From 21f63546585d8e2028001d3941b6fef41c8a13b9 Mon Sep 17 00:00:00 2001 From: muliangshu <283299495@qq.com> Date: Wed, 24 Dec 2025 08:18:19 +0800 Subject: [PATCH] =?UTF-8?q?stock=5Fnew=E6=9B=B4=E6=96=B0?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- dags/ai/stock/stock_new.py | 95 ++++++++++++++++++++++++++++++++++++++ 1 file changed, 95 insertions(+) create mode 100644 dags/ai/stock/stock_new.py diff --git a/dags/ai/stock/stock_new.py b/dags/ai/stock/stock_new.py new file mode 100644 index 0000000..8d985b5 --- /dev/null +++ b/dags/ai/stock/stock_new.py @@ -0,0 +1,95 @@ +import datetime +from sqlalchemy import create_engine +from dateutil.relativedelta import relativedelta +import pandas as pd +from sklearn.preprocessing import StandardScaler +from sklearn.mixture import GaussianMixture +import numpy as np +from collections import defaultdict +engine = create_engine('mysql+pymysql://root:Goojia123456@192.168.0.93:3306/stock') +mu_stocks_all = pd.read_sql(f"select * from mu_stocks_new", con = engine) +stock_basic_infos = pd.read_sql(f"select * from stock_basic", con=engine) +trade_dates = mu_stocks_all['trade_date'].tolist() +today = datetime.datetime.now() +def gmm_trade(get_data,df): + gmm = GaussianMixture(n_components=100, covariance_type='spherical', random_state=0) + data = df.values + data = np.array(data, dtype='float64') + x = data + gmm.fit(x) + y_hat = gmm.predict(x) + gmm_means_ = [i[0] for i in gmm.means_] + d = defaultdict(list) + for k, va in [(v, i) for i, v in enumerate(gmm_means_)]: + d[va] = k + gmm_means_list = [] + for label in y_hat: + mean_value = d[label] + gmm_means_list.append(mean_value) + get_data["rating"] = y_hat + get_data["mean"] = gmm_means_list + return get_data +def stock_gmm(last_day_str,last_month_str): + start = datetime.datetime.now() + stock_all = pd.read_sql(f"select * from stock_index_daily where trade_date>'{last_month_str}' and trade_date<='{last_day_str}'", con = engine) + end = datetime.datetime.now() + print('MySQL Query Running time: %s Seconds' % (end - start)) + stock_all = stock_all.sort_values(by="trade_date", ascending=False) + stock_select = stock_all[((stock_all['ts_code'].str.startswith('600')) | (stock_all['ts_code'].str.startswith('00')))] + stock_data = stock_select.dropna(subset=['turnover_rate_f', 'volume_ratio', 'pe', 'pe_ttm', 'pb', 'ps_ttm', 'ps']) + stock_group = stock_data.groupby(['ts_code']) + scaler = StandardScaler() + import_col = ['turnover_rate_f', 'volume_ratio', 'pe', 'pe_ttm', 'pb', 'ps_ttm', 'ps'] + scaler_col = [str(i) for i in range(len(import_col))] + res = [] + for ts_code, ts_detail in stock_group: + if len(ts_detail) > 3: + ts_detail[scaler_col] = ts_detail.apply(lambda x: (0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0), axis=1, + result_type="expand") + ts_detail[scaler_col] = scaler.fit_transform(ts_detail[import_col].values) + for i, row in ts_detail.iterrows(): + res.append(row) + stock_res = pd.DataFrame(res) + stock_all_newest = stock_all[stock_all.trade_date == stock_all.trade_date.max()] + stock_all_newest = stock_all_newest[['ts_code', 'turnover_rate', 'close', 'amount', 'pe', 'pe_ttm', 'total_mv', 'pct_chg']] + X = stock_res[['0', '1', '2', '3', '4', '5', '6']] + gmm_df = gmm_trade(stock_res, X) + gmm_group_df = gmm_df.groupby(['ts_code', 'rating'])['stock_name'].count().reset_index() + gmm_group_df.rename(columns={'stock_name': 'count'}, inplace=True) + gmm_df = gmm_df.merge(gmm_group_df, how='left', on=['ts_code', 'rating']) + min_mean = gmm_df['mean'].min() + slecct_df = gmm_df[gmm_df['mean'] == min_mean] + count_max = slecct_df['count'].max() + slecct_df = slecct_df[slecct_df['count'] == count_max] + slecct_df.drop_duplicates(subset=["ts_code"], keep="first", inplace=True) + block_list = list(slecct_df['stock_name']) + select_block = [] + for name in block_list: + status = 1 + for i in ['N', 'ST']: + if i in name: + status = 0 + if status == 1: + select_block.append(name) + return stock_all_newest,select_block,slecct_df + +def stock_select(i): + last_day = today+relativedelta(days=-i) + last_month = last_day+relativedelta(months=-2) + last_day_str = last_day.strftime('%Y%m%d') + if last_day_str in trade_dates: + return last_day_str + if int(last_day_str)<20250101: + return last_day_str + last_month_str = last_month.strftime('%Y%m%d') + stock_all_newest,select_block,slecct_df=stock_gmm(last_day_str,last_month_str) + db_stock = slecct_df[['ts_code', 'stock_name', 'open', 'close', 'high', 'low', + 'vol', 'amount', 'pre_close', 'change', 'pct_chg', 'turnover_rate', + 'turnover_rate_f', 'volume_ratio', 'pe', 'pe_ttm', 'pb', 'ps_ttm', 'ps', + 'total_share', 'float_share', 'free_share', 'total_mv', 'circ_mv']].copy() + db_stock['trade_date'] = last_day_str + db_stock['ts_code'] = db_stock['ts_code'].apply(lambda x: x[0:6]) + db_stock.to_sql("mu_stocks_new",con=engine, index=False, if_exists="append") + return last_day_str +if __name__ == '__main__': + stock_select(0) \ No newline at end of file