Skip to main content

A python wrapper for the osu! api. Includes api v2 support.

Project description

PyPI version

ossapi

ossapi is a python wrapper for the osu! api. ossapi supports both api v2 and api v1. See API v2 Usage for api v2 documentation, or API v1 Usage for api v1 documentation.

To install:

pip install ossapi

To upgrade:

pip install -U ossapi

If you need support or would like to contribute, feel free to join the circleguard discord: https://discord.gg/e84qxkQ.

API v2 Usage

Authenticating

The osu api provides two ways to authenticate, authorization code and client credentials. Authorization code grants full access to the api, but requires user interaction to authenticate the first time. Client credentials grants guest user access only, but authenticates automatically.

In either case you will need to create on oauth client on your settings page. Give it whatever name you want. However, your callback url *must* be a port on localhost. So http://localhost:3914/, http://localhost:727/, etc are all acceptable values. Make sure you're not taking a commonly used port.

Authorization Code

With the authorization code flow we will use the oauth client's id, secret, and redirect_uri (aka callback url) to authenticate. Copy these values from the oauth application you just created.

from ossapi import *
api = OssapiV2(client_id, client_secret, redirect_uri)

The first time you run this code, a page will open in your browser asking you to authenticate with osu! for your client. Once you do so, we cache the response, so you can use ossapi in the future without needing to re-authenticate.

As stated above, this flow grants full access to the api through your user.

Client Credentials

With the client credentials flow we will use the oauth client's id and secret to authenticate. Copy these values from the oauth application you just created.

from ossapi import *
api = OssapiV2(client_id, client_secret)

Unlike the authorization code flow, this authentication happens automatically and silently, and does not require user intervention. This is ideal for scripts which need to run without user interaction. As stated above however, this flow grants only guest user access to the api. This means you will not be able to use certain endpoints, like downloading replays.

Supported Endpoints

Here is a complete list of endpoints we currently have implemented. You can track our progress towards implementing all documented api v2 endpoints here: https://github.com/circleguard/ossapi/issues/14.

print(api.beatmapset_discussion_posts().discussions[0].message_type)
print(api.user_recent_activity(12092800)[0].created_at)
print(api.spotlights()[0].name)
print(api.user_beatmaps(user_id=12092800, type_="most_played")[0].count)
print(api.user_kudosu(user_id=3178418)[0].action)
print(api.beatmap_scores(beatmap_id=1981090).scores[0].id)
print(api.beatmap(beatmap_id=1981090).max_combo)
print(api.ranking("osu", RankingType.PERFORMANCE, country="US").ranking[0].user.username)
print(api.user_scores(12092800, "best")[0].accuracy)
print(api.beatmap(beatmap_id=221777).last_updated)
print(api.beatmap_user_score(beatmap_id=221777, user_id=2757689).score.mods)
print(api.beatmap_user_scores(beatmap_id=221777, user_id=2757689)[0].mods)
print(api.search(query="peppy").users.data[0].profile_colour)
print(api.comment(comment_id=1).comments[0].message)
print(api.download_score(mode="osu", score_id=2797309065))
print(api.search_beatmapsets(query="the big black").beatmapsets[0].title)
print(api.beatmapset_events(types=[BeatmapsetEventType.ISSUE_REOPEN]).events[0].type)
print(api.user(12092800).playstyle)
print(api.wiki_page("en", "Welcome").available_locales)
print(api.changelog_build("stable40", "20210520.2").users)
print(api.changelog_listing().builds[0].display_version)
print(api.changelog_lookup("lazer").changelog_entries[0].github_pull_request_id)
print(api.forum_topic(141240).posts[0].forum_id)
print(api.beatmapset_discussion_votes().votes[0].score)
print(api.score(mode="osu", score_id=2797309065).accuracy)
print(api.news_listing(year=2021).news_posts[0].author)
print(api.news_post(1025, key="id"))
print(api.friends()[0].username)
print(api.seasonal_backgrounds().backgrounds[0].url)

Note that although this code just prints a single attribute for each endpoint, you can obviously do more complicated things like iterate over arrays:

response = api.ranking("osu", RankingType.PERFORMANCE, country="US")
for ranking in response.ranking:
    print(f"global #{ranking.global_rank}: {ranking.user.username}")

Pagination

Some endpoints are paginated, and so you may need a way to access the 3rd, 5th, or 25th page of the results. The way to do this is with the Cursor class.

