Skip to content

Latest commit

 

History

History
407 lines (292 loc) · 11.4 KB

README.md

File metadata and controls

407 lines (292 loc) · 11.4 KB
Raw

PyPI version nhl-api-py workflow

NHL API & NHL Edge Stats

About

NHL-api-py is a Python package that provides a simple wrapper around the NHL API, allowing you to easily access and retrieve NHL data in your Python applications.

Note: This is very early maturing, I created this to help me with some machine learning projects around the NHL and the NHL data sets. Special thanks to https://github.com/erunion/sport-api-specifications/tree/master/nhl and https://gitlab.com/dword4/nhlapi/-/blob/master/stats-api.md.

Developer Note: This is being updated with the new, also undocumented, NHL API.

As of 10/5/24 I seem to have a majority of the endpoints added from what I can tell, but every once and awhile I come across one that needs to be added/changed. These will most likely be a minor ver bump.

If you find any, open a ticket or post in the discussions tab. I would love to hear more.

Contact

Im available on Bluesky for any questions or just general chats about enhancements.

Usage

pip install nhl-api-py
from nhlpy import NHLClient

client = NHLClient()
# Fore more verbose logging
client = NHLClient(verbose=True)
# OR Other available configurations:
client = NHLClient(verbose={bool}, timeout={int}, ssl_verify={bool}, follow_redirects={bool})

Stats with QueryBuilder

The skater stats endpoint can be accessed using the new query builder. It should make creating and understanding the queries a bit easier. Filters are being added as I go, and will match up to what the NHL API will allow.

The idea is to easily, and programatically, build up more complex queries using the query filters. A quick example below:

filters = [
    GameTypeQuery(game_type="2"),
    DraftQuery(year="2020", draft_round="2"),
    SeasonQuery(season_start="20202021", season_end="20232024"),
    PositionQuery(position=PositionTypes.ALL_FORWARDS)
]

Sorting

The sorting is a list of dictionaries similar to below. You can supply your own, otherwise it will default to the default sort properties that the stat dashboard uses. All sorting defaults are found in the nhl-api-py/nhlpy/api/query/sorting/sorting_options.py file.

Default Sorting 
skater_summary_default_sorting = [
    {"property": "points", "direction": "DESC"},
    {"property": "gamesPlayed", "direction": "ASC"},
    {"property": "playerId", "direction": "ASC"},
]

Report Types

The following report types are available. These are used to build the request url. So /summary, /bios, etc.

summary
bios
faceoffpercentages
faceoffwins
goalsForAgainst
realtime
penalties
penaltykill
penaltyShots
powerplay
puckPossessions
summaryshooting
percentages
scoringRates
scoringpergame
shootout
shottype
timeonice

Available Filters

from nhlpy.api.query.filters.franchise import FranchiseQuery
from nhlpy.api.query.filters.shoot_catch import ShootCatchesQuery
from nhlpy.api.query.filters.draft import DraftQuery
from nhlpy.api.query.filters.season import SeasonQuery
from nhlpy.api.query.filters.game_type import GameTypeQuery
from nhlpy.api.query.filters.position import PositionQuery, PositionTypes
from nhlpy.api.query.filters.status import StatusQuery
from nhlpy.api.query.filters.opponent import OpponentQuery
from nhlpy.api.query.filters.home_road import HomeRoadQuery
from nhlpy.api.query.filters.experience import ExperienceQuery
from nhlpy.api.query.filters.decision import DecisionQuery

filters = [
    GameTypeQuery(game_type="2"),
    DraftQuery(year="2020", draft_round="2"),
    SeasonQuery(season_start="20202021", season_end="20232024"),
    PositionQuery(position=PositionTypes.ALL_FORWARDS),
    ShootCatchesQuery(shoot_catch="L"),
    HomeRoadQuery(home_road="H"),
    FranchiseQuery(franchise_id="1"),
    StatusQuery(is_active=True),#for active players OR for HOF players StatusQuery(is_hall_of_fame=True),
    OpponentQuery(opponent_franchise_id="2"),
    ExperienceQuery(is_rookie=True), # for rookies || ExperienceQuery(is_rookie=False) #for veteran
    DecisionQuery(decision="W") # OR DecisionQuery(decision="L") OR DecisionQuery(decision="O")
]

Example

from nhlpy.api.query.builder import QueryBuilder, QueryContext
from nhlpy.nhl_client import NHLClient
from nhlpy.api.query.filters.draft import DraftQuery
from nhlpy.api.query.filters.season import SeasonQuery
from nhlpy.api.query.filters.game_type import GameTypeQuery
from nhlpy.api.query.filters.position import PositionQuery, PositionTypes

client = NHLClient(verbose=True)

filters = [
    GameTypeQuery(game_type="2"),
    DraftQuery(year="2020", draft_round="2"),
    SeasonQuery(season_start="20202021", season_end="20232024"),
    PositionQuery(position=PositionTypes.ALL_FORWARDS)
]

query_builder = QueryBuilder()
query_context: QueryContext = query_builder.build(filters=filters)

data = client.stats.skater_stats_with_query_context(
    report_type='summary',
    query_context=query_context,
    aggregate=True
)

Granular Filtering

Each API request uses an additional query parameter called factCayenneExp. This defaults to gamesPlayed>=1 but can be overridden by setting the fact_query parameter in the QueryContextObject object. These can be combined together with and to create a more complex query. It supports >, <, >=, <=. For example: shootingPct>=0.01 and timeOnIcePerGame>=60 and faceoffWinPct>=0.01 and shots>=1

