vf/vf_react
4
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

update app_vx

Browse Source
This commit is contained in:
zzlgreat
2025-11-13 07:40:46 +08:00
parent 926ffa1b8f
commit d64349b606
7 changed files with 13787 additions and 112 deletions
Download Patch File Download Diff File Expand all files Collapse all files

516
app_vx.py
View File

@@ -15,6 +15,7 @@ from pathlib import Path
import time
from sqlalchemy import create_engine, text, func, or_, case, event, desc, asc
from flask import Flask, has_request_context, render_template, request, jsonify, redirect, url_for, flash, session, render_template_string, current_app, send_from_directory
from flask_sqlalchemy import SQLAlchemy
from flask_login import LoginManager, UserMixin, login_user, logout_user, login_required, current_user
from flask_mail import Mail, Message
@@ -325,7 +326,7 @@ def subscription_required(level='pro'):
@subscription_required('pro') # 需要 Pro 或 Max 用户
@subscription_required('max') # 仅限 Max 用户
注意：此装饰器需要配合 @token_required 使用
注意：此装饰器需要配合 使用
"""
from functools import wraps
@@ -1052,8 +1053,6 @@ def get_clickhouse_client():
@app.route('/api/stock/<stock_code>/kline')
@token_required
@pro_or_max_required
def get_stock_kline(stock_code):
"""获取股票K线数据 - 仅限 Pro/Max 会员（小程序功能）"""
chart_type = request.args.get('chart_type', 'daily') # 默认改为daily
@@ -1628,7 +1627,7 @@ def add_comment(post_id):
@app.route('/post/comments/<int:post_id>')
@token_required
def get_comments(post_id):
"""获取帖子评论列表"""
page = request.args.get('page', 1, type=int)
@@ -2012,8 +2011,8 @@ def get_limit_rate(stock_code):
@app.route('/api/events', methods=['GET'])
@token_required
@pro_or_max_required
def api_get_events():
"""
获取事件列表API - 优化版本（保持完全兼容）
@@ -2555,11 +2554,7 @@ def api_get_events():
'week_change': week_change
}
# ==================== 获取整体统计信息 ====================
# 获取所有筛选条件下的事件和股票（用于统计）
all_filtered_events = query.limit(500).all()
all_event_ids = [e.id for e in all_filtered_events]
# ==================== 获取整体统计信息（应用所有筛选条件） ====================
overall_distribution = {
'limit_down': 0,
@@ -2573,40 +2568,105 @@ def api_get_events():
'limit_up': 0
}
if all_event_ids:
# 获取所有相关股票
all_stocks_for_stats = RelatedStock.query.filter(
RelatedStock.event_id.in_(all_event_ids)
# 使用当前筛选条件的query，但不应用分页限制，获取所有符合条件的事件
# 这样统计数据会跟随用户的筛选条件变化
all_filtered_events = query.limit(1000).all() # 限制最多1000个事件，避免查询过慢
week_event_ids = [e.id for e in all_filtered_events]
if week_event_ids:
# 获取这些事件的所有关联股票
week_related_stocks = RelatedStock.query.filter(
RelatedStock.event_id.in_(week_event_ids)
).all()
# 统计涨跌分布
for stock in all_stocks_for_stats:
clean_code = stock.stock_code.replace('.SH', '').replace('.SZ', '').replace('.BJ', '')
if clean_code in stock_changes:
daily_change = stock_changes[clean_code]['daily_change']
# 按事件ID分组
week_stocks_by_event = {}
for stock in week_related_stocks:
if stock.event_id not in week_stocks_by_event:
week_stocks_by_event[stock.event_id] = []
week_stocks_by_event[stock.event_id].append(stock)
# 计算涨跌停限制
limit_rate = get_limit_rate(stock.stock_code)
# 收集所有股票代码（用于批量查询行情）
week_stock_codes = []
week_code_mapping = {}
for stocks in week_stocks_by_event.values():
for stock in stocks:
clean_code = stock.stock_code.replace('.SH', '').replace('.SZ', '').replace('.BJ', '')
week_stock_codes.append(clean_code)
week_code_mapping[clean_code] = stock.stock_code