For example, the /rankings/ endpoint is paginated. If we wanted to get the top 1-50 players, we don't need a cursor at all, since paginated endpoints return the first page by default:

r = api.ranking("osu", RankingType.PERFORMANCE)
print(r.ranking[-1].global_rank) # 50

Accessing the subsequent page of results immediately afterwards is such a common use case that all paginated endpoints return a cursor attribute which is already pre-prepared to retrive the next page. Just pass it to a new api call:

r = api.ranking("osu", RankingType.PERFORMANCE)
cursor = r.cursor
print(r.ranking[-1].global_rank) # 50

r = api.ranking("osu", RankingType.PERFORMANCE, cursor=cursor)
print(r.ranking[-1].global_rank) # 100

However, this doesn't work so well if you want to skip a bunch of pages and go straight to eg the 20th page. To do so, construct your own Cursor object and use that:

cursor = Cursor(page=20)
r = api.ranking("osu", RankingType.PERFORMANCE, cursor=cursor)
print(r.ranking[-1].global_rank) # 1000

If there are no more pages, the cursor object of the response will be None:

cursor = Cursor(page=199)
r = api.ranking("osu", RankingType.PERFORMANCE, cursor=cursor)
print(r.cursor) # Cursor(page=200)

cursor = Cursor(page=200) # there are only 200 rankings pages
r = api.ranking("osu", RankingType.PERFORMANCE, cursor=cursor)
print(r.cursor) # None

Advanced Usage

User and Beatmap as parameters

Some functions, like api.beatmap_scores, take a beatmap_id (or user_id). We also allow passing a Beatmap / BeatmapCompact (or User / UserCompact) in place of the id as a convenience:

beatmap = api.beatmap(221777)
assert api.beatmap_scores(beatmap) == api.beatmap_scores(beatmap.id)

Internally, we simply take beatmap.id (or user.id) and supply that to the function.

Expandable Models

UserCompact and BeatmapCompact classes are "expandable" into User and Beatmap respectively. Some endpoints only return eg a UserCompact, but you may want attributes that are present on User. To expand such a class, call #expand:

compact_user = api.search(query="tybug").users.data[0]
# `statistics` is only available on `User` not `UserCompact`,
# so expansion is necessary
print(compact_user.expand().statistics.ranked_score)
# this is equivalent to
print(api.user(compact_user).statistics.ranked_score)

Similarly, beatmap = compact_beatmap.expand() is equivalent to beatmap = api.beatmap(compact_beatmap).

(Note that beatmapsets will also be expandable in the future; I am waiting for the beatmapset lookup endpoint to become documented before implementing this.)

Following Foreign Keys

Many models contain an id which references another model - for instance, Beatmap.beatmapset_id. These are called "foreign keys" in the database world. Just the id isn't very useful to us, though. We would like to be able to ask what the title, author, genre, nominations etc of the beatmapset is. What we need is the ability to "follow" this foreign key to retrieve a full Beatmapset instance. Ossapi provides a convenience method to do just that:

beatmap = api.beatmap(221777)
bmset = beatmap.beatmapset()

You can do the same for user() and beatmap(), in applicable models:

disc = api.beatmapset_discussion_posts(2641058).posts[0]
user = disc.user()

bm_playcount = api.user_beatmaps(user_id=12092800, type_="most_played")[0]
beatmap = bm_playcount.beatmap()

Note that the id field and corresponding method isn't always called beatmap, beatmapset, or user. For instance, a beatmapset discussion post has two foreign key fields last_editor_id and deleted_by_id. So we expose two methods last_editor and deleted_by to follow these foreign keys:

disc = api.beatmapset_discussion_posts(2641058).posts[0]
last_editor = disc.last_editor()
deleted_by = disc.deleted_by()
print(last_editor.username, deleted_by)

The attentive reader may notice that the api docs say that a Beatmapset attribute is optionally returned inside a Beatmap object. Why do we need a method to access this then, instead of accessing directly? Well, the key factor here is that this beatmapset could be null (None). If it's not, and you call beatmapset(), we actually just return the beatmapset given by the api. However, if it is null, we make a single api request to api.beatmapset to retrieve the beatmapset from the beatmapset id, and return that.