This should support the following filters:

  • gamesPlayed
  • points
  • goals
  • pointsPerGame
  • penaltyMinutes
  • plusMinus
  • ppGoals # power play goals
  • evGoals # even strength goals
  • pointsPerGame
  • penaltyMinutes
  • evPoints # even strength points
  • ppPoints # power play points
  • gameWinningGoals
  • otGoals
  • shPoints # short handed points
  • shGoals # short handed goals
  • shootingPct
  • timeOnIcePerGame
  • faceoffWinPct
  • shots
.....
query_builder = QueryBuilder()
query_context: QueryContext = query_builder.build(filters=filters)

query_context.fact_query = "gamesPlayed>=1 and goals>=10"  # defaults to gamesPlayed>=1

data = client.stats.skater_stats_with_query_context(
    report_type='summary',
    query_context=query_context,
    aggregate=True
)

Invalid Query / Errors

The QueryContext object will hold the result of the built query with the supplied queries. In the event of an invalid query (bad data, wrong option, etc), the QueryContext object will hold all the errors that were encountered during the build process. This should help in debugging.

You can quickly check the QueryContext object for errors by calling query_context.is_valid(). Any "invalid" filters will be removed from the output query, but anything that is still valid will be included.

...
query_context: QueryContext = query_builder.build(filters=filters)
query_context.is_valid() # False if any of the filters fails its validation check
query_context.errors

Additional Stats Endpoints (In development)

client.stats.club_stats_season(team_abbr="BUF") # kinda weird endpoint.

client.stats.player_career_stats(player_id="8478402")

client.stats.player_game_log(player_id="8478402", season_id="20242025", game_type="2")

# Team Summary Stats.
#   These have lots of available parameters.  You can also tap into the apache cayenne expressions to build custom
#   queries, if you have that knowledge.
client.stats.team_summary(start_season="20202021", end_season="20212022", game_type_id=2)
client.stats.team_summary(start_season="20202021", end_season="20212022")


# Skater Summary Stats.
#   Queries for skaters for year ranges, filterable down by franchise.
client.stats.skater_stats_summary_simple(start_season="20232024", end_season="20232024")
client.stats.skater_stats_summary_simple(franchise_id=10, start_season="20232024", end_season="20232024")

# For the following query context endpoints, see the above section
client.stats.skater_stats_with_query_context(...)

# Goalies
client.stats.goalie_stats_summary_simple(start_season="20242025", stats_type="summary")

Schedule Endpoints

# Returns the games for the given date.
client.schedule.get_schedule(date="2021-01-13")

# Return games for the week of (date)
client.schedule.get_weekly_schedule(date="2021-01-13")

client.schedule.get_schedule_by_team_by_month(team_abbr="BUF")
client.schedule.get_schedule_by_team_by_month(team_abbr="BUF", month="2021-01")

client.schedule.get_schedule_by_team_by_week(team_abbr="BUF")
client.schedule.get_schedule_by_team_by_week(team_abbr="BUF", date="2024-01-01")

client.schedule.get_season_schedule(team_abbr="BUF", season="20212022")

client.schedule.schedule_calendar(date="2023-11-23")

Standings Endpoints

client.standings.get_standings()
client.standings.get_standings(date="2021-01-13")
client.standings.get_standings(season="20222023")

# standings manifest.  This returns a ton of information for every season ever it seems like
# This calls the API for this info, I also cache this in /data/seasonal_information_manifest.json
# for less API calls since this only changes yearly.
client.standings.season_standing_manifest()

Teams Endpoints

client.teams.teams_info() # returns id + abbrevation + name of all teams

Game Center

client.game_center.boxscore(game_id="2023020280")

client.game_center.play_by_play(game_id="2023020280")

client.game_center.landing(game_id="2023020280")

client.game_center.score_now()

# this is used via the website to provide additional related game information
client.game_center.right_rail(game_id="2023020280")

Misc Endpoints

client.misc.glossary()

client.misc.config()

client.misc.countries()

client.misc.season_specific_rules_and_info()

client.misc.draft_year_and_rounds()

Insomnia Rest Client Export

Insomnia Rest Client is a great tool for testing

nhl_api-{ver}.json in the root folder is an export of the endpoints I have been working through using the Insomnia Rest Client. You can import this directly into the client and use it to test the endpoints. I will be updating this as I go

Developers

  1. Install Poetry

curl -sSL https://install.python-poetry.org | python3 -

or using pipx

pipx install poetry

  1. poetry install --with dev

  2. poetry shell

Build Pipeline

The build pipeline will run black, ruff, and pytest. Please make sure these are passing before submitting a PR.

$ poetry shell

# You can then run the following
$ pytest
$ ruff .
$ black .

pypi test net

poetry build
poetry publish -r test-pypi

Poetry version management

# View current version
poetry version

# Bump version
poetry version patch  # 0.1.0 -> 0.1.1
poetry version minor  # 0.1.0 -> 0.2.0
poetry version major  # 0.1.0 -> 1.0.0

# Set specific version
poetry version 2.0.0

# Set pre-release versions
poetry version prepatch  # 0.1.0 -> 0.1.1-alpha.0
poetry version preminor  # 0.1.0 -> 0.2.0-alpha.0
poetry version premajor  # 0.1.0 -> 1.0.0-alpha.0

# Specify pre-release identifier
poetry version prerelease  # 0.1.0 -> 0.1.0-alpha.0
poetry version prerelease beta  # 0.1.0-alpha.0 -> 0.1.0-beta.0