# 分类统计
if daily_change <= -limit_rate + 0.01:
overall_distribution['limit_down'] += 1
elif daily_change >= limit_rate - 0.01:
overall_distribution['limit_up'] += 1
elif daily_change > 5:
overall_distribution['up_over_5'] += 1
elif daily_change > 1:
overall_distribution['up_1_to_5'] += 1
elif daily_change > 0.1:
overall_distribution['up_within_1'] += 1
elif daily_change >= -0.1:
overall_distribution['flat'] += 1
elif daily_change > -1:
overall_distribution['down_within_1'] += 1
elif daily_change > -5:
overall_distribution['down_5_to_1'] += 1
else:
overall_distribution['down_over_5'] += 1
week_stock_codes = list(set(week_stock_codes))
# 批量查询这些股票的最新行情数据
week_stock_changes = {}
if week_stock_codes:
codes_str = "'" + "', '".join(week_stock_codes) + "'"
recent_trades_sql = f"""
SELECT
SECCODE,
SECNAME,
F010N as daily_change,
ROW_NUMBER() OVER (PARTITION BY SECCODE ORDER BY TRADEDATE DESC) as rn
FROM ea_trade
WHERE SECCODE IN ({codes_str})
AND F010N IS NOT NULL
AND TRADEDATE >= DATE_SUB(CURDATE(), INTERVAL 3 DAY)
ORDER BY SECCODE, TRADEDATE DESC
"""
result = db.session.execute(text(recent_trades_sql))
for row in result.fetchall():
sec_code = row[0]
if row[3] == 1: # 只取最新的数据（rn=1）
week_stock_changes[sec_code] = {
'stock_name': row[1],
'daily_change': float(row[2]) if row[2] else 0
}
# 按事件统计平均涨跌幅分布
event_avg_changes = {}
for event in all_filtered_events:
event_stocks = week_stocks_by_event.get(event.id, [])
if not event_stocks:
continue
total_change = 0
valid_count = 0
for stock in event_stocks:
clean_code = stock.stock_code.replace('.SH', '').replace('.SZ', '').replace('.BJ', '')
if clean_code in week_stock_changes:
daily_change = week_stock_changes[clean_code]['daily_change']
total_change += daily_change
valid_count += 1
if valid_count > 0:
avg_change = total_change / valid_count
event_avg_changes[event.id] = avg_change
# 统计事件平均涨跌幅的分布
for event_id, avg_change in event_avg_changes.items():
# 对于事件平均涨幅，不使用涨跌停分类，使用通用分类
if avg_change <= -10:
overall_distribution['limit_down'] += 1
elif avg_change >= 10:
overall_distribution['limit_up'] += 1
elif avg_change > 5:
overall_distribution['up_over_5'] += 1
elif avg_change > 1:
overall_distribution['up_1_to_5'] += 1
elif avg_change > 0.1:
overall_distribution['up_within_1'] += 1
elif avg_change >= -0.1:
overall_distribution['flat'] += 1
elif avg_change > -1:
overall_distribution['down_within_1'] += 1
elif avg_change > -5:
overall_distribution['down_5_to_1'] += 1
else:
overall_distribution['down_over_5'] += 1
# ==================== 构建响应数据 ====================
@@ -2839,8 +2899,8 @@ def get_event_class(count):
else:
return 'bg-gradient-success'
@app.route('/api/calendar-event-counts')
@token_required
@pro_or_max_required
def get_calendar_event_counts():
"""获取整月的事件数量统计，仅统计type为event的事件"""
try:
@@ -2930,8 +2990,8 @@ def to_dict(self):
# 1. 首页接口
@app.route('/api/home', methods=['GET'])
@token_required
@pro_or_max_required
def api_home():
try:
seven_days_ago = datetime.now() - timedelta(days=7)
@@ -3620,17 +3680,107 @@ def api_login_email():
# 5. 事件详情-相关标的接口
@app.route('/api/event/<int:event_id>/related-stocks-detail', methods=['GET'])
@token_required
@pro_or_max_required
def api_event_related_stocks(event_id):
"""事件相关标的详情接口 - 仅限 Pro/Max 会员"""
try:
from datetime import datetime, timedelta, time as dt_time
from sqlalchemy import text
