From bb3caa0cd41bd56f496125934491759b4d865733 Mon Sep 17 00:00:00 2001
From: rp <rp@outlook.com>
Date: 星期三, 09 四月 2025 21:58:26 +0800
Subject: [PATCH] 修复备份
---
xg_优化备份.py | 611 +++++++++++++++++++++++++++++++++++++++++++++++++++++++
1 files changed, 611 insertions(+), 0 deletions(-)
diff --git "a/xg_\344\274\230\345\214\226\345\244\207\344\273\275.py" "b/xg_\344\274\230\345\214\226\345\244\207\344\273\275.py"
new file mode 100644
index 0000000..7e8061c
--- /dev/null
+++ "b/xg_\344\274\230\345\214\226\345\244\207\344\273\275.py"
@@ -0,0 +1,611 @@
+import os
+import pickle
+import pandas as pd
+import numpy as np
+import tkinter as tk
+import tkinter.font as tkfont
+from tkinter import ttk
+from datetime import timedelta
+from time import time
+import matplotlib.pyplot as plt
+from matplotlib.backends.backend_tkagg import FigureCanvasTkAgg, NavigationToolbar2Tk
+from xgboost import XGBRegressor
+from lunardate import LunarDate
+from sklearn.model_selection import train_test_split
+from sklearn.metrics import mean_squared_error, mean_absolute_error
+import matplotlib
+
+# 閰嶇疆 matplotlib 涓枃鏄剧ず
+matplotlib.rcParams['font.sans-serif'] = ['SimHei', 'Microsoft YaHei', 'SimSun', 'Arial Unicode MS']
+matplotlib.rcParams['axes.unicode_minus'] = False
+matplotlib.rcParams['font.family'] = 'sans-serif'
+
+# 鍏ㄥ眬缂撳瓨鍙橀噺鍙婄壒寰佸悕绉帮紙姝ゅ feature_columns 浠呬负鍗犱綅锛�
+cached_model = None
+last_training_time = None
+feature_columns = None
+
+# -------------------------------
+# 鏁版嵁鍔犺浇涓庨澶勭悊鍑芥暟
+# -------------------------------
+def load_data(upstream_file, downstream_file, qinglong_lake_file=None):
+ try:
+ upstream_df = pd.read_csv(upstream_file)
+ downstream_df = pd.read_csv(downstream_file)
+ if qinglong_lake_file:
+ qinglong_lake_df = pd.read_csv(qinglong_lake_file)
+ except FileNotFoundError:
+ print("鏂囦欢鏈壘鍒帮紝璇锋鏌ヨ矾寰�")
+ return None
+
+ # 鍋囪鍘熷鏁版嵁鍒椾緷娆′负 ['DateTime', 'TagName', 'Value']
+ upstream_df.columns = ['DateTime', 'TagName', 'Value']
+ downstream_df.columns = ['DateTime', 'TagName', 'Value']
+ if qinglong_lake_file:
+ qinglong_lake_df.columns = ['DateTime', 'TagName', 'Value']
+
+ # 杞崲鏃堕棿鏍煎紡鍙婃暟鍊煎鐞�
+ upstream_df['DateTime'] = pd.to_datetime(upstream_df['DateTime'])
+ downstream_df['DateTime'] = pd.to_datetime(downstream_df['DateTime'])
+ if qinglong_lake_file:
+ qinglong_lake_df['DateTime'] = pd.to_datetime(qinglong_lake_df['DateTime'])
+ upstream_df['Value'] = pd.to_numeric(upstream_df['Value'], errors='coerce')
+ downstream_df['Value'] = pd.to_numeric(downstream_df['Value'], errors='coerce')
+ if qinglong_lake_file:
+ qinglong_lake_df['Value'] = pd.to_numeric(qinglong_lake_df['Value'], errors='coerce')
+
+ # 杩囨护鐩愬害灏忎簬5鐨勬暟鎹�
+ upstream_df = upstream_df[upstream_df['Value'] >= 5]
+ downstream_df = downstream_df[downstream_df['Value'] >= 5]
+ if qinglong_lake_file:
+ qinglong_lake_df = qinglong_lake_df[qinglong_lake_df['Value'] >= 5]
+
+ # 灏�0鏇挎崲涓篘aN锛屽苟鍒╃敤3鍊嶆爣鍑嗗樊娉曞鐞嗗紓甯稿��
+ for df in [upstream_df, downstream_df]:
+ df.loc[df['Value'] == 0, 'Value'] = np.nan
+ mean_val, std_val = df['Value'].mean(), df['Value'].std()
