Merge pull request #824 from a-luna/patch-release/v0.5.4

v0.5.4

package.json

@@ -1,6 +1,6 @@
 {
   "name": "vig-node-deps",
-  "version": "0.5.3",
+  "version": "0.5.4",
   "description": "This file is used for heroku deployment, the postbuild script below is called after each push to install the dependencies listed in the REAL package.json in the nightmarejs folder. Yes, I know that it is lame and hacky to have this root package.json file that exists solely to tap on the shoulder of another package.json and ask sheepishly to run freaking 'npm install' simply because heroku couldn't possibly manage the herculean feat of locating a package.json file anywhere but the root folder, but that's where we are and that's why this file exists.",
   "scripts": {
     "heroku-postbuild": "cd src/vigorish/nightmarejs && npm install"

requirements-dev.txt

@@ -10,14 +10,14 @@ backports.entry-points-selectable==1.1.0
 black==21.5b1
 bleach==3.3.1
 blessed==1.18.1
-boto3==1.18.10
-botocore==1.21.10
+boto3==1.18.16
+botocore==1.21.16
 bullet==2.2.0
 certifi==2021.5.30
 cffi==1.14.6
 chardet==4.0.0
-charset-normalizer==2.0.3
-cheap-repr==0.4.5
+charset-normalizer==2.0.4
+cheap-repr==0.4.6
 click==7.1.2
 cognitive-complexity==1.2.0
 cohesion==1.0.0
@@ -49,9 +49,9 @@ flake8-use-fstring==1.1
 flake8-variables-names==0.0.4
 getversion==1.0.2
 halo==0.0.31
-identify==2.2.11
+identify==2.2.13
 idna==2.10
-importlib-metadata==4.6.1
+importlib-metadata==4.6.3
 iniconfig==1.1.1
 ipdb==0.13.9
 ipykernel==6.0.3
@@ -80,7 +80,7 @@ pathspec==0.9.0
 pexpect==4.8.0
 pickleshare==0.7.5
 pip-upgrader==1.4.15
-pipdeptree==2.0.0
+pipdeptree==2.1.0
 pkginfo==1.7.1
 platformdirs==2.2.0
 pluggy==0.13.1
@@ -110,15 +110,15 @@ python-dateutil==2.8.2
 python-dotenv==0.19.0
 pytz==2021.1
 PyYAML==5.4.1
-pyzmq==22.1.0
+pyzmq==22.2.1
 rapidfuzz==1.4.1
 readme-renderer==29.0
-regex==2021.7.6
+regex==2021.8.3
 requests==2.26.0
 requests-toolbelt==0.9.1
 rfc3986==1.5.0
 s3transfer==0.5.0
-scipy==1.7.0
+scipy==1.7.1
 SecretStorage==3.3.1
 setuptools-scm==6.0.1
 six==1.16.0
@@ -134,15 +134,15 @@ termcolor==1.1.0
 terminaltables==3.1.0
 toml==0.10.2
 tornado==6.1
-tox==3.24.0
-tqdm==4.61.2
+tox==3.24.1
+tqdm==4.62.0
 traitlets==5.0.5
 twine==3.4.2
 typed-ast==1.4.3
 typepy==1.2.0
 typing-extensions==3.10.0.0
 urllib3==1.26.6
-virtualenv==20.6.0
+virtualenv==20.7.0
 w3lib==1.22.0
 wcwidth==0.2.5
 webencodings==0.5.1

requirements.txt

@@ -1,7 +1,7 @@
 aenum==3.1.0
 blessed==1.18.1
-boto3==1.18.10
-botocore==1.21.10
+boto3==1.18.16
+botocore==1.21.16
 bullet==2.2.0
 certifi==2021.5.30
 chardet==4.0.0
@@ -25,14 +25,14 @@ PyYAML==5.4.1
 rapidfuzz==1.4.1
 requests==2.26.0
 s3transfer==0.5.0
-scipy==1.7.0
+scipy==1.7.1
 six==1.16.0
 spinners==0.0.24
 SQLAlchemy>=1.3.23,<1.4.0
 SQLAlchemy-Utils==0.37.8
 tabulate==0.8.9
 termcolor==1.1.0
-tqdm==4.61.2
+tqdm==4.62.0
 urllib3==1.26.6
 w3lib==1.22.0
 wcwidth==0.2.5

src/vigorish/data/metrics/pfx_metrics.py

@@ -36,14 +36,6 @@ def __repr__(self):
     def pitch_name(self) -> str:
         return self.pitch_type.print_name
 