event = Event.query.get_or_404(event_id)
related_stocks = event.related_stocks.order_by(RelatedStock.correlation.desc()).all()
# 获取ClickHouse客户端用于分时数据查询
client = get_clickhouse_client()
# 获取事件时间（如果事件有开始时间，使用开始时间；否则使用创建时间）
event_time = event.start_time if event.start_time else event.created_at
current_time = datetime.now()
# 定义交易日和时间范围计算函数（与 app.py 中的逻辑完全一致）
def get_trading_day_and_times(event_datetime):
event_date = event_datetime.date()
event_time_only = event_datetime.time()
# Trading hours
market_open = dt_time(9, 30)
market_close = dt_time(15, 0)
with engine.connect() as conn:
# First check if the event date itself is a trading day
is_trading_day = conn.execute(text("""
SELECT 1
FROM trading_days
WHERE EXCHANGE_DATE = :date
"""), {"date": event_date}).fetchone() is not None
if is_trading_day:
# If it's a trading day, determine time period based on event time
if event_time_only < market_open:
# Before market opens - use full trading day
return event_date, market_open, market_close
elif event_time_only > market_close:
# After market closes - get next trading day
next_trading_day = conn.execute(text("""
SELECT EXCHANGE_DATE
FROM trading_days
WHERE EXCHANGE_DATE > :date
ORDER BY EXCHANGE_DATE LIMIT 1
"""), {"date": event_date}).fetchone()
# Convert to date object if we found a next trading day
return (next_trading_day[0].date() if next_trading_day else None,
market_open, market_close)
else:
# During trading hours
return event_date, event_time_only, market_close
else:
# If not a trading day, get next trading day
next_trading_day = conn.execute(text("""
SELECT EXCHANGE_DATE
FROM trading_days
WHERE EXCHANGE_DATE > :date
ORDER BY EXCHANGE_DATE LIMIT 1
"""), {"date": event_date}).fetchone()
# Convert to date object if we found a next trading day
return (next_trading_day[0].date() if next_trading_day else None,
market_open, market_close)
trading_day, start_time, end_time = get_trading_day_and_times(event_time)
if not trading_day:
# 如果没有交易日，返回空数据
return jsonify({
'code': 200,
'message': 'success',
'data': {
'event_id': event_id,
'event_title': event.title,
'event_desc': event.description,
'event_type': event.event_type,
'event_importance': event.importance,
'event_status': event.status,
'event_created_at': event.created_at.strftime("%Y-%m-%d %H:%M:%S"),
'event_start_time': event.start_time.isoformat() if event.start_time else None,
'event_end_time': event.end_time.isoformat() if event.end_time else None,
'keywords': event.keywords,
'view_count': event.view_count,
'post_count': event.post_count,
'follower_count': event.follower_count,
'related_stocks': [],
'total_count': 0
}
})
# For historical dates, ensure we're using actual data
start_datetime = datetime.combine(trading_day, start_time)
end_datetime = datetime.combine(trading_day, end_time)
# If the trading day is in the future relative to current time, return only names without data
if trading_day > current_time.date():
start_datetime = datetime.combine(trading_day, start_time)
end_datetime = datetime.combine(trading_day, end_time)
print(f"事件时间: {event_time}, 交易日: {trading_day}, 时间范围: {start_datetime} - {end_datetime}")
def get_minute_chart_data(stock_code):
"""获取股票分时图数据"""
try:
@@ -3705,39 +3855,126 @@ def api_event_related_stocks(event_id):
stocks_data = []
for stock in related_stocks:
# 获取股票基本信息 - 也使用灵活匹配
print(f"正在处理股票 {stock.stock_code} 的价格数据...")