+ lower_bound, upper_bound = mean_val - 3 * std_val, mean_val + 3 * std_val
+ df.loc[(df['Value'] < lower_bound) | (df['Value'] > upper_bound), 'Value'] = np.nan
+ if qinglong_lake_file:
+ qinglong_lake_df.loc[qinglong_lake_df['Value'] == 0, 'Value'] = np.nan
+ mean_val, std_val = qinglong_lake_df['Value'].mean(), qinglong_lake_df['Value'].std()
+ lower_bound, upper_bound = mean_val - 3 * std_val, mean_val + 3 * std_val
+ qinglong_lake_df.loc[(qinglong_lake_df['Value'] < lower_bound) | (qinglong_lake_df['Value'] > upper_bound), 'Value'] = np.nan
+
+ # 閲嶅懡鍚� Value 鍒楀苟淇濈暀闇�瑕佺殑鍒�
+ upstream_df = upstream_df.rename(columns={'Value': 'upstream'})[['DateTime', 'upstream']]
+ downstream_df = downstream_df.rename(columns={'Value': 'downstream'})[['DateTime', 'downstream']]
+ if qinglong_lake_file:
+ qinglong_lake_df = qinglong_lake_df.rename(columns={'Value': 'qinglong_lake'})[['DateTime', 'qinglong_lake']]
+
+ # 鍚堝苟鏁版嵁
+ merged_df = pd.merge(upstream_df, downstream_df, on='DateTime', how='inner')
+ if qinglong_lake_file:
+ merged_df = pd.merge(merged_df, qinglong_lake_df, on='DateTime', how='left')
+
+ print(f"鍚堝苟鍓嶆暟鎹鏁�: {len(merged_df)}")
+ merged_df = merged_df.set_index('DateTime')
+
+ # 鎻掑�硷細鍏堢敤绾挎�э紝鍐嶇敤鏃堕棿鎻掑�硷紝鏈�鍚庣敤鍓嶅悜鍚庡悜濉厖
+ merged_df['upstream'] = merged_df['upstream'].interpolate(method='linear', limit=4)
+ merged_df['downstream'] = merged_df['downstream'].interpolate(method='linear', limit=4)
+ if qinglong_lake_file:
+ merged_df['qinglong_lake'] = merged_df['qinglong_lake'].interpolate(method='linear', limit=4)
+ merged_df['upstream'] = merged_df['upstream'].interpolate(method='time', limit=24)
+ merged_df['downstream'] = merged_df['downstream'].interpolate(method='time', limit=24)
+ if qinglong_lake_file:
+ merged_df['qinglong_lake'] = merged_df['qinglong_lake'].interpolate(method='time', limit=24)
+ merged_df['upstream'] = merged_df['upstream'].fillna(method='ffill').fillna(method='bfill')
+ merged_df['downstream'] = merged_df['downstream'].fillna(method='ffill').fillna(method='bfill')
+ if qinglong_lake_file:
+ merged_df['qinglong_lake'] = merged_df['qinglong_lake'].fillna(method='ffill').fillna(method='bfill')
+
+ # 骞虫粦澶勭悊锛氫娇鐢ㄦ粦鍔ㄧ獥鍙ｇЩ鍔ㄥ钩鍧�
+ merged_df['upstream_smooth'] = merged_df['upstream'].rolling(window=24, min_periods=1, center=True).mean()
+ merged_df['downstream_smooth'] = merged_df['downstream'].rolling(window=24, min_periods=1, center=True).mean()
+ if qinglong_lake_file:
+ merged_df['qinglong_lake_smooth'] = merged_df['qinglong_lake'].rolling(window=24, min_periods=1, center=True).mean()
+ # 瀵逛綆鐩愬害閮ㄥ垎鐢ㄦ洿澶х獥鍙ｅ钩婊�
+ low_sal_mask = merged_df['upstream'] < 50
+ if low_sal_mask.any():
+ merged_df.loc[low_sal_mask, 'upstream_smooth'] = merged_df.loc[low_sal_mask, 'upstream']\
+ .rolling(window=48, min_periods=1, center=True).mean()
+
+ merged_df = merged_df.dropna()
+ merged_df = merged_df[merged_df['upstream'].apply(np.isfinite)]
+ merged_df = merged_df[merged_df['downstream'].apply(np.isfinite)]
+ if qinglong_lake_file:
