Installation

Requirements

Before you begin, ensure you have the following installed:

• C++23 compatible compiler — GCC 13+ or Clang 16+
• Make — For building the project
• Git — For cloning the repository
• Linux x86-64 — Currently the primary supported platform

Clone the Repository

First, clone the Nevaarize repository:

git clone https://github.com/gtkrshnaaa/nevaarize.git
cd nevaarize

Build from Source

Nevaarize uses a simple Makefile for building. Run:

make

This will compile all source files and create the nevaarize binary in the bin/ directory.

Verify Installation

Test that the build was successful:

./bin/nevaarize --version

You should see output like:

Nevaarize v0.2.3
Native JIT Compiler for the Nevaarize Programming Language
Built with C++23, Zero External Dependencies

Run Your First Program

Try running the hello world example:

./bin/nevaarize examples/basics.nva

Output (excerpt):

Integer: 42
Float: 3.14
...
add(7, 3): 10
...
=== All Features Demonstrated ===

Optional: Add to PATH

For convenience, add Nevaarize to your system PATH:

# Add to ~/.bashrc or ~/.zshrc
export PATH="$PATH:/path/to/nevaarize/bin"

# Reload your shell
source ~/.bashrc

Now you can run nevaarize from anywhere:

nevaarize script.nva

Project Structure

Here's an overview of the Nevaarize directory structure:

nevaarize/
├── bin/              # Compiled binary
├── build/            # Object files
├── core/             # Core compiler implementation
│   ├── include/      # Header files
│   ├── src/          # Source files
│   └── stdlib/       # Standard library (inside core)
├── examples/         # Example programs (64 verified)
│   ├── algorithm/    # Sorting, searching, math
│   ├── benchmarks/   # Performance benchmarks
│   ├── modules/      # Module system demos
│   ├── richcodesample/ # Advanced code samples
│   └── stdlibtest/   # Stdlib module tests
├── docs/             # Documentation (you are here)
└── Makefile          # Build configuration

Model Commands

Nevaarize includes CLI commands for AI model training and inference:

Train a Model

Run a training script and save the resulting model:

./bin/nevaarize model train script.nva to model.nmod

The output path is resolved relative to the training script's location.

Run Model Inference

Load a trained model and run inference:

# View model info
./bin/nevaarize model run model.nmod

# Run with input data
./bin/nevaarize model run model.nmod --input "[1.0, 0.0, 1.0, 0.0]"