# 获取股票基本信息
stock_info = StockBasicInfo.query.filter_by(SECCODE=stock.stock_code).first()
if not stock_info:
base_code = stock.stock_code.split('.')[0]
stock_info = StockBasicInfo.query.filter_by(SECCODE=base_code).first()
# 获取最新交易数据 - 使用灵活匹配
latest_trade = None
search_codes = [stock.stock_code, stock.stock_code.split('.')[0]]
for code in search_codes:
latest_trade = TradeData.query.filter_by(SECCODE=code) \
.order_by(TradeData.TRADEDATE.desc()).first()
if latest_trade:
break
# 获取前一交易日数据
prev_trade = None
if latest_trade:
prev_trade = TradeData.query.filter_by(SECCODE=latest_trade.SECCODE) \
.filter(TradeData.TRADEDATE < latest_trade.TRADEDATE) \
.order_by(TradeData.TRADEDATE.desc()).first()
# 计算涨跌幅
# 使用与 get_stock_quotes 完全相同的逻辑计算涨跌幅
latest_price = None
first_price = None
change_pct = None
change_amount = None
if latest_trade and prev_trade:
if prev_trade.F007N and prev_trade.F007N != 0:
change_amount = float(latest_trade.F007N) - float(prev_trade.F007N)
change_pct = (change_amount / float(prev_trade.F007N)) * 100
elif latest_trade and latest_trade.F010N:
change_pct = float(latest_trade.F010N)
change_amount = float(latest_trade.F009N) if latest_trade.F009N else None
open_price = None
high_price = None
low_price = None
volume = None
amount = None
trade_date = trading_day
try:
# 使用与 get_stock_quotes 完全相同的 SQL 查询
# 获取事件时间点的第一个价格 (first_price) 和当前时间的最后一个价格 (last_price)
data = client.execute("""
WITH first_price AS (
SELECT close
FROM stock_minute
WHERE code = %(code)s
AND timestamp >= %(start)s
AND timestamp <= %(end)s
ORDER BY timestamp
LIMIT 1
),
last_price AS (
SELECT close
FROM stock_minute
WHERE code = %(code)s
AND timestamp >= %(start)s
AND timestamp <= %(end)s
ORDER BY timestamp DESC
LIMIT 1
)
SELECT
last_price.close as last_price,
(last_price.close - first_price.close) / first_price.close * 100 as change,
first_price.close as first_price
FROM last_price
CROSS JOIN first_price
WHERE EXISTS (SELECT 1 FROM first_price)
AND EXISTS (SELECT 1 FROM last_price)
""", {
'code': stock.stock_code,
'start': start_datetime,
'end': end_datetime
})
print(f"股票 {stock.stock_code} 查询结果: {data}")
if data and data[0] and data[0][0] is not None:
latest_price = float(data[0][0])
change_pct = float(data[0][1]) if data[0][1] is not None else None
first_price = float(data[0][2]) if len(data[0]) > 2 and data[0][2] is not None else None
# 计算涨跌额
if latest_price is not None and first_price is not None:
change_amount = latest_price - first_price
# 获取额外的价格信息（开盘价、最高价、最低价等）
extra_data = client.execute("""
SELECT
open, high, low, volume, amt
FROM stock_minute
WHERE code = %(code)s
AND timestamp >= %(start)s
AND timestamp <= %(end)s
ORDER BY timestamp DESC
LIMIT 1
""", {
'code': stock.stock_code,
'start': start_datetime,
'end': end_datetime
})
if extra_data and extra_data[0]:
open_price = float(extra_data[0][0]) if extra_data[0][0] else None
high_price = float(extra_data[0][1]) if extra_data[0][1] else None
low_price = float(extra_data[0][2]) if extra_data[0][2] else None
volume = int(extra_data[0][3]) if extra_data[0][3] else None
amount = float(extra_data[0][4]) if extra_data[0][4] else None
except Exception as e:
print(f"Error fetching price data for {stock.stock_code}: {e}")
# 如果 ClickHouse 查询失败，尝试使用 TradeData 作为降级方案
try:
