wsba-hockey 1.2.3__tar.gz → 1.2.5__tar.gz

{wsba_hockey-1.2.3/src/wsba_hockey.egg-info → wsba_hockey-1.2.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: wsba_hockey
-Version: 1.2.3
+Version: 1.2.5
 Summary: WeakSide Breakout's complete Python package of access to hockey data, primairly including the scraping of National Hockey League schedule, play-by-play, and shifts information.
 Author-email: Owen Singh <owenbksingh@gmail.com>
 Project-URL: Homepage, https://github.com/owensingh38/wsba_hockey/

{wsba_hockey-1.2.3 → wsba_hockey-1.2.5}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "wsba_hockey"
-version = "1.2.3"
+version = "1.2.5"
 authors = [
   { name="Owen Singh", email="owenbksingh@gmail.com" },
 ]

{wsba_hockey-1.2.3 → wsba_hockey-1.2.5}/src/wsba_hockey/tools/columns.py

@@ -4,7 +4,9 @@
 def col_map():
     return {
         'schedule':{
+            'season':'season',
             'id':'game_id',
+            'game_title':'game_title',
             'game_date':'game_date',
             'gameType':'season_type',
             'neutralSite':'neutral_site',
@@ -36,6 +38,7 @@ def col_map():
             'periodDescriptor.maxRegulationPeriods':'period_max_regulation',
             'periodDescriptor.periodType':'period_type',
             'gameOutcome.lastPeriodType':'period_type_last',
+            'gameOutcome.otPeriods':'ot_periods',
             'seriesUrl':'series_url',
             'seriesStatus.round':'series_round',
             'seriesStatus.seriesAbbrev':'series_abbr',
@@ -46,7 +49,11 @@ def col_map():
             'seriesStatus.topSeedTeamAbbrev':'top_seed_team_abbr',
             'seriesStatus.bottomSeedTeamAbbrev':'bottom_seed_team_abbr',
             'seriesStatus.topSeedWins':'top_seed_wins',
-            'seriesStatus.bottomSeedWins':'bottom_seed_wins'
+            'seriesStatus.bottomSeedWins':'bottom_seed_wins',
+            'clock.timeRemaining':'period_time_remaining',
+            'clock.secondsRemaining':'period_seconds_remaining',
+            'clock.running':'period_clock_running',
+            'clock.inIntermission':'game_in_intermission'
         },
         'season_info':{
             'id':'season',
@@ -155,7 +162,35 @@ def col_map():
             "placeName.default": "place_name",
             "teamName.default": "team_name",
             "teamCommonName.default": "team_common_name",
-            "teamAbbrev.default": "team_abbr"
+            "teamAbbrev.default": "team_abbr",
+            "seriesUrl": "series_url",
+            "seriesTitle": "series_title",
+            "seriesAbbrev": "series_abbrev",
+            "seriesLetter": "series_letter",
+            "playoffRound": "playoff_round",
+            "topSeedRank": "top_seed_rank",
+            "topSeedRankAbbrev": "top_seed_rank_abbr",
+            "topSeedWins": "top_seed_wins",
+            "bottomSeedRank": "bottom_seed_rank",
+            "bottomSeedRankAbbrev": "bottom_seed_rank_abbr",
+            "bottomSeedWins": "bottom_seed_wins",
+            "winningTeamId": "winning_team_id",
+            "losingTeamId": "losing_team_id",
+            "topSeedTeam.id": "top_seed_team_id",
+            "topSeedTeam.abbrev": "top_seed_team_abbr",
+            "topSeedTeam.name.default": "top_seed_team_name",
+            "topSeedTeam.commonName.default": "top_seed_team_common_name",
+            "topSeedTeam.placeNameWithPreposition.default": "top_seed_team_place_name",
+            "topSeedTeam.logo": "top_seed_team_logo",
+            "topSeedTeam.darkLogo": "top_seed_team_dark_logo",
+            "bottomSeedTeam.id": "bottom_seed_team_id",
+            "bottomSeedTeam.abbrev": "bottom_seed_team_abbr",
+            "bottomSeedTeam.name.default": "bottom_seed_team_name",
+            "bottomSeedTeam.commonName.default": "bottom_seed_team_common_name",
+            "bottomSeedTeam.placeNameWithPreposition.default": "bottom_seed_team_place_name",
+            "bottomSeedTeam.logo": "bottom_seed_team_logo",
+            "bottomSeedTeam.darkLogo": "bottom_seed_team_dark_logo",
+            "seriesLogo": "series_logo",
         },
         'roster':{
             "id": "player_id",

{wsba_hockey-1.2.3 → wsba_hockey-1.2.5}/src/wsba_hockey/wsba_main.py

@@ -85,7 +85,7 @@ CONVERT_TEAM_ABBR = {'L.A':'LAK',
                      'T.B':'TBL',
                      'PHX':'ARI'}
 
-PER_SIXTY = ['Fi','xGi','Gi','A1','A2','P1','P','Si','OZF','NZF','DZF','FF','FA','xGF','xGA','GF','GA','SF','SA','CF','CA','HF','HA','Give','Take','Penl','Penl2','Penl5','Draw','Block','GSAx']
+PER_SIXTY = ['Fi','xGi','Gi','A1','A2','P1','P','Si','OZF','NZF','DZF','FF','FA','xGF','xGA','GF','GA','SF','SA','CF','CA','HF','HA','Give','Take','Penl','Penl2','Penl5','Draw','PIM','Block','GSAx']
 