In short, calling such a foreign key method will cost an api call if the backing model attribute is not present, and is free otherwise. You can check for the presence of the backing model by accessing the attribute with the same name as the method, but prefixed with a underscore:

beatmap = api.beatmap(2217777)
print(beatmap._beatmapset)
# not None, so beatmap.beatmapset() would just return this attribute

Note that in some cases, the backing model is a *Compact version instead of the full version of the model. In this case, if you require the full version, I recommend calling .expand() after the foreign key method:

beatmap = api.beatmap(2217777)
# guaranteed to be a Beatmapset, not a BeatmapsetCompact
bmset = beatmap.beatmapset().expand()

Serializing Models

If you need to access the original json returned by the api, you can serialize the models back into a json string with serialize_model:

from ossapi import serialize_model
print(serialize_model(api.user("tybug2")))

Note that this is not guaranteed to be identical to the json returned by the api. For instance, there may be additional attributes in the serialized json which are optional in the api spec, not returned by the api, and set to null. But it should be essentially the same.

There are various reasons why this approach was chosen over storing the raw json returned by the api, or some other solution. Please open an issue if this approach is not sufficient for your use case.

API v1 Usage

You can get your api v1 key at https://osu.ppy.sh/p/api/. Note that due to a redirection bug, you may need to log in and wait 30 seconds before being able to access the api page through the above link.

Basic usage:

from ossapi import Ossapi

api = Ossapi("key")
print(api.get_beatmaps(user=53378)[0].submit_date)
print(api.get_match(69063884).games[0].game_id)
print(api.get_scores(221777)[0].username)
print(len(api.get_replay(beatmap_id=221777, user=6974470)))
print(api.get_user(12092800).playcount)
print(api.get_user_best(12092800)[0].pp)
print(api.get_user_recent(12092800)[0].beatmap_id)

For convenience when working with mods, we provide a Mod class, which is used wherever the api returns a mod value. An overview of its methods, in example format:

from ossapi import Mod, Ossapi

api = Ossapi("key")

mods = api.get_scores(221777)[0].mods
# Mod's __str__ uses short_name()
print(mods)
print(mods.short_name())

# to break down a mod into its component mods (eg if you want ["HD", "DT"] from "HDDT")
print(mods.decompose())

# to get the long form name (HD -> Hidden)
print(mods.long_name())

# to access the underlying value
print(mods.value)

# to add or remove a mod from the mod combination, use + and -
print(mods + Mod.FL)
print(mods - Mod.HD)
# you can also add or remove multiple mods at a time
print(mods - Mod.HDHR)

# common mod combinations are stored as static variables under `Mod` for convenience
print(Mod.HDDT, Mod.HDHR, Mod.HDDTHR)
# otherwise, the preferred way to build up mods is by adding them together
print(Mod.HD + Mod.FL + Mod.EZ)
# alternatively, you can instantiate with the raw value
print(Mod(1034))
assert Mod.HD + Mod.FL + Mod.EZ == Mod(1034)

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

ossapi-2.4.0.tar.gz (61.4 kB view details)

Uploaded Source

Built Distribution

ossapi-2.4.0-py3-none-any.whl (61.9 kB view details)

Uploaded Python 3

File details

Details for the file ossapi-2.4.0.tar.gz.

File metadata

  • Download URL: ossapi-2.4.0.tar.gz
  • Upload date:
  • Size: 61.4 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.1 CPython/3.10.5

File hashes

Hashes for ossapi-2.4.0.tar.gz
Algorithm Hash digest
SHA256 543d7932597d720883908deaec7ac2e230718c74edc17450f96ceed4d28c6a4a
MD5 baf3f42b418048b1d3a478946f82c0ce
BLAKE2b-256 f66c7dfa300c5bada9a12e0ee08ce0c94acac056cd50887970130307e4c70988

See more details on using hashes here.

File details

Details for the file ossapi-2.4.0-py3-none-any.whl.

File metadata

  • Download URL: ossapi-2.4.0-py3-none-any.whl
  • Upload date:
  • Size: 61.9 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.1 CPython/3.10.5

File hashes

Hashes for ossapi-2.4.0-py3-none-any.whl
Algorithm Hash digest
SHA256 67c15392d5f8739ec28fff7a77c03279dae9c32ae4bbac732922d5f5344f2c23
MD5 13190cf24579433e8b09dc6c8f0925d1
BLAKE2b-256 f3133ca1b41c1a09f4400435d6bd923eb372fad5a66d13117af9231836023bdf

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page