+ merged_df = merged_df[merged_df['qinglong_lake'].apply(np.isfinite)]
+ merged_df = merged_df.reset_index()
+ print(f"娓呮礂鍚庢暟鎹鏁�: {len(merged_df)}")
+ print(f"涓婃父鐩愬害鑼冨洿: {merged_df['upstream'].min()} - {merged_df['upstream'].max()}")
+ print(f"涓嬫父鐩愬害鑼冨洿: {merged_df['downstream'].min()} - {merged_df['downstream'].max()}")
+ if qinglong_lake_file:
+ print(f"闈掗緳婀栫洂搴﹁寖鍥�: {merged_df['qinglong_lake'].min()} - {merged_df['qinglong_lake'].max()}")
+ merged_df = merged_df.sort_values('DateTime')
+ return merged_df
+
+# -------------------------------
+# 娣诲姞鍐滃巻锛堟疆姹愶級鐗瑰緛
+# -------------------------------
+def add_lunar_features(df):
+ lunar_day, lunar_phase_sin, lunar_phase_cos, is_high_tide = [], [], [], []
+ for dt in df['DateTime']:
+ ld = LunarDate.fromSolarDate(dt.year, dt.month, dt.day)
+ lunar_day.append(ld.day)
+ lunar_phase_sin.append(np.sin(2 * np.pi * ld.day / 15))
+ lunar_phase_cos.append(np.cos(2 * np.pi * ld.day / 15))
+ is_high_tide.append(1 if (ld.day <= 5 or (ld.day >= 16 and ld.day <= 20)) else 0)
+ df['lunar_day'] = lunar_day
+ df['lunar_phase_sin'] = lunar_phase_sin
+ df['lunar_phase_cos'] = lunar_phase_cos
+ df['is_high_tide'] = is_high_tide
+ return df
+
+# -------------------------------
+# 鎵归噺鐢熸垚寤惰繜鐗瑰緛锛堝悜閲忓寲锛屽埄鐢� shift锛�
+# -------------------------------
+def batch_create_delay_features(df, delay_hours):
+ for delay in delay_hours:
+ df[f'upstream_delay_{delay}h'] = df['upstream'].shift(delay)
+ df[f'downstream_delay_{delay}h'] = df['downstream'].shift(delay)
+ return df
+
+# -------------------------------
+# 鍚戦噺鍖栨瀯閫犺缁冩牱鏈紙浼樺寲鐗瑰緛宸ョ▼锛�
+# -------------------------------
+def create_features_vectorized(df, look_back=96, forecast_horizon=5):
+ """
+ 鍒╃敤 numpy 鐨� sliding_window_view 瀵瑰巻鍙茬獥鍙ｃ�佷笅娓哥獥鍙ｃ�佹爣绛捐繘琛屾壒閲忓垏鐗囷紝
+ 鍏朵粬鐗瑰緛锛堟椂闂淬�佸啘鍘嗐�佺粺璁°�佸欢杩熺壒寰侊級鐩存帴鎵归噺璇诲彇鍚庢嫾鎺�
+ """
+ # 杩欓噷瀹氫箟 total_samples 涓猴細
+ total_samples = len(df) - look_back - forecast_horizon + 1
+ if total_samples <= 0:
+ print("鏁版嵁涓嶈冻浠ュ垱寤虹壒寰�")
+ return np.array([]), np.array([])
+
+ # 鍒╃敤 sliding_window_view 鏋勯�犲巻鍙茬獥鍙ｏ紙涓婃父杩炵画 look_back 涓暟鎹級
+ upstream_array = df['upstream'].values # shape (n,)
+ # 婊戝姩绐楀彛锛岀粨鏋� shape (n - look_back + 1, look_back)
+ from numpy.lib.stride_tricks import sliding_window_view
+ window_up = sliding_window_view(upstream_array, window_shape=look_back)[:total_samples, :]
+
+ # 涓嬫父鏈�杩� 24 灏忔椂锛氬埄鐢ㄦ粦鍔ㄧ獥鍙ｆ瀯閫狅紝绐楀彛澶у皬涓� 24
+ downstream_array = df['downstream'].values
+ window_down_full = sliding_window_view(downstream_array, window_shape=24)
+ # 瀵逛簬鏍囩鍜屼笅娓哥獥鍙ｏ紝鍘熼�昏緫锛氬彇 df['downstream'].iloc[i+look_back-24:i+look_back]
+ # 鍒欏搴旂储寮曚负 i+look_back-24锛� i 浠� 0 鍒� total_samples-1
+ window_down = window_down_full[look_back-24 : look_back-24 + total_samples, :]
+
+ # 鏃堕棿鐗瑰緛涓庡啘鍘嗙壒寰佺瓑锛氬彇鏍峰尯闂翠负 df.iloc[look_back: len(df)-forecast_horizon+1]
+ sample_df = df.iloc[look_back: len(df)-forecast_horizon+1].copy()
+ basic_time = sample_df['DateTime'].dt.hour.values.reshape(-1, 1) / 24.0
+ weekday = sample_df['DateTime'].dt.dayofweek.values.reshape(-1, 1) / 7.0
+ month = sample_df['DateTime'].dt.month.values.reshape(-1, 1) / 12.0
