A standalone crate to calculate star ratings and performance points for all osu! gamemodes. Peace edition
For Python: peace-performance-python
Fork of rosu-pp | MIT License
Conversions between gamemodes are generally not supported.
Async is supported through features, see below.
use std::fs::File;
use peace_performance::{Beatmap, BeatmapExt};
let file = match File::open("/path/to/file.osu") {
Ok(file) => file,
Err(why) => panic!("Could not open file: {}", why),
};
// Parse the map yourself
let map = match Beatmap::parse(file) {
Ok(map) => map,
Err(why) => panic!("Error while parsing map: {}", why),
};
// If `BeatmapExt` is included, you can make use of
// some methods on `Beatmap` to make your life simpler.
// If the mode is known, it is recommended to use the
// mode's pp calculator, e.g. `TaikoPP`, manually.
let result = map.pp()
.mods(24) // HDHR
.combo(1234)
.misses(2)
.accuracy(99.2)
.calculate();
println!("PP: {}", result.pp());
// If you intend to reuse the current map-mod combination,
// make use of the previous result!
// If attributes are given, then stars & co don't have to be recalculated.
let next_result = map.pp()
.mods(24) // HDHR
.attributes(result) // recycle
.combo(543)
.misses(5)
.n50(3)
.passed_objects(600)
.accuracy(96.5)
.calculate();
println!("Next PP: {}", next_result.pp());
let stars = map.stars(16, None).stars(); // HR
let max_pp = map.max_pp(16).pp();
println!("Stars: {} | Max PP: {}", stars, max_pp);
If either the async_tokio
or async_std
feature is enabled, beatmap parsing will be async.
use async_std::fs::File;
// use tokio::fs::File;
let file = match File::open("/path/to/file.osu").await {
Ok(file) => file,
Err(why) => panic!("Could not open file: {}", why),
};
// Parse the map asynchronously
let map = match Beatmap::parse(file).await {
Ok(map) => map,
Err(why) => panic!("Error while parsing map: {}", why),
};
// The rest stays the same
let result = map.pp()
.mods(24) // HDHR
.combo(1234)
.misses(2)
.accuracy(99.2)
.calculate();
println!("PP: {}", result.pp());
all_included
: Both stack leniency & slider paths are considered so that the difficulty and pp calculation immitates osu! as close as possible. Pro: Most precise; Con: Least performant.no_leniency
: The positional offset of notes created by stack leniency is not considered. This means the jump distance inbetween notes might be slightly off, resulting in small inaccuracies. Since calculating these offsets is relatively expensive though, this version is considerably faster thanall_included
.no_sliders_no_leniency
(i.e. oppai): In addition to not considering the positional offset caused by stack leniency, slider paths are also ignored. This means the travel distance of notes is completely omitted which may cause further inaccuracies. Since the slider paths don't have to be computed though, it is generally faster thanno_leniency
.
Note: If the fruits
feature is enabled, sliders will be parsed regardless, resulting in a reduced performance advantage of no_sliders_no_leniency
.
Flag | Description |
---|---|
default |
Enable async_tokio, all modes and choose the all_included version for osu!standard. Set default_features = false to disable. |
score_v2_buff |
Buff ScoreV2 (STD) - acc *= 1.25 |
ppysb_edition |
Special changes for RELAX and AUTOPILOT |
relax_nerf |
Nerf relax and autopilot pp. Relax: aim * 0.9, spd * 0.3, acc *0.8 ; Autopilot: aim * 0.3, spd * 0.9, acc * 0.8 |
taiko |
Enable osu!taiko. |
fruits |
Enable osu!ctb. |
mania |
Enable osu!mania. |
osu |
Enable osu!standard. Requires to also enable exactly one of the features no_leniency , no_sliders_no_leniency , or all_included . |
no_leniency |
When calculating difficulty attributes in osu!standard, ignore stack leniency but consider sliders. Solid middleground between performance and precision, hence the default version. |
no_sliders_no_leniency |
When calculating difficulty attributes in osu!standard, ignore stack leniency and sliders. Best performance but slightly less precision than no_leniency . |
all_included |
When calculating difficulty attributes in osu!standard, consider both stack leniency and sliders. Best precision but significantly worse performance than no_leniency . |
async_tokio |
Beatmap parsing will be async through tokio |
async_std |
Beatmap parsing will be async through async-std |
Comparing the PP calculation speed between osu-perf (alternative rust pp calculculation crate), an oppai-ng rust binding, and rosu-pp's no_sliders_no_leniency
:
Comparing the PP calculation speed between rosu-pp's all_included
, no_leniency
, and no_sliders_no_leniency
versions:
Comparing the PP (in)accuracy between rosu-pp's all_included
, no_leniency
, and no_sliders_no_leniency
versions:
Comparing the stars (in)accuracy between rosu-pp's all_included
, no_leniency
, and no_sliders_no_leniency
versions:
- osu sr versions
- all included
- no_leniency
- no_sliders_no_leniency
- taiko sr
- ctb sr
- mania sr
- osu pp
- taiko pp
- ctb pp
- mania pp
- refactoring
- benchmarking
- async