REX

A lightweight domain-specific language for high-performance scientific research modeling. Transpiles .rex source files to native Zig code — producing small (<85KB), fast binaries with zero runtime overhead.

---

Features

• 7 scientific domains — quantum chaos, spectral graph theory, random matrix theory, chemistry, genomics, fluid dynamics, thermal physics
• Native Zig output — compiles to small, fast binaries with no garbage collector
• Lab-ready — CSV export, DAQ piping, embedded no_std support
• Zero dependencies — transpiler is pure Python, runtime is a single Zig file

---

Quick Start

# Transpile a Rex model
python rex_transpiler.py examples/example_chemistry.rex

# Compile and run (requires Zig 0.15.2)
zig run examples/example_chemistry.zig

---

Domain Reference

Keyword	Domain	What it computes
magnetic_quantum_walk	Quantum chaos	CTQW return probability with magnetic flux
loschmidt_echo	Quantum chaos	Gaussian fidelity decay
participation_ratio	Quantum chaos	Inverse participation ratio
spacing_analysis	Quantum chaos	GOE level spacing ratio
chaos_stats	Quantum chaos	Brody parameter
laplacian_spectrum	Spectral graphs	Trace of Laplacian
fiedler_value	Spectral graphs	Algebraic connectivity
eigenvalue_stats	Spectral graphs	Mean level spacing
quantum_walk	Spectral graphs	Bessel approximation on cycle
rmt	Random matrices	GOE Wigner surmise spacing
spectrum	Random matrices	Spectral radius estimate
residue	Complex analysis	Residue at simple pole
contour_integral	Complex analysis	Contour integral value
molecule_energy	Chemistry	Hückel MO pi energy
reaction_rate	Chemistry	Arrhenius rate at 298K
quantum_chemistry_hamiltonian	Chemistry	H2 STO-3G ground state
dna_sequence	Genomics	GC content
crispr_offtarget	Genomics	Off-target probability
gene_expression	Genomics	Fold change
population_genetics	Genomics	Neutral drift fixation
vector_titer	Gene therapy	AAV9 Poisson titer
delivery_efficiency	Gene therapy	Multi-compartment efficiency
heat_equation_1d	Fluid dynamics	FTCS finite difference
burgers_1d	Fluid dynamics	Upwind scheme L2 norm
navier_stokes_2d_simple	Fluid dynamics	Kinetic energy decay
turbulence_stats	Fluid dynamics	Kolmogorov -5/3 law
thermo_efficiency	Thermodynamics	Carnot efficiency
phase_transition	Thermodynamics	Mean-field order parameter
validate	Lab	Validation check
export_csv	Lab	CSV file export
lab_daq_output	Lab	DAQ instrument piping
embedded_no_std	Lab	no_std hardened binary

---

Language Reference

// Graphs
graph g = (undirected, [(0,1),(1,2),(2,0)]);
laplacian_spectrum g;
fiedler_value g;

// Quantum walks
magnetic_quantum_walk ctqw on g time 10 flux 0.314;
loschmidt_echo g;

// Random matrices
rmt ensemble size 100;

// Chemistry
molecule_energy benzene using huckel;
reaction_rate arrhenius;

// Genomics
dna_sequence brca1 = "ATGCGT";
crispr_offtarget brca1 mismatches 2 pam "NGG";

// Fluids
heat_equation_1d rod length 1.0 dt 0.01 steps 100;
burgers_1d shock velocity 1.0 viscosity 0.01;

// Lab output
export_csv all_results to "results.csv";
lab_daq_output "data.bin" to stdout;

---

Installation

Requirements: Python 3.8+, Zig 0.15.2

git clone https://github.com/Troq20909/rex.git
cd rex
python rex_transpiler.py examples/example_full.rex
zig run examples/example_full.zig

---

Project Structure

rex/
├── rex_transpiler.py   # Rex → Zig transpiler (pure Python)
├── rex_runtime.zig     # Runtime library (35 functions)
├── build.zig           # Zig build system integration
├── CONTRIBUTING.md     # How to add new keywords
├── LICENSE             # MIT
└── examples/
    ├── example_chemistry.rex
    ├── example_fluids.rex
    ├── example_full.rex
    ├── example_genomics.rex
    ├── example_lab.rex
    ├── example_quantum.rex
    └── example_rmt.rex

---

Known Limitations

• Function bodies support only let and print statements
• Domain keywords must appear at top level
• Comments are single-line only (//)
• Zig 0.15.2 required

---

Roadmap

• Type inference
• Multi-file models
• WebAssembly target
• Jupyter notebook integration
• More domain keywords

---

License

MIT — use, modify, distribute freely for research.