 #Some games in the API are specifically known to cause errors in scraping.
 #This list is updated as frequently as necessary
@@ -289,7 +289,7 @@ def nhl_scrape_schedule(season:int, start:str = '', end:str = ''):
             A DataFrame containing the schedule data for the specified season and date range.
     """
 
-    api = "https://api-web.nhle.com/v1/schedule/"
+    api = "https://api-web.nhle.com/v1/score/"
 
     #If either start or end are blank then find start and endpoints for specified season
     if start == '' or end == '':
@@ -325,7 +325,7 @@ def nhl_scrape_schedule(season:int, start:str = '', end:str = ''):
         print(f'Scraping games on {str(inc)[:10]}...')
         
         get = rs.get(f'{api}{str(inc)[:10]}').json()
-        gameWeek = pd.json_normalize(list(pd.json_normalize(get['gameWeek'])['games'])[0])
+        gameWeek = pd.json_normalize(get['games']).drop(columns=['goals'],errors='ignore')
         
         #Return nothing if there's nothing
         if gameWeek.empty:
@@ -382,35 +382,54 @@ def nhl_scrape_season(season:int, split_shifts:bool = False, season_types:list[i
     """
      
     #Determine whether to use schedule data in repository or to scrape
+    local_failed = False
+
     if local:
-        load = pd.read_csv(local_path)
-        load['date'] = pd.to_datetime(load['date'])
-
-        if start == '' or end == '':
-            season_data = rs.get('https://api.nhle.com/stats/rest/en/season').json()['data']
-            season_data = [s for s in season_data if s['id'] == season][0]
-            start = season_data['startDate'][0:10]
-            end = season_data['endDate'][0:10]
-            
+        try:
+            load = pd.read_csv(local_path)
+            load['game_date'] = pd.to_datetime(load['game_date'])
+
+            if start == '' or end == '':
+                season_data = rs.get('https://api.nhle.com/stats/rest/en/season').json()['data']
+                season_data = [s for s in season_data if s['id'] == season][0]
+                
+                season_start = season_data['startDate'][0:10]
+                season_end = season_data['endDate'][0:10]
+
+            else:   
+                season_start = f'{(str(season)[0:4] if int(start[0:2])>=9 else str(season)[4:8])}-{start[0:2]}-{start[3:5]}'
+                season_end =  f'{(str(season)[0:4] if int(end[0:2])>=9 else str(season)[4:8])}-{end[0:2]}-{end[3:5]}'
+
             form = '%Y-%m-%d'
 
             #Create datetime values from dates
-            start = datetime.strptime(start,form)
-            end = datetime.strptime(end,form)
-
-        else:   
-            start = f'{(str(season)[0:4] if int(start[0:2])>=9 else str(season)[4:8])}-{start[0:2]}-{start[3:5]}'
-            end =  f'{(str(season)[0:4] if int(end[0:2])>=9 else str(season)[4:8])}-{end[0:2]}-{end[3:5]}'
+            start_date = datetime.strptime(season_start,form)
+            end_date = datetime.strptime(season_end,form)
+
+            load = load.loc[(load['season']==season)&
+                            (load['season_type'].isin(season_types))&
+                            (load['game_date']>=start_date)&(load['game_date']<=end_date)&
+                            (load['game_schedule_state']=='OK')&
+                            (load['game_state']!='FUT')
+                            ]
             
+            game_ids = load['game_id'].to_list()
+        except KeyError:
+            #If loading games locally fails then force a scrape
+            local_failed = True
+            print('Loading games locally has failed.  Loading schedule data with a scrape...')
+    else:
+        local_failed = True
+
+    if local_failed:
+        load = nhl_scrape_schedule(season,start,end)
         load = load.loc[(load['season']==season)&
                         (load['season_type'].isin(season_types))&
-                        (load['game_date']>=start)&(load['game_date']<=end)]
+                        (load['game_schedule_state']=='OK')&
+                        (load['game_state']!='FUT')
+                        ]
         
         game_ids = load['game_id'].to_list()
-    else:
-        load = nhl_scrape_schedule(season,start,end)
-        load = load.loc[(load['season']==season)&(load['season_type'].isin(season_types))]
-        game_ids = load['game_id'].to_list()
 
     #If no games found, terminate the process
     if not game_ids:
@@ -504,8 +523,11 @@ def nhl_scrape_standings(arg:int | list[int] | Literal['now'] = 'now', season_ty
             data = rs.get(api).json()['series']
             dfs.append(pd.json_normalize(data))
 
+        #Combine and standardize columns
+        df = pd.concat(dfs).rename(columns=COL_MAP['standings'])
+
         #Return: playoff bracket
-        return pd.concat(dfs)
+        return df[[col for col in COL_MAP['standings'].values() if col in df.columns]]
 
     else:
         if arg == "now":
@@ -664,6 +686,10 @@ def nhl_scrape_player_info(player_ids:list[int]):
     """
 
     print(f'Retreiving player information for {player_ids}...')
+
+    #Wrap game_id in a list if only a single game_id is provided
+    player_ids = [player_ids] if type(player_ids) != list else player_ids
+
     infos = []
     for player_id in player_ids:
         player_id = int(player_id)

{wsba_hockey-1.2.3 → wsba_hockey-1.2.5/src/wsba_hockey.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: wsba_hockey
-Version: 1.2.3
+Version: 1.2.5
 Summary: WeakSide Breakout's complete Python package of access to hockey data, primairly including the scraping of National Hockey League schedule, play-by-play, and shifts information.
 Author-email: Owen Singh <owenbksingh@gmail.com>
 Project-URL: Homepage, https://github.com/owensingh38/wsba_hockey/