+ basic_time_feats = np.hstack([basic_time, weekday, month])
+
+ lunar_feats = sample_df[['lunar_phase_sin','lunar_phase_cos','is_high_tide']].values
+ # 缁熻鐗瑰緛锛堥鍏堝埄鐢� rolling 宸茶绠楀ソ锛屾敞鎰忓彇鍑哄搴旇锛�
+ try:
+ stats_up = sample_df[['mean_1d_up','mean_3d_up','std_1d_up','max_1d_up','min_1d_up']].values
+ stats_down = sample_df[['mean_1d_down','mean_3d_down','std_1d_down','max_1d_down','min_1d_down']].values
+ except KeyError as e:
+ print(f"缁熻鐗瑰緛鍒椾笉瀛樺湪: {e}锛岃纭繚鍏堣绠楃粺璁＄壒寰�")
+ return np.array([]), np.array([])
+
+ # 寤惰繜鐗瑰緛锛氬亣璁炬墍鏈夊欢杩熺壒寰佸垪鍚嶅潎浠� "upstream_delay_" 鎴� "downstream_delay_" 寮�澶�
+ delay_cols = [col for col in sample_df.columns if col.startswith('upstream_delay_') or col.startswith('downstream_delay_')]
+ delay_feats = sample_df[delay_cols].values
+
+ # 鎷兼帴鎵�鏈夌壒寰侊細鍏堝皢鍘嗗彶绐楀彛锛坵indow_up锛変笌涓嬫父绐楀彛锛坵indow_down锛夋嫾鎺ワ紝鍐嶆嫾鎺ュ叾浠栫壒寰�
+ X = np.hstack([window_up, window_down, basic_time_feats, lunar_feats, stats_up, stats_down, delay_feats])
+
+ # 鏋勯�犳爣绛撅細鍒╃敤婊戝姩绐楀彛鏋勯�� forecast_horizon 鍐呯殑涓嬫父鏁版嵁
+ label_full = sliding_window_view(downstream_array, window_shape=forecast_horizon)
+ # 鏍囩鍖洪棿瀵瑰簲浠� index = look_back 鍒� look_back + total_samples
+ y = label_full[look_back: look_back + total_samples, :]
+ global feature_columns
+ feature_columns = ["combined_vector_features"]
+ print(f"鍚戦噺鍖栫壒寰佸伐绋嬪畬鎴愶紝鏈夋晥鏍锋湰鏁�: {X.shape[0]}")
+ return X, y
+
+# -------------------------------
+# 鑾峰彇妯″瀷鍑嗙‘搴︽寚鏍�
+# -------------------------------
+def get_model_metrics():
+ """鑾峰彇淇濆瓨鍦ㄦā鍨嬬紦瀛樹腑鐨勫噯纭害鎸囨爣"""
+ model_cache_file = 'salinity_model.pkl'
+ if os.path.exists(model_cache_file):
+ try:
+ with open(model_cache_file, 'rb') as f:
+ model_data = pickle.load(f)
+ return {
+ 'rmse': model_data.get('rmse', None),
+ 'mae': model_data.get('mae', None)
+ }
+ except Exception as e:
+ print(f"鑾峰彇妯″瀷鎸囨爣澶辫触: {e}")
+ return None
+
+# -------------------------------
+# 妯″瀷璁粌涓庨娴嬶紝灞曠ず楠岃瘉鍑嗙‘搴︼紙RMSE, MAE锛�
+# -------------------------------
+def train_and_predict(df, start_time, force_retrain=False):
+ global cached_model, last_training_time
+ model_cache_file = 'salinity_model.pkl'
+ model_needs_training = True
+
+ if os.path.exists(model_cache_file) and force_retrain:
+ try:
+ os.remove(model_cache_file)
+ print("宸插垹闄ゆ棫妯″瀷缂撳瓨锛堝己鍒堕噸鏂拌缁冿級")
+ except Exception as e:
+ print("鍒犻櫎缂撳瓨寮傚父:", e)
+
+ train_df = df[df['DateTime'] < start_time].copy()
+ if not force_retrain and cached_model is not None and last_training_time is not None:
+ if last_training_time >= train_df['DateTime'].max():
+ model_needs_training = False
+ print(f"浣跨敤缂撳瓨妯″瀷锛岃缁冩椂闂�: {last_training_time}")
+ elif not force_retrain and os.path.exists(model_cache_file):
+ try:
+ with open(model_cache_file, 'rb') as f:
+ model_data = pickle.load(f)
+ cached_model = model_data['model']
+ last_training_time = model_data['training_time']
+ if last_training_time >= train_df['DateTime'].max():
+ model_needs_training = False
+ print(f"浠庢枃浠跺姞杞芥ā鍨嬶紝璁粌鏃堕棿: {last_training_time}")
+ except Exception as e:
+ print("鍔犺浇妯″瀷澶辫触:", e)
+
+ if model_needs_training:
+ print("寮�濮嬭缁冩柊妯″瀷...")
