Added NPB Schedule Data

.github/workflows/get_npb_schedules.yml

@@ -1,9 +1,9 @@
-name: Update NPB Schedules
+name: Update NPB Data
 on:
   schedule:
     - cron: "0 12 * * *"
-  # push:
-  #   branches: [main]
+  push:
+    branches: [main]
   # pull_request:
   #   branches: [main]
 
@@ -34,6 +34,7 @@ jobs:
       - name: run Python Script
         run: |
           python get_npb_schedule.py
+          python get_npb_standings.py
 
       - uses: xresloader/upload-to-github-release@main
         env:
@@ -44,4 +45,14 @@ jobs:
           #update_latest_release: true
           overwrite: true
           verbose: true
-          default_release_name: "Schedule"
+          default_release_name: "Schedule"
+      - uses: xresloader/upload-to-github-release@main
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        with:
+          file: "standings/*"
+          branches: "main"
+          #update_latest_release: true
+          overwrite: true
+          verbose: true
+          default_release_name: "Game_Standings"

get_npb_schedule.py

@@ -1,8 +1,6 @@
-from datetime import datetime, timedelta
+from datetime import datetime
 import time
 import pandas as pd
-import requests
-from tqdm import tqdm
 
 
 from utls import get_json_from_url
@@ -13,65 +11,86 @@ def get_npb_schedule(season: int, save_results=False):
 
     """
     sched_df = pd.DataFrame()
-    row_df = pd.DataFrame()
+    columns = [
+        'season',
+        'ID',
+        'SeqNo',
+        'game_id',
+        'game_kind_id',
+        'date_jpn',
+        'time_jpn',
+        'week_day_jpn',
+        'stadium_id',
+        'stadium_name_jpn',
+        'round',
+        'DhF',
+        'game_state',
+        'game_result',
+
+        'home_score',
+        'home_team_id',
+        'home_team_short_name',
+        'home_team_name_en',
+        'home_team_name_en_short',
+        'home_team_initial',
+        'home_section',
+        'home_text_area',
+
+        'away_score',
+        'away_team_id',
+        'away_team_short_name',
+        'away_team_name_en',
+        'away_team_name_en_short',
+        'away_team_initial',
+        'away_section',
+        'away_text_area',
+
+        'stadium_name_short',
+        'last_updated',
+        'creation_date'
+    ]
+
     url = f"https://spaia.jp/baseball/npb/api/schedules?Year={season}"
     json_data = get_json_from_url(url=url)
 
-    for game in tqdm(json_data):
-        game_id = game['GameID']
-        row_df = pd.DataFrame(
-            {"game_id":game_id},
-            index=[0]
-        )
-        row_df['game_kind_id'] = game['Year']
-
-        date_str = f"{game['DateJPN']} {game['TimeJPN']}"
-        game_date = datetime.strptime(date_str, "%Y%m%d %H%M")
-        del date_str
-        row_df['season'] = game['Year']
-        row_df['game_datetime_jpn'] = game_date
-        row_df['game_datetime_iso'] = game_date - timedelta(hours=9)
-        row_df['stadium_id'] = game['StadiumID']
-        row_df['stadium_name_jp'] = game['StadiumName']
-        row_df['round'] = game['Round']
-        row_df['dhf'] = game['DhF']
-        row_df['game_state'] = game['GameState']
-        row_df['game_result'] = game['GameResult']
-        row_df['home_score'] = game['HScore']
-        row_df['away_score'] = game['VScore']
-
-        row_df['home_team_id'] = game['HTeamID']
-        row_df['home_team_short_name'] = game['HTeamNameS']
-        row_df['home_team_en_name'] = game['HomeTeamNameE']
-        row_df['home_team_initial'] = game['VisitorTeamInitial']
-        row_df['home_team_en_initial'] = game['HomeTeamNameES']
-
-        row_df['away_team_id'] = game['VTeamID']
-        row_df['away_team_short_name'] = game['VTeamNameS']
-        row_df['away_team_en_name'] = game['VisitorTeamNameE']
-        row_df['away_team_initial'] = game['VisitorTeamInitial']
-        row_df['away_team_en_initial'] = game['VisitorTeamNameES']
-
-        row_df['stadium_short_name'] = game['StadiumNameS'] # ?
+    sched_df = pd.DataFrame(json_data)
 