latest_trade = None
search_codes = [stock.stock_code, stock.stock_code.split('.')[0]]
for code in search_codes:
latest_trade = TradeData.query.filter_by(SECCODE=code) \
.order_by(TradeData.TRADEDATE.desc()).first()
if latest_trade:
break
if latest_trade:
latest_price = float(latest_trade.F007N) if latest_trade.F007N else None
open_price = float(latest_trade.F003N) if latest_trade.F003N else None
high_price = float(latest_trade.F005N) if latest_trade.F005N else None
low_price = float(latest_trade.F006N) if latest_trade.F006N else None
first_price = float(latest_trade.F002N) if latest_trade.F002N else None
volume = float(latest_trade.F004N) if latest_trade.F004N else None
amount = float(latest_trade.F011N) if latest_trade.F011N else None
trade_date = latest_trade.TRADEDATE
# 计算涨跌幅
if latest_trade.F010N:
change_pct = float(latest_trade.F010N)
if latest_trade.F009N:
change_amount = float(latest_trade.F009N)
except Exception as fallback_error:
print(f"Fallback query also failed for {stock.stock_code}: {fallback_error}")
# 获取分时图数据
minute_chart_data = get_minute_chart_data(stock.stock_code)
@@ -3755,17 +3992,19 @@ def api_event_related_stocks(event_id):
# 交易数据
'trade_data': {
'latest_price': float(latest_trade.F007N) if latest_trade and latest_trade.F007N else None,
'open_price': float(latest_trade.F003N) if latest_trade and latest_trade.F003N else None,
'high_price': float(latest_trade.F005N) if latest_trade and latest_trade.F005N else None,
'low_price': float(latest_trade.F006N) if latest_trade and latest_trade.F006N else None,
'prev_close': float(latest_trade.F002N) if latest_trade and latest_trade.F002N else None,
'change_amount': change_amount,
'latest_price': latest_price,
'first_price': first_price, # 事件发生时的价格
'open_price': open_price,
'high_price': high_price,
'low_price': low_price,
'change_amount': round(change_amount, 2) if change_amount is not None else None,
'change_pct': round(change_pct, 2) if change_pct is not None else None,
'volume': float(latest_trade.F004N) if latest_trade and latest_trade.F004N else None,
'amount': float(latest_trade.F011N) if latest_trade and latest_trade.F011N else None,
'trade_date': latest_trade.TRADEDATE.isoformat() if latest_trade else None,
} if latest_trade else None,
'volume': volume,
'amount': amount,
'trade_date': trade_date.isoformat() if trade_date else None,
'event_start_time': start_datetime.isoformat() if start_datetime else None, # 事件开始时间
'event_end_time': end_datetime.isoformat() if end_datetime else None, # 查询结束时间
} if latest_price is not None else None,
# 分时图数据
'minute_chart_data': minute_chart_data,
@@ -3809,8 +4048,8 @@ def api_event_related_stocks(event_id):
@app.route('/api/stock/<stock_code>/minute-chart', methods=['GET'])
@token_required
@pro_or_max_required
def get_minute_chart_data(stock_code):
"""获取股票分时图数据 - 仅限 Pro/Max 会员"""
client = get_clickhouse_client()
@@ -3885,8 +4124,8 @@ def get_minute_chart_data(stock_code):
@app.route('/api/event/<int:event_id>/stock/<stock_code>/detail', methods=['GET'])
@token_required
@pro_or_max_required
def api_stock_detail(event_id, stock_code):
"""个股详情接口 - 仅限 Pro/Max 会员"""
try:
@@ -4146,8 +4385,8 @@ def get_stock_minute_chart_data(stock_code):
# 7. 事件详情-相关概念接口
@app.route('/api/event/<int:event_id>/related-concepts', methods=['GET'])
