ollama_boolean

A console application that uses Ollama to classify text prompts into binary or ternary responses.

Quick Start

Download Pre-built Binaries (Recommended)

📥 Download Latest Release

Supported Platforms:

• Linux (x64): ollama_boolean-linux-x64
• Windows (x64): ollama_boolean-windows-x64.exe
• macOS Intel: ollama_boolean-macos-x64
• macOS Apple Silicon: ollama_boolean-macos-arm64

Quick Install (Linux/macOS):

# One-line installer
curl -sSL https://raw.githubusercontent.com/codecaine-zz/ollama_boolean/main/install.sh | bash

# Or manual download
curl -L -o ollama_boolean https://github.com/codecaine-zz/ollama_boolean/releases/latest/download/ollama_boolean-linux-x64
chmod +x ollama_boolean
sudo mv ollama_boolean /usr/local/bin/

Windows:

# Download from releases page
Invoke-WebRequest -Uri "https://github.com/codecaine-zz/ollama_boolean/releases/latest/download/ollama_boolean-windows-x64.exe" -OutFile "ollama_boolean.exe"

Or Build from Source

Installation

To install dependencies:

bun install

Compilation to Binary

You can compile the script to a standalone binary executable:

# Using npm script (recommended)
bun run build

# Or directly with bun build
bun build --compile --outfile ollama_boolean index.ts

# This creates an executable named 'ollama_boolean' in the current directory

Cross-Platform Compilation

Build for different platforms:

# Build for current platform
bun run build

# Build for Windows
bun run build:win

# Build for Linux
bun run build:linux

# Build for macOS
bun run build:mac

📖 For more details on cross-platform compilation with Bun, see: Bun Build Documentation

Using the Compiled Binary

Once compiled, you can run the binary directly without needing Bun installed:

# Verbose mode
./ollama_boolean "Is the sky blue?"

# Quiet mode
./ollama_boolean --quiet "Can pigs fly?"

# With custom model
./ollama_boolean "Is this art beautiful?" llama3.2

# Show help
./ollama_boolean --help

Distribution

The compiled binary is self-contained and can be:

• Copied to other machines with the same architecture
• Added to your PATH for global access
• Distributed without requiring Bun or Node.js installation

# Example: Copy to /usr/local/bin for global access (macOS/Linux)
sudo cp ollama_boolean /usr/local/bin/

GUI Integration

For rapid application development (RAD) with a graphical user interface, you can integrate the compiled binary with Neutralino.js:

• Subprocess Integration: Call the ollama_boolean binary as a subprocess from your Neutralino.js app
• IPC Communication: Use Inter-Process Communication to exchange data between the GUI and the CLI tool
• Cross-Platform: Deploy your GUI app across Windows, macOS, and Linux using Neutralino's lightweight framework

This allows you to create a modern, cross-platform desktop application while leveraging the existing CLI functionality.

Usage

The app classifies user prompts and returns:

• 1 for Yes/Positive answers
• 0 for No/Negative answers
• 2 for Unknown/Subjective answers

Basic Usage

bun run index.ts "<your_prompt>"

With Custom Model

bun run index.ts "<your_prompt>" <model_name>

Quiet Mode (Number Only)

For scripts or when you only need the raw classification number:

bun run index.ts --quiet "<your_prompt>"
bun run index.ts -q "<your_prompt>" <model_name>

Examples

# Verbose output (default)
bun run index.ts "Is the sky blue?"
# Output:
# Prompt: "Is the sky blue?"
# Model: qwen3
# Classifying...
# 
# Result: 1
# Interpretation: Yes/Positive

# Quiet mode - only returns the number
bun run index.ts --quiet "Is the sky blue?"
# Output: 1

# Using a specific model with quiet mode
bun run index.ts -q "Can pigs fly?" llama3.2
# Output: 0

# Subjective question
bun run index.ts "Is this painting beautiful?"
# Output: 2 (with full interpretation in verbose mode)

Help

bun run index.ts --help

Requirements

• Bun runtime
• Ollama running locally on port 11434
• At least one Ollama model installed (default: qwen3)

Model Setup

Make sure you have Ollama running and the desired model installed:

# Install and run Ollama
ollama pull qwen3
ollama serve

This project was created using bun init in bun v1.2.19.

Testing

The project includes comprehensive Bun tests covering:

Run Tests

# Run all tests
bun test

# Run tests with a timeout of 3000ms
bun test --timeout 3000

# Run tests in watch mode (re-run on file changes)
bun run test:watch

# Run tests with coverage report
bun run test:coverage

Test Coverage

The test suite covers:

• CLI Interface: Help commands, argument parsing, error handling
• Input Validation: Empty prompts, special characters, edge cases
• Response Validation: JSON parsing, result validation, error scenarios
• Performance: Argument parsing efficiency, long prompts
• Binary Compilation: Successful compilation verification

Test Categories

• Unit Tests: Core logic and validation functions
• Integration Tests: CLI interface and argument handling
• Performance Tests: Efficiency and edge case handling
• Compilation Tests: Binary build verification

The tests are designed to run without requiring Ollama to be running, using mocked responses and logic validation where appropriate.