-        row_df['home_wins'] = game['Win']
-        row_df['home_losses'] = game['Lose']
-        row_df['home_draws'] = game['Draw']
-        row_df['home_batting_avg'] = game['Avg']
-        row_df['home_pitching_era'] = game['Era']
-        row_df['home_section'] = game['Home_Section']
-        row_df['home_jp_description'] = game['Home_TextArea']
+    sched_df.rename(
+        columns={
+            "GameID": "game_id",
+            "GameKindID": "game_kind_id",
+            "DateJPN": "date_jpn",
+            "TimeJPN": "time_jpn",
+            "WeekDayJPN": "week_day_jpn",
+            "StadiumID": "stadium_id",
+            "StadiumName": "stadium_name_jpn",
+            "Round": "round",
+            "GameState": "game_state",
+            "GameResult": "game_result",
+            "HScore": "home_score",
+            "VScore": "away_score",
+            "HTeamID": "home_team_id",
+            "HTeamNameS": "home_team_short_name",
+            "VTeamID": "away_team_id",
+            "VTeamNameS": "away_team_short_name",
+            "StadiumNameS": "stadium_name_short",
+            "HomeTeamNameE": "home_team_name_en",
+            "HomeTeamNameES": "home_team_name_en_short",
+            "HomeTeamInitial": "home_team_initial",
+            "VisitorTeamNameE": "away_team_name_en",
+            "VisitorTeamNameES": "away_team_name_en_short",
+            "VisitorTeamInitial": "away_team_initial",
+            "Home_Section": "home_section",
+            "Home_TextArea": "home_text_area",
+            "Visitor_Section": "away_section",
+            "Visitor_TextArea": "away_text_area",
+            "UpdatedAt": "last_updated",
+            "Year": "season",
+            "CreatedAt": "creation_date"
+        }, inplace=True
+    )
 
-        row_df['away_wins'] = game['Visitor_Win']
-        row_df['away_losses'] = game['Visitor_Lose']
-        row_df['away_draws'] = game['Visitor_Draw']
-        row_df['away_batting_avg'] = game['Visitor_Avg']
-        row_df['away_pitching_era'] = game['Visitor_Era']
-        row_df['away_section'] = game['Visitor_Section']
-        row_df['away_jp_description'] = game['Visitor_TextArea']
-
-        row_df['last_updated'] = game['UpdatedAt']
-        sched_df = pd.concat([sched_df,row_df],ignore_index=True)
+    sched_df = sched_df[columns]
 
     if save_results == True:
         sched_df.to_csv(f"schedules/{season}_npb_schedule.csv", index=False)
@@ -88,4 +107,5 @@ def get_npb_schedule(season: int, save_results=False):
             season=i,
             save_results=True
         )
+        print(df)
         time.sleep(1)

get_npb_standings.py

@@ -0,0 +1,46 @@
+from datetime import datetime
+import time
+import pandas as pd
+from tqdm import tqdm
+
+
+from utls import get_json_from_url
+
+
+def get_npb_standings_by_game(season: int, save_results=False):
+    """
+    """
+    standings_df = pd.DataFrame()
+    cl_df = pd.DataFrame()
+    pl_df = pd.DataFrame()
+
+    url_1 = f"https://spaia.jp/baseball/npb/api/official_stats_history?GameAssortment=1&Year={season}"
+    url_2 = f"https://spaia.jp/baseball/npb/api/official_stats_history?GameAssortment=2&Year={season}"
+
+    # Central League (1)
+    json_data = get_json_from_url(url=url_1)
+    cl_df = pd.DataFrame(json_data)
+
+    # Pacific League (2)
+    json_data = get_json_from_url(url=url_2)
+    pl_df = pd.DataFrame(json_data)
+
+    standings_df = pd.concat([cl_df, pl_df], ignore_index=True)
+    if save_results == True:
+        standings_df.to_csv(f"standings/{season}_game_standings.csv",index=False)
+        standings_df.to_parquet(
+            f"standings/{season}_game_standings.parquet",index=False)
+
+    return standings_df
+
+
+if __name__ == "__main__":
+    now = datetime.now()
+    f_year = now.year - 2
+    c_year = now.year + 1
+    for i in tqdm(range(f_year,c_year)):
+        df =get_npb_standings_by_game(
+            season=i,
+            save_results=True
+        )
+    print(df)

spaia_api_endpoints.md

@@ -36,7 +36,7 @@ Base: https://spaia.jp/baseball/npb/api/related_players?player_id=1400101
 ## Same Draft Year
 Base: https://spaia.jp/baseball/npb/api/same_draft_year_players?player_id=1400101
 
-# Schedule
+# Schedule (done)
 
 ## Upcomming Games
 Base: https://spaia.jp/baseball/npb/api/game_schedule
@@ -64,7 +64,7 @@ Base: https://spaia.jp/baseball/npb/api/official_stats_history?GameAssortment=1&
 # Teams
 
 ## Full Team Rosters
-Base: https://spaia.jp/baseball/npb/api/directory?TeamID=2
+Base: https://spaia.jp/baseball/npb/api/directory?TeamID=2&year=2023
 
 ## Batters
 Base: https://spaia.jp/baseball/npb/api/batter_list?team=2&year=2023
@@ -73,10 +73,10 @@ Base: https://spaia.jp/baseball/npb/api/batter_list?team=2&year=2023
 Base: https://spaia.jp/baseball/npb/api/pitcher_list?team=2&year=2023
 
 ## Coaching Staff
-Base: https://spaia.jp/baseball/npb/api/staff_list?team=2
+Base: https://spaia.jp/baseball/npb/api/staff_list?team=2&year=2023
 
 ## Full System: 
-Base: https://spaia.jp/baseball/npb/api/players_by_team?team_id=2
+Base: https://spaia.jp/baseball/npb/api/players_by_team?team_id=2&year=2023
 
 # NPB Draft