+ if len(train_df) < 100:
+ print("璁粌鏁版嵁涓嶈冻")
+ return None, None, None, None
+
+ start_train = time()
+ X, y = create_features_vectorized(train_df, look_back=96, forecast_horizon=5)
+ if len(X) == 0 or len(y) == 0:
+ print("鏍锋湰鐢熸垚涓嶈冻锛岃缁冪粓姝�")
+ return None, None, None, None
+ print(f"璁粌鏍锋湰鏁伴噺: {X.shape[0]}")
+ X_train, X_val, y_train, y_val = train_test_split(X, y, test_size=0.2, random_state=42)
+ model = XGBRegressor(
+ n_estimators=300,
+ learning_rate=0.03,
+ max_depth=5,
+ min_child_weight=2,
+ subsample=0.85,
+ colsample_bytree=0.85,
+ gamma=0.1,
+ reg_alpha=0.2,
+ reg_lambda=1.5,
+ n_jobs=-1,
+ random_state=42
+ )
+ try:
+ model.fit(X_train, y_train,
+ eval_set=[(X_val, y_val)], eval_metric='rmse',
+ early_stopping_rounds=20, verbose=False)
+ # 鍦ㄩ獙璇侀泦涓婅绠� RMSE 鍜� MAE
+ y_val_pred = model.predict(X_val)
+ rmse = np.sqrt(mean_squared_error(y_val, y_val_pred))
+ mae = mean_absolute_error(y_val, y_val_pred)
+ print(f"楠岃瘉闆� RMSE: {rmse:.4f}, MAE: {mae:.4f}")
+ last_training_time = start_time
+ cached_model = model
+ with open(model_cache_file, 'wb') as f:
+ pickle.dump({
+ 'model': model,
+ 'training_time': last_training_time,
+ 'feature_columns': feature_columns,
+ 'rmse': rmse,
+ 'mae': mae
+ }, f)
+ print(f"妯″瀷璁粌瀹屾垚锛岃�楁椂: {time() - start_train:.2f}绉�")
+ except Exception as e:
+ print("妯″瀷璁粌寮傚父:", e)
+ return None, None, None, None
+ else:
+ model = cached_model
+
+ # 棰勬祴閮ㄥ垎锛氭瀯閫犲崟涓娴嬫牱鏈紙涓庤缁冩椂鐗瑰緛鏋勯�犱竴鑷达級
+ try:
+ # 杩欓噷閲囩敤涓� create_features_vectorized 绫讳技鐨勬�濊矾鏋勯�犻娴嬫牱鏈�
+ # 鍙栨渶杩戞暟鎹冻澶熸瀯鎴愬巻鍙茬獥鍙ｅ拰鍏朵粬鐗瑰緛
+ n = len(df)
+ if n < 96 + 5:
+ print("棰勬祴鏁版嵁涓嶈冻")
+ return None, None, None, None
+
+ # 浣跨敤 sliding_window_view 鏋勯�犳渶鏂扮殑涓婃父鍜屼笅娓哥獥鍙�
+ upstream_array = df['upstream'].values
+ window_up = np.lib.stride_tricks.sliding_window_view(upstream_array, window_shape=96)[-1, :]
+ downstream_array = df['downstream'].values
+ window_down = np.lib.stride_tricks.sliding_window_view(downstream_array, window_shape=24)[-1, :]
+
+ # 鏃堕棿鐗瑰緛鍜屽啘鍘嗙壒寰佸熀浜庡綋鍓嶉娴嬪紑濮嬫椂鍒�
+ hour_norm = start_time.hour / 24.0
+ weekday_norm = start_time.dayofweek / 7.0
+ month_norm = start_time.month / 12.0
+ basic_time_feats = np.array([hour_norm, weekday_norm, month_norm]).reshape(1, -1)
+ ld = LunarDate.fromSolarDate(start_time.year, start_time.month, start_time.day)
+ lunar_feats = np.array([np.sin(2*np.pi*ld.day/15),
+ np.cos(2*np.pi*ld.day/15),
+ 1 if (ld.day <=5 or (ld.day >=16 and ld.day<=20)) else 0]).reshape(1, -1)
+
+ # 缁熻鐗瑰緛锛氱敤鏈�鏂� 24/72 灏忔椂鏁版嵁锛堝彇鏈熬24/72锛�
+ try:
+ # 浼樺厛浣跨敤DataFrame涓凡璁＄畻鐨勭粺璁＄壒寰�
+ stats_up = df[['mean_1d_up','mean_3d_up','std_1d_up','max_1d_up','min_1d_up']].iloc[-1:].values
+ stats_down = df[['mean_1d_down','mean_3d_down','std_1d_down','max_1d_down','min_1d_down']].iloc[-1:].values
+ except KeyError:
+ # 濡傛灉涓嶅瓨鍦紝鍒欑洿鎺ヨ绠�
+ recent_up = df['upstream'].values[-24:]
+ stats_up = np.array([np.mean(recent_up),
+ np.mean(df['upstream'].values[-72:]),
+ np.std(recent_up),
+ np.max(recent_up),
+ np.min(recent_up)]).reshape(1, -1)
+ recent_down = df['downstream'].values[-24:]
+ stats_down = np.array([np.mean(recent_down),
+ np.mean(df['downstream'].values[-72:]),
+ np.std(recent_down),
+ np.max(recent_down),
+ np.min(recent_down)]).reshape(1, -1)
+
+ # 寤惰繜鐗瑰緛锛氱洿鎺ヤ粠鏈�鍚庝竴琛屽欢杩熺壒寰佸彇鍊�
+ delay_cols = [col for col in df.columns if col.startswith('upstream_delay_') or col.startswith('downstream_delay_')]
+ delay_feats = df[delay_cols].iloc[-1:].values # shape (1, ?)
