Terrain development in Minecraft relies on large, deeply nested JSON structures that quickly become difficult to read and maintain. Rhombus introduces powerful abstractions and a comprehensive Python API, making world generation code cleaner, more modular, and significantly faster to develop.

Rhombus is an eDSL delivered as a Python package that allows worldgen developers to comfortably write expressions resembling the abstract syntax trees of density functions for Minecraft while benefiting from Python's forgiving syntax.

_{This project is pretty similar to misode/gaia-beet, which you might also find useful. Know that I started developing Rhombus before I knew about it, the similarities in concept are quite frightening though. The biggest difference to Misode's gaia-beet is that Rhombus does not see itself as a mere Beet plugin — although we heavily rely on it — but rather as a coherent, mostly separated language and it will be further developed and improved in exactly this sense.}

Key Advantages

• 📦 Object-oriented Design
  Full use of Python’s object model for composing, reusing, and structuring density functions.
• 📝 Native Comments
  Since Python code is in use, comments work naturally without any custom syntax.
• 📖 Integrated Documentation
  Functions and classes provide docstrings describing behavior, parameters, and usage.
• ⚡ Performance Optimization
  Recurring expressions and resource intense operations can be cached automatically.
• 🗄️ Macro Library
  Growing collection of macros for common patterns and complex operations, reducing boilerplate and improving readability.

Features

Unveil

This list only includes features that are principially not neccessary, but make up the real strengths of Rhombus.

• Rhombus Preview
  A lightweight local frontend with file-watching for previewing density functions
  • Visualizer (built on deepslate)
    • Supports all major worldgen mods (see Mod Support)
  • File Explorer & Viewer (built on monaco)
  • Spline Visualization
• Macro Library
  • General Math
    • sum, prod
    • Infinity, NaN
    • round, floor, ceil, mod
    • sgn, heaviside, monus, ramp
  • Spline approximations
    • Sampler for any Python function
    • sin, cos, tan, atan, tanh, coth
    • exp (with arbitrary base)
    • smoothstep, normalPDF, normalCDF, erf, logistic
  • Performance optimization
    • evaluate the number of unique nodes of a function
    • automatically cache specific sub functions of a function
    • automatically cache all recurring sub functions of a function
  • Expensive iterative Methods
    • sqrt
    • sin, cos, tan, exp, ln
    • round, ceil, floor, mod
  • Working with Maps
    • extrude_heightmap
  • Fluent Interface for conditionality
  • Coordinate reconstruction
• Mod Support
  • More Density Functions
  • Lithostitched
  • Tectonic
  • En-sityFunctions

Impressions

pip install rhombus

from rhombus import *

continent_noise = Noise(-10, [2, 1, 2, 2, 2, 1, 1, 1, 1])
erosion_noise   = Noise(-9, [3.5, 0, 2, 4, 2, 2, 3])

erosion = clamp(noise(erosion_noise, xz_scale=1.3, y_scale=0) - 0.1, min=-1, max=1)

height_map = spline(erosion, [
    (-1,    -1,     0),
    (-0.6,  -0.95,  0),
    (-0.61, -0.6,   0),
    (-0.2,  -0.55,  0),
    (-0.21, -0.2,   0),
    ( 0.05, -0.2,   0),
    ( 0.21,  0.2,   0),
    ( 0.6,   0.3,   0),
    ( 0.61,  0.8,   0),
    ( 1,     0.8,   0)
])

FINAL = maps.extrude_heightmap(height_map, (-1, 0.8), (64, 256))