feat: 10장 주식 데이터 조회 추가

example/10-stock-data.py

@@ -0,0 +1,119 @@
+import logging
+import re
+from io import BytesIO
+
+import numpy as np
+import pandas as pd
+import requests as rq
+from bs4 import BeautifulSoup
+
+GEN_OTP_URL = 'http://data.krx.co.kr/comm/fileDn/GenerateOTP/generate.cmd'
+DOWN_URL = 'http://data.krx.co.kr/comm/fileDn/download_csv/download.cmd'
+
+# 최근 영업일을 가져옴
+def get_latest_biz_day():
+    url = 'https://finance.naver.com/sise/sise_deposit.nhn'
+    data = rq.post(url)
+    data_html = BeautifulSoup(data.content, 'lxml')
+    parse_day = data_html.select_one('div.subtop_sise_graph2 > ul.subtop_chart_note > li > span.tah').text
+    biz_day = re.findall('[0-9]+', parse_day)
+    biz_day = ''.join(biz_day)
+    return biz_day
+
+
+# 업종 분류 현황 가져옴
+def get_stock_data(biz_day, mkt_id):
+    # logging.basicConfig(level=logging.DEBUG)
+    gen_otp_data = {
+        'locale': 'ko_KR',
+        'mktId': mkt_id, # STK: 코스피, KSQ: 코스닥
+        'trdDd': biz_day,
+        'money': '1',
+        'csvxls_isNo': 'false',
+        'name': 'fileDown',
+        'url': 'dbms/MDC/STAT/standard/MDCSTAT03901'
+    }
+    headers = {
+        'Referer': 'http://data.krx.co.kr/contents/MDC/MDI/mdiLoader/index.cmd?menuId=MDC0201050201',
+        'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/131.0.0.0 Safari/537.36'
+    }
+
+    otp = rq.post(url=GEN_OTP_URL, data=gen_otp_data, headers=headers, verify=False)
+    # # 요청 디버깅
+    # print("===== Request Details =====")
+    # print(f"Method: {otp.request.method}")
+    # print(f"URL: {otp.request.url}")
+    # print(f"Headers: {otp.request.headers}")
+    # print(f"Body: {otp.request.body}")
+    #
+    # # 응답 디버깅
+    # print("===== Response Details =====")
+    # print(f"Status Code: {otp.status_code}")
+    # print(f"Headers: {otp.headers}")
+    # print(f"Body: {otp.text}")
+
+    down_sector = rq.post(url=DOWN_URL, data={'code': otp.text}, headers=headers)
+    return pd.read_csv(BytesIO(down_sector.content), encoding='EUC-KR')
+
+
+# 개별 지표 조회
+def get_ind_stock_data(biz_day):
+    gen_otp_data = {
+        'locale': 'ko_KR',
+        'searchType': '1',
+        'mktId': 'ALL',
+        'trdDd': biz_day,
+        'csvxls_isNo': 'false',
+        'name': 'fileDown',
+        'url': 'dbms/MDC/STAT/standard/MDCSTAT03501'
+    }
+    headers = {
+        'Referer': 'http://data.krx.co.kr/contents/MDC/MDI/mdiLoader/index.cmd?menuId=MDC0201050201',
+        'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/131.0.0.0 Safari/537.36'
+    }
+
+    otp = rq.post(url=GEN_OTP_URL, data=gen_otp_data, headers=headers, verify=False)
+
+    down_ind_sector = rq.post(url=DOWN_URL, data={'code': otp.text}, headers=headers)
+    return pd.read_csv(BytesIO(down_ind_sector.content), encoding='EUC-KR')
+
+
+def get_total_stock_data(biz_day):
+    # 업종 분류 현황(코스피, 코스닥)
+    sector_stk = get_stock_data(biz_day, 'stk')
+    sector_ksq = get_stock_data(biz_day, 'ksq')
+    # 각각 조회 후 합침
+    krx_sector = pd.concat([sector_stk, sector_ksq]).reset_index(drop=True)
+    krx_sector['종목명'] = krx_sector['종목명'].str.strip()
+    krx_sector['기준일'] = biz_day
+
+    # 개별 지표 조회
+    krx_ind = get_ind_stock_data(biz_day)
+    krx_ind['종목명'] = krx_ind['종목명'].str.strip()
+    krx_ind['기준일'] = biz_day
+
+    # 데이터 정리
+    # 종목, 개별 중 한군데만 있는 데이터 삭제(선박펀드, 광물펀드, 해외종목 등)
+    diff = list(set(krx_sector['종목명']).symmetric_difference(set(krx_ind['종목명'])))
+    kor_ticket = pd.merge(krx_sector, krx_ind, on=krx_sector.columns.intersection(krx_ind.columns).tolist(), how='outer')
+    # 일반적인 종목과 SPAC, 우선주, 리츠, 기타 주식을 구분
+    kor_ticket['종목구분'] = np.where(kor_ticket['종목명'].str.contains('스팩|제[0-9]+호'), '스팩',
+            np.where(kor_ticket['종목코드'].str[-1:] != '0', '우선주',
+                     np.where(kor_ticket['종목명'].str.endswith('리츠'), '리츠',
+                              np.where(kor_ticket['종목명'].isin(diff), '기타',
+                                       '보통주'
+            ))))
+    kor_ticket = kor_ticket.reset_index(drop=True)
+    kor_ticket.columns = kor_ticket.columns.str.replace(' ', '')
+    kor_ticket = kor_ticket[['종목코드', '종목명', '시장구분', '종가',
+                             '시가총액', '기준일', 'EPS', '선행EPS', 'BPS', '주당배당금', '종목구분']]
+    kor_ticket = kor_ticket.replace({np.nan: None})
+    kor_ticket['기준일'] = pd.to_datetime(kor_ticket['기준일'])
+    return kor_ticket
+
+if __name__ == '__main__':
+    latest_biz_day = get_latest_biz_day()
+    data = get_total_stock_data(latest_biz_day)
+    print(data)
+
+