+
+ # 鎷兼帴鎵�鏈夐娴嬬壒寰�
+ X_pred = np.hstack([window_up.reshape(1, -1),
+ window_down.reshape(1, -1),
+ basic_time_feats, lunar_feats, stats_up, stats_down, delay_feats])
+ if np.isnan(X_pred).any() or np.isinf(X_pred).any():
+ X_pred = np.nan_to_num(X_pred, nan=0.0, posinf=1e6, neginf=-1e6)
+ predictions = model.predict(X_pred)
+ # 鐢熸垚鏈潵鏃ユ湡鏍囩锛堥娴嬫湭鏉� 5 澶╋級
+ future_dates = [start_time + timedelta(days=i) for i in range(5)]
+ print("棰勬祴瀹屾垚")
+
+ # 鑾峰彇妯″瀷鎸囨爣
+ metrics = None
+ if os.path.exists(model_cache_file):
+ try:
+ with open(model_cache_file, 'rb') as f:
+ model_data = pickle.load(f)
+ metrics = {
+ 'rmse': model_data.get('rmse', None),
+ 'mae': model_data.get('mae', None)
+ }
+ except Exception as e:
+ print(f"鑾峰彇妯″瀷鎸囨爣澶辫触: {e}")
+
+ return future_dates, predictions.flatten(), model, metrics
+ except Exception as e:
+ print("棰勬祴杩囩▼寮傚父:", e)
+ return None, None, None, None
+
+# -------------------------------
+# GUI鐣岄潰閮ㄥ垎
+# -------------------------------
+def run_gui():
+ def configure_gui_fonts():
+ font_names = ['寰蒋闆呴粦', 'Microsoft YaHei', 'SimSun', 'SimHei']
+ for font_name in font_names:
+ try:
+ default_font = tkfont.nametofont("TkDefaultFont")
+ default_font.configure(family=font_name)
+ text_font = tkfont.nametofont("TkTextFont")
+ text_font.configure(family=font_name)
+ fixed_font = tkfont.nametofont("TkFixedFont")
+ fixed_font.configure(family=font_name)
+ return True
+ except Exception as e:
+ continue
+ return False
+
+ def on_predict():
+ try:
+ predict_start = time()
+ status_label.config(text="棰勬祴涓�...")
+ root.update()
+ start_time_dt = pd.to_datetime(entry.get())
+ force_retrain = retrain_var.get()
+ future_dates, predictions, model, metrics = train_and_predict(df, start_time_dt, force_retrain)
+ if future_dates is None or predictions is None:
+ status_label.config(text="棰勬祴澶辫触")
+ return
+
+ # 鑾峰彇骞舵樉绀烘ā鍨嬪噯纭害鎸囨爣
+ if metrics:
+ metrics_text = f"妯″瀷鍑嗙‘搴� - RMSE: {metrics['rmse']:.4f}, MAE: {metrics['mae']:.4f}"
+ metrics_label.config(text=metrics_text)
+
+ ax.clear()
+ # 缁樺埗鍘嗗彶鏁版嵁锛堟渶杩� 120 澶╋級
+ history_end = min(start_time_dt, df['DateTime'].max())
+ history_start = history_end - timedelta(days=120)
+ hist_data = df[(df['DateTime'] >= history_start) & (df['DateTime'] <= history_end)]
+ ax.plot(hist_data['DateTime'], hist_data['downstream'], label='涓�鍙栨按(涓嬫父)鐩愬害', color='blue', linewidth=1.5)
+ ax.plot(hist_data['DateTime'], hist_data['upstream_smooth'], label='闈掗緳娓�(涓婃父)鐩愬害', color='purple', linewidth=1.5, alpha=0.7)
+ if 'qinglong_lake_smooth' in hist_data.columns:
+ ax.plot(hist_data['DateTime'], hist_data['qinglong_lake_smooth'], label='闈掗緳婀栫洂搴�', color='green', linewidth=1.5, alpha=0.7)
+ ax.plot(future_dates, predictions, marker='o', linestyle='--', label='棰勬祴鐩愬害', color='red', linewidth=2)
+ actual_data = df[(df['DateTime'] >= start_time_dt) & (df['DateTime'] <= future_dates[-1])]
+ if not actual_data.empty:
+ ax.plot(actual_data['DateTime'], actual_data['downstream'], marker='s', linestyle='-', label='瀹為檯鐩愬害', color='orange', linewidth=2)
+ std_dev = hist_data['downstream'].std() * 0.5
+ ax.fill_between(future_dates, predictions - std_dev, predictions + std_dev, color='red', alpha=0.2)