-    @property
-    def triple_slash(self) -> str:
-        return (
-            f"{format_decimal_bat_stat(self.avg)}/"
-            f"{format_decimal_bat_stat(self.obp)}/"
-            f"{format_decimal_bat_stat(self.slg)} ({format_decimal_bat_stat(self.ops)})"
-        )
-
     @cached_property
     def pitch_type_int(self):
         return sum({p.pitch_type_int for p in self.pfx})
@@ -56,6 +48,14 @@ def pitch_type(self) -> PitchType:
     def pitch_type_abbrevs(self) -> List[str]:
         return [str(pt) for pt in self.pitch_type]
 
+    @property
+    def triple_slash(self) -> str:
+        return (
+            f"{format_decimal_bat_stat(self.avg)}/"
+            f"{format_decimal_bat_stat(self.obp)}/"
+            f"{format_decimal_bat_stat(self.slg)} ({format_decimal_bat_stat(self.ops)})"
+        )
+
     @cached_property
     def total_pitches(self) -> int:
         return len(self.pfx)
@@ -443,50 +443,3 @@ def get_batted_ball_stats(self, include_bip_count: bool = False) -> Dict[str, st
     def get_usage_stats(self, include_pitch_count: bool = False) -> str:
         pitch_count = f" ({self.total_pitches})" if include_pitch_count else ""
         return f"{self.percent:.0%}{pitch_count} {self.avg_speed:.1f}mph"
-
-
-# from vigorish.app import Vigorish
-# from vigorish.data.metrics.pfx_metrics import PitchFxMetrics
-# mlb_id = 621111
-# game_id = "LAN202106130"
-# app = Vigorish()
-# pfx_for_game = PitchFxMetrics.for_pitcher_game(app.db_session, mlb_id, game_id)
-
-# from vigorish.app import Vigorish
-# from vigorish.data.metrics.pfx_metrics import PitchFxMetrics
-# from vigorish.data.player_data import PlayerData
-# mlb_id = 621111
-# game_id = "LAN202106130"
-# app = Vigorish()
-# pd = PlayerData(app, mlb_id)
-# pfx = pd.get_pfx_bat_metrics_for_game(game_id)
-
-# import json
-# from pathlib import Path
-# from vigorish.app import Vigorish
-# from vigorish.data.metrics.pfx_metrics import PitchFxMetrics
-# from vigorish.util.datetime_util import format_timedelta_str
-# mlb_id = 425794
-# app = Vigorish()
-# pfx_by_year = PitchFxMetrics.for_pitcher_by_year(app.db_session, mlb_id)
-# Path("pfx_metrics_example.json").write_text(json.dumps(pfx_by_year, indent=2, sort_keys=False))
-
-# import json
-# from pathlib import Path
-# from vigorish.app import Vigorish
-# from vigorish.data.player_data import PlayerData
-# app = Vigorish()
-# pd = PlayerData(app, 621111)
-# career_pfx=pd.get_all_pfx_career_data()
-# career_pfx["all"]["metrics"] = career_pfx["all"]["metrics"].as_dict()
-# career_pfx["rhb"]["metrics"] = career_pfx["rhb"]["metrics"].as_dict()
-# career_pfx["lhb"]["metrics"] = career_pfx["lhb"]["metrics"].as_dict()
-# yearly_pfx=pd.get_all_pfx_yearly_data()
-# for year, pfx_stats_for_year in yearly_pfx["all"]["metrics"].items():
-#     yearly_pfx["all"]["metrics"][year] = pfx_stats_for_year.as_dict()
-# for year, pfx_stats_for_year in yearly_pfx["rhb"]["metrics"].items():
-#     yearly_pfx["rhb"]["metrics"][year] = pfx_stats_for_year.as_dict()
-# for year, pfx_stats_for_year in yearly_pfx["lhb"]["metrics"].items():
-#     yearly_pfx["lhb"]["metrics"][year] = pfx_stats_for_year.as_dict()
-# Path("career_pfx.json").write_text(json.dumps(career_pfx, indent=2, sort_keys=False))
-# Path("yearly_pfx.json").write_text(json.dumps(yearly_pfx, indent=2, sort_keys=False))

