Skip to content

README.md

Latest commit

 

History

History
205 lines (152 loc) · 4.97 KB

README.md

File metadata and controls

205 lines (152 loc) · 4.97 KB

Nonogram Solver (Picross/Griddlers Solver)

Rust License

A Rust application that solves nonogram puzzles (picross/griddlers/hanjie/logimage) through logical deduction. Analyzes an image, automatically extracts constraints, and outputs the solution in ASCII/JSON/Array format or generates a marked image.

🎉 What's New in v1.0.0

✅ Simplified Workflow: Image → Solution

  • Automatic constraint extraction from filled cells (no OCR needed!)
  • Multi-format export: ASCII (console), JSON, Array2D
  • Interactive mode for manual constraint entry
  • Simplified command: nonogram-solver image.png just works
  • Console output by default with solved grid
  • 100% pure Rust: No non-Rust dependencies for main features

🎯 Features

  • ✨ Automatic extraction: Analyzes filled cells to derive constraints
  • 🖼️ Multi-format support: JPEG, PNG, BMP, GIF, TIFF, WebP
  • 📤 Multi-format export: ASCII (█/✕/·), JSON structured, Array2D (1/0/-1)
  • 🎮 Interactive mode: Manual constraint entry in terminal
  • 🧩 Pure logical deduction: Solving algorithms without guessing
  • 🎨 Visual marking (optional): Generate image with deductions marked in red
  • 🏆 Three solvers: Basic, Advanced, Ultimate (backtracking + parallelization)

📋 Prerequisites

  • Rust 1.91 or higher
  • C compiler (gcc/clang) for native dependencies

🚀 Installation

# Clone the project
git clone https://github.com/your-username/nonogram-solver.git
cd nonogram-solver

# Build in release mode
cargo build --release

# Binary will be in target/release/nonogram-solver

📖 Usage

Simplest usage

# Automatic analysis + console ASCII output
./target/release/nonogram-solver image.png

Export the solution

# Export as JSON
./target/release/nonogram-solver -i image.png --export solution.json

# Export as 2D array
./target/release/nonogram-solver -i image.png --export solution.txt

# Export as ASCII explicitly
./target/release/nonogram-solver -i image.png --export solution.txt --export-format ascii

Constraint extraction modes

# Automatic extraction from filled cells (default)
./target/release/nonogram-solver -i image.png --extract-filled

# Interactive mode (manual entry)
./target/release/nonogram-solver -i image.png --interactive

# With JSON file (classic method)
./target/release/nonogram-solver -i image.png -c constraints.json

Main options

Option Description
-i, --input <FILE> Input nonogram image
-o, --output <FILE> Output image (optional)
-e, --export <FILE> Export grid to file
--export-format <FMT> Format: ascii, json, array
--extract-filled Auto-extract from filled cells
--interactive Interactive constraint input
--extract-only Extract and display without solving
--advanced Advanced solver
--ultimate Ultimate solver (backtracking + parallel)
-v, --verbose Verbose mode

Complete examples

# Auto extraction + ultimate solver + JSON export
./target/release/nonogram-solver \
  -i puzzle.png \
  --extract-filled \
  --ultimate \
  --export solution.json \
  --verbose

# Interactive mode with output image
./target/release/nonogram-solver \
  -i puzzle.png \
  --interactive \
  -o solution.png

# Extract only (no solving)
./target/release/nonogram-solver \
  -i puzzle.png \
  --extract-only \
  --export grid.txt

📤 Export Formats

ASCII (console)

✕ ✕ █ █ █
✕ ✕ █ █ █
█ █ █ █ █

JSON

{
  "width": 5,
  "height": 3,
  "cells": [
    ["crossed", "crossed", "filled", "filled", "filled"],
    ...
  ]
}

Array2D

[
  [-1, -1, 1, 1, 1],
  [-1, -1, 1, 1, 1],
  [1, 1, 1, 1, 1]
]

🏗️ Architecture

Core modules

  • grid/: Grid representation (CellState, Grid, Constraints)
  • solver/: Solving algorithms (NonogramSolver, AdvancedSolver, UltimateSolver)
  • image_parser/: Image analysis and grid detection
  • image_generator/: Image generation with markings
  • grid_output/: Multi-format export (ASCII, JSON, Array2D)
  • interactive/: Interactive constraint input
  • ocr (optional): Tesseract-based constraint extraction

Available solvers

  1. Basic: Line-by-line deduction with cache
  2. Advanced: Cross-analysis + advanced heuristics
  3. Ultimate: All techniques + backtracking + parallelization

🔧 Development

# Build
cargo build

# Run tests
cargo test

# Format code
cargo fmt

# Run linter
cargo clippy

# Release build
cargo build --release

📊 Performance

  • 5x5 grid: < 0.1 seconds
  • 10x10 grid: < 0.5 seconds
  • 20x20 grid: < 3 seconds
  • 30x30 grid: < 10 seconds

📝 Supported image formats

JPEG, PNG, BMP, GIF, TIFF, WebP

📄 License

MIT

Built with ❤️ in Rust