+ ax.set_xlabel('鏃ユ湡')
+ ax.set_ylabel('鐩愬害')
+ ax.set_title(f"浠� {start_time_dt.strftime('%Y-%m-%d %H:%M:%S')} 寮�濮嬬殑鐩愬害棰勬祴")
+ ax.legend(loc='upper left')
+ fig.tight_layout()
+ canvas.draw()
+ predict_time = time() - predict_start
+ status_label.config(text=f"棰勬祴瀹屾垚 (鑰楁椂: {predict_time:.2f}绉�)")
+ result_text = "棰勬祴缁撴灉:\n"
+ for i, (date, pred) in enumerate(zip(future_dates, predictions)):
+ result_text += f"绗� {i+1} 澶� ({date.strftime('%Y-%m-%d')}): {pred:.2f}\n"
+ result_label.config(text=result_text)
+ except Exception as e:
+ status_label.config(text=f"閿欒: {str(e)}")
+
+ def on_scroll(event):
+ xlim = ax.get_xlim()
+ ylim = ax.get_ylim()
+ zoom_factor = 1.1
+ x_data = event.xdata if event.xdata is not None else (xlim[0]+xlim[1])/2
+ y_data = event.ydata if event.ydata is not None else (ylim[0]+ylim[1])/2
+ x_rel = (x_data - xlim[0]) / (xlim[1] - xlim[0])
+ y_rel = (y_data - ylim[0]) / (ylim[1] - ylim[0])
+ if event.step > 0:
+ new_width = (xlim[1]-xlim[0]) / zoom_factor
+ new_height = (ylim[1]-ylim[0]) / zoom_factor
+ x0 = x_data - x_rel * new_width
+ y0 = y_data - y_rel * new_height
+ ax.set_xlim([x0, x0+new_width])
+ ax.set_ylim([y0, y0+new_height])
+ else:
+ new_width = (xlim[1]-xlim[0]) * zoom_factor
+ new_height = (ylim[1]-ylim[0]) * zoom_factor
+ x0 = x_data - x_rel * new_width
+ y0 = y_data - y_rel * new_height
+ ax.set_xlim([x0, x0+new_width])
+ ax.set_ylim([y0, y0+new_height])
+ canvas.draw_idle()
+
+ def update_cursor(event):
+ if event.inaxes == ax:
+ canvas.get_tk_widget().config(cursor="fleur")
+ else:
+ canvas.get_tk_widget().config(cursor="")
+
+ def reset_view():
+ display_history()
+ status_label.config(text="鍥捐〃瑙嗗浘宸查噸缃�")
+
+ root = tk.Tk()
+ root.title("闈掗緳娓�-闄堣鐩愬害棰勬祴绯荤粺")
+ try:
+ configure_gui_fonts()
+ except Exception as e:
+ print("瀛椾綋閰嶇疆寮傚父:", e)
+ input_frame = ttk.Frame(root, padding="10")
+ input_frame.pack(fill=tk.X)
+ control_frame = ttk.Frame(root, padding="5")
+ control_frame.pack(fill=tk.X)
+ result_frame = ttk.Frame(root, padding="10")
+ result_frame.pack(fill=tk.BOTH, expand=True)
+ ttk.Label(input_frame, text="杈撳叆寮�濮嬫椂闂� (YYYY-MM-DD HH:MM:SS)").pack(side=tk.LEFT)
+ entry = ttk.Entry(input_frame, width=25)
+ entry.pack(side=tk.LEFT, padx=5)
+ predict_button = ttk.Button(input_frame, text="棰勬祴", command=on_predict)
+ predict_button.pack(side=tk.LEFT)
+ status_label = ttk.Label(input_frame, text="鎻愮ず: 绗竴娆¤繍琛岃鍕鹃��'寮哄埗閲嶆柊璁粌妯″瀷'")
+ status_label.pack(side=tk.LEFT, padx=10)
+ retrain_var = tk.BooleanVar(value=False)
+ ttk.Checkbutton(control_frame, text="寮哄埗閲嶆柊璁粌妯″瀷", variable=retrain_var).pack(side=tk.LEFT)
+ legend_label = ttk.Label(control_frame, text="鍥句緥: 绱壊=闈掗緳娓笂娓告暟鎹�, 钃濊壊=涓�鍙栨按涓嬫父鏁版嵁, 绾㈣壊=棰勬祴鍊�, 缁胯壊=瀹為檯鍊�")
+ legend_label.pack(side=tk.LEFT, padx=10)
+ reset_button = ttk.Button(control_frame, text="閲嶇疆瑙嗗浘", command=reset_view)
+ reset_button.pack(side=tk.LEFT, padx=5)
+
+ # 娣诲姞鏄剧ず妯″瀷鍑嗙‘搴︾殑鏍囩
+ metrics_frame = ttk.Frame(root, padding="5")
+ metrics_frame.pack(fill=tk.X)
+ model_metrics = get_model_metrics()
+ metrics_text = "妯″瀷鍑嗙‘搴�: 鏈煡" if not model_metrics else f"妯″瀷鍑嗙‘搴� - RMSE: {model_metrics['rmse']:.4f}, MAE: {model_metrics['mae']:.4f}"
+ metrics_label = ttk.Label(metrics_frame, text=metrics_text)