src/vigorish/data/scraped_data.py

@@ -13,6 +13,7 @@
 )
 from vigorish.data.name_search import PlayerNameSearch
 from vigorish.enums import DataSet, DefensePosition, PitchType, VigFile
+from vigorish.util.dt_format_strings import DATE_ONLY_TABLE_ID
 from vigorish.util.list_helpers import flatten_list2d, group_and_sort_list
 from vigorish.util.regex import URL_ID_CONVERT_REGEX, URL_ID_REGEX
 from vigorish.util.result import Result
@@ -260,14 +261,34 @@ def get_all_seasons_with_data_for_player(self, player_mlb_id: int) -> List[db.Se
     def player_name_search(self, query):
         return self.name_search.fuzzy_match(query)
 
-    def get_season_standings(self, year=None):
-        if not year:
-            most_recent_season = max(s.year for s in db.Season.get_all_regular_seasons(self.db_session))
-            year = most_recent_season
+    def get_pitch_apps_for_player_up_to_date(self, mlb_id, game_date):
+        player_id = db.PlayerId.find_by_mlb_id(self.db_session, mlb_id)
+        if not player_id:
+            return []
+        date_id = game_date.strftime(DATE_ONLY_TABLE_ID)
+        date_status = self.db_session.query(db.DateScrapeStatus).get(int(date_id))
+        return (
+            self.db_session.query(db.PitchStats)
+            .filter(db.PitchStats.player_id == player_id.db_player_id)
+            .filter(db.PitchStats.date_id <= date_id)
+            .filter(db.PitchStats.season_id == date_status.season_id)
+        )
+
+    def get_pitcher_record_on_date(self, mlb_id, game_date):
+        pitch_apps = self.get_pitch_apps_for_player_up_to_date(mlb_id, game_date)
+        wins = sum(p.is_wp for p in pitch_apps)
+        losses = sum(p.is_lp for p in pitch_apps)
+        return (wins, losses)
+
+    def get_pitcher_total_saves_on_date(self, mlb_id, game_date):
+        pitch_apps = self.get_pitch_apps_for_player_up_to_date(mlb_id, game_date)
+        return sum(p.is_sv for p in pitch_apps)
+
+    def get_season_standings(self, year, game_date=None):
         standings = []
         all_teams = db.Team.get_all_teams_for_season(self.db_session, year)
         for team in all_teams:
-            all_games = db.GameScrapeStatus.get_all_games_for_team(self.db_session, team.team_id_br, year)
+            all_games = db.GameScrapeStatus.get_all_games_for_team(self.db_session, team.team_id_br, year, game_date)
             away_games = list(filter(lambda x: x.away_team_id_br == team.team_id_br, all_games))
             home_games = list(filter(lambda x: x.home_team_id_br == team.team_id_br, all_games))
             away_results = self._get_away_game_results(away_games)

src/vigorish/models/player.py

@@ -42,6 +42,10 @@ def __repr__(self):
     def as_dict(self):
         return {c.name: getattr(self, c.name) for c in self.__table__.columns}
 
