic computing and Bayesian inference.

[jordanrule]

hi @carolinefrasca - this should be ready to re-run in CI. apologies for delay, learning process on my end!