競馬予想アプリ

競馬が好きなので予想アプリを作ってみました！！

pythonでスクレイピングをして作りました。

keiba.py

import streamlit as st
import pandas as pd
import requests
from bs4 import BeautifulSoup
import re
import time


def get_url(kai,day,l,get_num):
    tg_year = 2022
    df_race_li=[]
    url_ids = []

    for year in range(tg_year-1, tg_year+1):
        for z in range(1, 6):#開催、6まで十分だけど保険で7
            for y in range(1, 3):#日、12まで十分だけど保険で14
                    url_id = f"{year}{l}{z:0>2}{y:0>2}"
                    url_ids.append(url_id)

    tg_id = f"{tg_year}{l}{kai:0>2}{day:0>2}"
    tg_ids = [u for u in url_ids if u < tg_id]

    for race_id in reversed(tg_ids[-get_num:]):
        for x in range(1, 13):
            url=f"https://race.netkeiba.com/race/shutuba.html?race_id={race_id}{x:0>2}"
            print(url)
            r=requests.get(url)
            time.sleep(0.2)#サーバーの負荷を減らすため0.2秒待機する
            soup = BeautifulSoup(r.content, "html.parser")   
            race=soup.find_all("div",class_="RaceList_Item02")
            if len(race)==0:
                print("break")
                break
            racename=race[0].find_all("div",class_="RaceName")[0].text.strip()
            racedata=race[0].find_all("div",class_="RaceData01")[0].text.strip()
            racedata2=race[0].find_all("div",class_="RaceData02")[0].text.strip()
            race=soup.find_all("tr",class_="HorseList")
            uma_li=[]
            hito_li=[]
            for i in range(0,len(race)):
                uma=race[i].find_all("td",class_="HorseInfo")[0].text.strip()
                uma_li.append(uma)
                hito=race[i].find_all("td",class_="Jockey")[0].text.strip()
                hito_li.append(hito)
            df_race=pd.DataFrame({"uma":uma_li,"hito":hito_li})
            df_race["racename"]=racename
            df_race["racedata1"]=racedata
            df_race["racedata2"]=racedata2  
            df_race["racecourse"]=race_id[4:6]
            df_race["racenumber"]=race_id[6:8]
            df_race["racedate"]=race_id[8:10]   
            df_race["racenumber2"]=x
            df_race = df_race.reset_index()
            df_race_li.append(df_race)
    df_race_all = pd.concat(df_race_li)
    return df_race_all

url = "https://www.jra.go.jp"
top_r = requests.get(url)
top_soup = BeautifulSoup(top_r.content, "html.parser")

link_dict ={}
for i in range(1, 10):
    top_select = top_soup.select(f"#kaisai_area > div.kaisai_block > div:nth-child({i}) > div.main > div > div.race > p > a")
    if len(top_select)==0:
        break
    race_name = top_select[0].text
    link = top_select[0].get('href')
    link = f"https://www.jra.go.jp/{link}"
    link_dict[race_name] = link

st.title("競馬")
st.markdown("### 今週の注目レース")
for name, link in link_dict.items():
    if st.button(f'{name}予想'):
        url_race = link
        if "syutsuba.html" in url_race:
            syutsuba = requests.get(url_race)
            syutsuba_soup = BeautifulSoup(syutsuba.content, "html.parser")
            syutsuba_result = syutsuba_soup.select("#syutsuba")
            racejyou = syutsuba_result[0].find_all("div",class_="cell date")[0].text.strip().split(" ")[1]
            cap=syutsuba_result[0].find_all("div",class_="cell course")[0].text.strip()
            uma_td=syutsuba_result[0].find_all("td",class_="horse")
            hito_td=syutsuba_result[0].find_all("td",class_="jockey")
            uma_li=[]
            hito_li=[]
            for i in range(0,len(uma_td)):
                uma=uma_td[i].text.strip().split()[0]
                uma_li.append(uma)
                hito=" ".join(re.split("\n|\.| ", hito_td[i].text.strip())[-2:])
                hito_li.append(hito)
            df_race=pd.DataFrame({"uma":uma_li,"hito":hito_li})
            kai = re.findall("\d*回", racejyou)[0].replace("回", "")
            day = re.findall("\d*日", racejyou)[0].replace("日", "")
            race_jyou=re.sub("\d*日|\d*回", "", racejyou)
            race_jyou_dict = {
                "札幌":"01",
                "函館":"02",
                "福島":"03",
                "新潟":"04",
                "東京":"05",
                "中山":"06",
                "中京":"07",
                "京都":"08",
                "阪神":"09",
                "小倉":"10"
                }
            l=race_jyou_dict[race_jyou]
            st.table(df_race)
            st.markdown(racejyou)
            st.markdown(cap)
            
            st.markdown("### 過去の成績検索")
            s = cap.replace("コース：", "").replace(",", "")
            mm = int(re.search("\d*",s).group())
            shiba_dato = re.search("（.*）",s).group()[1]
            shiba_dato= "ダ" if shiba_dato =="ダート" else shiba_dato
            df_race_all=get_url(kai, day, l, 3)
            tg_race=df_race_all[df_race_all["racedata1"].str.contains(f".*{shiba_dato}.*{mm}")]
            d=tg_race[tg_race["index"] <=2][["index","uma","hito","racename", "racenumber", "racedate" , "racenumber2"]]
            st.subheader('過去の成績')
            st.table(d)
        else:
            st.markdown("出馬情報がありません。")

実行方法

 streamlit run keiba.py

参考

https://introduction1.com/2021/09/30/python-scrp1/
https://www.jra.go.jp
https://www.netkeiba.com