+    @property
+    def name(self):
+        return f"{self.name_first} {self.name_last}"
+
     @classmethod
     def find_by_bbref_id(cls, db_session, bbref_id):
         return db_session.query(cls).filter_by(bbref_id=bbref_id).first()

src/vigorish/models/season.py

@@ -1,5 +1,8 @@
 """Db model that describes a MLB season and tracks data scraping progress."""
+from __future__ import annotations
+
 from datetime import date, datetime
+from typing import List
 
 from sqlalchemy import Column, DateTime, Integer
 from sqlalchemy.ext.hybrid import hybrid_property
@@ -522,9 +525,8 @@ def is_date_in_season(cls, db_session, check_date, season_type=SeasonType.REGULA
                 f'{season_type.replace("_", " ").title()}'
             )
             return Result.Fail(error)
-        date_str = check_date.strftime(DATE_ONLY)
         if check_date < season.start_date or check_date > season.end_date:
-            error = f"{date_str} is not within the scope of the {season.name}"
+            error = f"{check_date.strftime(DATE_ONLY)} is not within the scope of the {season.name}"
             return Result.Fail(error)
         return Result.Ok(season)
 
@@ -575,7 +577,7 @@ def validate_date_range(cls, db_session, start, end):
         return Result.Ok(season)
 
     @classmethod
-    def get_all_regular_seasons(cls, db_session):
+    def get_all_regular_seasons(cls, db_session) -> List[db.Season]:
         return db_session.query(cls).filter_by(season_type=SeasonType.REGULAR_SEASON).all()
 
     @classmethod

src/vigorish/models/status_game.py

@@ -6,9 +6,9 @@
 from sqlalchemy import Column, DateTime, ForeignKey, Integer, String
 from sqlalchemy.ext.hybrid import hybrid_property
 from sqlalchemy.orm import relationship
+from sqlalchemy.sql.elements import or_
 
 import vigorish.database as db
-from vigorish.models.season import Season
 from vigorish.util.dt_format_strings import (
     DATE_ONLY,
     DATE_ONLY_2,
@@ -328,7 +328,7 @@ def get_all_brooks_game_ids_for_date(cls, db_session, date):
 
     @classmethod
     def get_all_bbref_game_ids_for_team(cls, db_session, team_id_br, year):
-        season = Season.find_by_year(db_session, year)
+        season = db.Season.find_by_year(db_session, year)
         if not season:
             return None
         games_for_season = list(db_session.query(cls).filter_by(season_id=season.id).all())
@@ -341,19 +341,24 @@ def get_all_bbref_game_ids_for_team(cls, db_session, team_id_br, year):
         ]
 
     @classmethod
-    def get_all_games_for_team(cls, db_session, team_id_br, year):
-        season = Season.find_by_year(db_session, year)
+    def get_all_games_for_team(cls, db_session, team_id_br, year, game_date=None):
+        if not game_date:
+            game_date = db.Season.get_most_recent_scraped_date(db_session, year)
+        if game_date.year != year:
+            raise ValueError('"year" and "game_date" parameters must belong to the same season')
+        date_id = game_date.strftime(DATE_ONLY_TABLE_ID)
+        season = db.Season.find_by_year(db_session, year)
         if not season:
             return None
-        games_for_season = list(db_session.query(cls).filter_by(season_id=season.id).all())
-        games_for_season.sort(key=lambda x: x.scrape_status_date_id)
-        bb_team_id = get_brooks_team_id(team_id_br)
-        return [
-            game
-            for game in games_for_season
-            if (game.away_team_id_bb == bb_team_id or game.home_team_id_bb == bb_team_id)
-            and game.scraped_bbref_boxscore
-        ]
+        return (
+            db_session.query(cls)
+            .filter(cls.season_id == season.id)
+            .filter(cls.scrape_status_date_id <= date_id)
+            .filter(or_(cls.away_team_id_br == team_id_br, cls.home_team_id_br == team_id_br))
+            .filter(cls.scraped_bbref_boxscore == 1)
+            .order_by(cls.scrape_status_date_id)
+            .all()
+        )
 
     @classmethod
     def get_game_id_map(cls, db_session):

src/vigorish/nightmarejs/package.json

@@ -1,6 +1,6 @@
 {
   "name": "vig-nodejs",
-  "version": "0.5.3",
+  "version": "0.5.4",
   "description": "NightmareJS and other node modules used to scrape html from websites",
   "author": "Aaron Luna <[email protected]> (https://aaronluna.dev)",
   "license": "MIT",