NTE Auto-Fishing

English | 简体中文 | 日本語

A visual, configurable auto-fishing assistant built for responsive game control.

Built with Python, OpenCV, MSS, PyDirectInput, and DearPyGui.

---

Highlights

• Live control center: Track bot state, fish count, session time, FPS, PID output, ROI data, and vision health in one GUI.
• Safer controls: The GUI starts paused, handles stop commands urgently, and releases held keys during pause, stop, recalibration, and shutdown.
• Runtime tuning: PID, HSV thresholds, timing, input keys, hotkeys, calibration, and debug options can be adjusted from the Settings tab. (Settings Guide)
• Resolution adaptation: Ratio-based ROI calibration with resolution-proportional fallback supports 1080p, 2K, 4K, and custom display sizes automatically.
• Fast capture and input: mss captures screen regions efficiently, while PyDirectInput sends game-friendly input events.
• Portable builds: GitHub Actions produces a single-file Windows EXE for GUI and a lightweight source package for CLI.

Project Structure

Path	Description
start_gui.py	Recommended GUI entry point.
main.py	Headless entry point and core bot loop.
config.py	Runtime configuration for PID, HSV, keys, timing, and calibration.
gui/	DearPyGui control center, panels, and thread-safe bridge.
modules/	Capture, input, vision, and fishing logic modules.
templates/	Ratio data for automatic ROI calibration.
tools/ratio_annotator.py	Utility for creating ratio-based ROI JSON from screenshots.

Getting Started

Option 1: Prebuilt Executable (GUI)

1. Download the latest NTE-Auto-Fish-v*.*.*-Windows-x64.exe from Releases.
2. Run it as Administrator so simulated input can reach the game.

Option 2: CLI Package

1. Download NTE-Auto-Fish-v*.*.*-CLI.zip from Releases and extract it.
2. Install dependencies once with python -m pip install -r requirements.txt.
3. For more options, run python main.py --help in a terminal. Available commands: start, calibrate, config show/set, reset.

Option 3: Run From Source

git clone https://github.com/Chizukuo/NTE-auto-fish.git
cd NTE-auto-fish
pip install -r requirements.txt

Launch the GUI:

python start_gui.py

Launch headless mode:

python main.py

Notes

• Run from an elevated terminal on Windows.
• Borderless window or windowed fullscreen usually gives the most reliable capture behavior.
• Hotkeys are configurable in the GUI and re-register after editing.
• Debug logging writes extra tracking data to fishing_data.csv.

Known Issues

• Dawn/dusk lighting: During in-game sunrise and sunset, the warm yellow ambient lighting significantly interferes with HSV-based cursor detection, causing tracking failures and drastically reduced fishing success rates. This is an inherent limitation of the current color-based detection approach. If you encounter this, consider adjusting the cursor HSV thresholds in Settings to compensate for the shifted lighting conditions.

---

Built for practical, low-friction automation.