Skip to content

Hinamizawa-Forks/taiko-replay-analyzer

 
 

Repository files navigation

AI slop. polluting the environment so you can see fullalt %

osu!Taiko Replay Analyzer

Visualize and analyze osu!Taiko replays. Detects playstyle (full-alt / singletap / hybrid), tracks BPM comfort zones, shows pattern weak spots, and builds per-player profiles across many replays.


Windows

  1. Install Python 3.11+ from https://www.python.org/downloads/ — tick "Add Python to PATH" during setup.

  2. Double-click run.bat — it sets everything up on first launch.

That's it. Every launch after that goes straight to the app.


Linux

You need: Python 3.11+ and pygame 2.x

Install pygame via your package manager (python3-pygame) or pip (pip3 install pygame), then run:

python3 main.py

Setup

Songs folder

The app needs your osu! Songs folder (or osu! lazer files/ directory) to load beatmaps.

It tries to detect the location automatically. If that fails, it will prompt you once — enter the path and it saves to config.txt.

You can also set it manually by opening config.txt and editing the line:

songs_folder = C:\Users\you\AppData\Local\osu!\Songs

Common locations:

Stable (Songs folder)

  • Windows: %LOCALAPPDATA%\osu!\Songs
  • Linux: ~/.local/share/osu-wine/osu!/Songs

Lazer (files directory)

  • Windows: %APPDATA%\osu\files
  • Linux: ~/.local/share/osu/files

Replays

To open a single replay, put the .osr file in the same folder as the app and run. If there is only one .osr file it opens automatically — if there are multiple you will be asked to pick one.

To import many replays at once into profiles, use --mass-add (see below).


Usage

python3 main.py [replay.osr] [options]

Windows: replace python3 with python, or just use run.bat.

Options

  • (no arguments) — auto-detects a single .osr in the current folder
  • replay.osr — open a specific replay file
  • --songs <path> — override the Songs / lazer files folder for this run
  • --portable — run without a Songs folder (playstyle analysis only, no beatmap data)
  • --profile — open the profile viewer
  • --profile "Name" — open a specific player's profile directly
  • --create-profile "Name" — create a new player profile
  • --create-profile "Name" --layout DDKK — create profile with a specific key layout
  • --create-profile "Name" --aliases "Alt,Old" — create profile with username aliases
  • --mass-add — scan a folder of replays and import them into matching profiles
  • --delete-profile "Name" — delete a player profile

Profiles

Profiles store replay history and aggregate stats across sessions.

# Create
python3 main.py --create-profile "PlayerName"
python3 main.py --create-profile "PlayerName" --layout DDKK

# Import many replays at once
python3 main.py --mass-add

# Delete
python3 main.py --delete-profile "PlayerName"

Saved profiles are listed in config.txt. Profile data is stored at:

  • Windows: %APPDATA%\taiko-replay-analyzer\profiles\
  • Linux: ~/.local/share/taiko-replay-analyzer/profiles\

Key Layouts

  • KDDK — Kat · Don · Don · Kat (default)
  • DDKK — Don · Don · Kat · Kat
  • KKDD — Kat · Kat · Don · Don

Layout is set when creating a profile. Press [L] in the Profile Viewer → Playstyle tab to change it at any time.


Controls

Replay Viewer

  • Space — play / pause
  • ← → — seek ±5 seconds
  • Shift + ← → — seek ±30 seconds
  • D — toggle data panel
  • P — open player profile
  • Esc — quit

Profile Viewer

  • Tab — cycle tabs (Overview / Patterns / Playstyle)
  • ↑ ↓ — scroll
  • L — cycle key layout
  • Enter — open selected replay
  • Esc — back / quit

Skin

The skin/ folder holds the visual assets (notes, drum, hit results, lane background). These are standard osu! skin files — replace them with any skin you like. The app falls back to plain shapes if an asset is missing.

About

use for rust mirror service & future frontend staging osu headless client replay output to give taiko replay analyzer, for none Josu replays (none std) while osu headless linux client causes flashbang effects

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages

  • Python 99.4%
  • Batchfile 0.6%