+ metrics_label.pack(side=tk.LEFT, padx=10)
+
+ result_label = ttk.Label(result_frame, text="", justify=tk.LEFT)
+ result_label.pack(side=tk.RIGHT, fill=tk.Y)
+ fig, ax = plt.subplots(figsize=(10,5), dpi=100)
+ canvas = FigureCanvasTkAgg(fig, master=result_frame)
+ canvas.get_tk_widget().pack(side=tk.LEFT, fill=tk.BOTH, expand=True)
+ toolbar_frame = ttk.Frame(result_frame)
+ toolbar_frame.pack(side=tk.BOTTOM, fill=tk.X)
+ toolbar = NavigationToolbar2Tk(canvas, toolbar_frame)
+ toolbar.update()
+ canvas.mpl_connect('scroll_event', on_scroll)
+ canvas.mpl_connect('motion_notify_event', update_cursor)
+
+ def display_history():
+ ax.clear()
+ end_date = df['DateTime'].max()
+ start_date = max(df['DateTime'].min(), end_date - timedelta(days=60))
+ hist_data = df[(df['DateTime']>=start_date)&(df['DateTime']<=end_date)]
+ ax.plot(hist_data['DateTime'], hist_data['downstream'], label='涓�鍙栨按(涓嬫父)鐩愬害', color='blue', linewidth=1.5)
+ ax.plot(hist_data['DateTime'], hist_data['upstream_smooth'], label='闈掗緳娓�(涓婃父)鐩愬害', color='purple', linewidth=1.5, alpha=0.7)
+ ax.set_xlabel('鏃ユ湡')
+ ax.set_ylabel('鐩愬害')
+ ax.set_title('鍘嗗彶鐩愬害鏁版嵁瀵规瘮')
+ ax.legend()
+ fig.tight_layout()
+ canvas.draw()
+
+ display_history()
+ root.mainloop()
+
+# -------------------------------
+# 涓荤▼搴忓叆鍙ｏ細鍔犺浇鏁版嵁銆佹坊鍔犵壒寰併�佺敓鎴愬欢杩熺壒寰佸悗鍚姩GUI
+# -------------------------------
+def save_processed_data(df, filename='processed_data.pkl'):
+ try:
+ df.to_pickle(filename)
+ print(f"宸蹭繚瀛樺鐞嗗悗鐨勬暟鎹埌 {filename}")
+ return True
+ except Exception as e:
+ print(f"淇濆瓨鏁版嵁澶辫触: {e}")
+ return False
+
+def load_processed_data(filename='processed_data.pkl'):
+ try:
+ if os.path.exists(filename):
+ df = pd.read_pickle(filename)
+ print(f"宸蹭粠 {filename} 鍔犺浇澶勭悊鍚庣殑鏁版嵁")
+ return df
+ else:
+ print(f"鎵句笉鍒板鐞嗗悗鐨勬暟鎹枃浠� {filename}")
+ return None
+ except Exception as e:
+ print(f"鍔犺浇鏁版嵁澶辫触: {e}")
+ return None
+
+# 灏濊瘯鍔犺浇澶勭悊鍚庣殑鏁版嵁锛屽鏋滀笉瀛樺湪鍒欓噸鏂板鐞�
+processed_data = load_processed_data()
+if processed_data is not None:
+ df = processed_data
+else:
+ df = load_data('闈掗緳娓�1.csv', '涓�鍙栨按.csv')
+ if df is not None:
+ df = add_lunar_features(df)
+ delay_hours = [1,2,3,4,6,12,24,36,48,60,72,84,96,108,120]
+ df = batch_create_delay_features(df, delay_hours)
+
+ # 娣诲姞缁熻鐗瑰緛
+ df['mean_1d_up'] = df['upstream'].rolling(window=24, min_periods=1).mean()
+ df['mean_3d_up'] = df['upstream'].rolling(window=72, min_periods=1).mean()
+ df['std_1d_up'] = df['upstream'].rolling(window=24, min_periods=1).std()
+ df['max_1d_up'] = df['upstream'].rolling(window=24, min_periods=1).max()
+ df['min_1d_up'] = df['upstream'].rolling(window=24, min_periods=1).min()
+
+ df['mean_1d_down'] = df['downstream'].rolling(window=24, min_periods=1).mean()
+ df['mean_3d_down'] = df['downstream'].rolling(window=72, min_periods=1).mean()
+ df['std_1d_down'] = df['downstream'].rolling(window=24, min_periods=1).std()
+ df['max_1d_down'] = df['downstream'].rolling(window=24, min_periods=1).max()
+ df['min_1d_down'] = df['downstream'].rolling(window=24, min_periods=1).min()
+
+ # 淇濆瓨澶勭悊鍚庣殑鏁版嵁
+ save_processed_data(df)
+
+if df is not None:
+ run_gui()
+else:
+ print("鏁版嵁鍔犺浇澶辫触锛屾棤娉曡繍琛岄娴嬨��")
--
Gitblit v1.9.3