@token_required
@pro_or_max_required
def api_event_related_concepts(event_id):
"""事件相关概念接口"""
try:
@@ -4189,8 +4428,8 @@ def api_event_related_concepts(event_id):
# 8. 事件详情-历史事件接口
@app.route('/api/event/<int:event_id>/historical-events', methods=['GET'])
@token_required
@pro_or_max_required
def api_event_historical_events(event_id):
"""事件历史事件接口"""
try:
@@ -4290,8 +4529,8 @@ def api_event_historical_events(event_id):
@app.route('/api/event/<int:event_id>/comments', methods=['GET'])
@token_required
@pro_or_max_required
def get_event_comments(event_id):
"""获取事件的所有评论和帖子（嵌套格式）
@@ -4545,8 +4784,8 @@ def get_event_comments(event_id):
@app.route('/api/comment/<int:comment_id>/replies', methods=['GET'])
@token_required
@pro_or_max_required
def get_comment_replies(comment_id):
"""获取某条评论的所有回复
@@ -4689,10 +4928,64 @@ def get_comment_replies(comment_id):
}), 500
# 工具函数：清理 Markdown 文本
def clean_markdown_text(text):
"""清理文本中的 Markdown 符号和多余的换行符
Args:
text: 原始文本（可能包含 Markdown 符号）
Returns:
清理后的纯文本
"""
if not text:
return text
import re
# 1. 移除 Markdown 标题符号 (### , ## , # )
text = re.sub(r'^#{1,6}\s+', '', text, flags=re.MULTILINE)
# 2. 移除 Markdown 加粗符号 (**text** 或 __text__)
text = re.sub(r'\*\*(.+?)\*\*', r'\1', text)
text = re.sub(r'__(.+?)__', r'\1', text)
# 3. 移除 Markdown 斜体符号 (*text* 或 _text_)
text = re.sub(r'\*(.+?)\*', r'\1', text)
text = re.sub(r'_(.+?)_', r'\1', text)
# 4. 移除 Markdown 列表符号 (- , * , + , 1. )
text = re.sub(r'^[\s]*[-*+]\s+', '', text, flags=re.MULTILINE)
text = re.sub(r'^[\s]*\d+\.\s+', '', text, flags=re.MULTILINE)
# 5. 移除 Markdown 引用符号 (> )
text = re.sub(r'^>\s+', '', text, flags=re.MULTILINE)
# 6. 移除 Markdown 代码块符号 (``` 或 `)
text = re.sub(r'```[\s\S]*?```', '', text)
text = re.sub(r'`(.+?)`', r'\1', text)
# 7. 移除 Markdown 链接 ([text](url) -> text)
text = re.sub(r'\[(.+?)\]\(.+?\)', r'\1', text)
# 8. 清理多余的换行符
# 将多个连续的换行符（\n\n\n...）替换为单个换行符
text = re.sub(r'\n{3,}', '\n\n', text)
# 9. 清理行首行尾的空白字符
text = re.sub(r'^\s+|\s+$', '', text, flags=re.MULTILINE)
# 10. 移除多余的空格（连续多个空格替换为单个空格）
text = re.sub(r' {2,}', ' ', text)
# 11. 清理首尾空白
text = text.strip()
return text
# 10. 投资日历-事件接口（增强版）
@app.route('/api/calendar/events', methods=['GET'])
@token_required
@pro_or_max_required
def api_calendar_events():
"""投资日历事件接口 - 连接 future_events 表 (修正版)"""
try:
@@ -4895,10 +5188,15 @@ def api_calendar_events():
elif search_query.lower() in str(related_concepts).lower():
highlight_match = 'concepts'
# 清理 Markdown 符号和多余的换行符
cleaned_former = clean_markdown_text(event.former)
cleaned_forecast = clean_markdown_text(event.forecast)
cleaned_fact = clean_markdown_text(event.fact)
event_dict = {
'id': event.data_id,
'title': event.title,
'description': f"前值: {event.former}, 预测: {event.forecast}, 实际: {event.fact}" if event.former or event.forecast or event.fact else "",
'description': f"前值: {cleaned_former}, 预测: {cleaned_forecast}, 实际: {cleaned_fact}" if cleaned_former or cleaned_forecast or cleaned_fact else "",
'start_time': event.calendar_time.isoformat() if event.calendar_time else None,
'end_time': None, # future_events 表没有结束时间
'category': {
@@ -4914,9 +5212,9 @@ def api_calendar_events():
'related_avg_chg': round(related_avg_chg, 2),
'related_max_chg': round(related_max_chg, 2),
'related_week_chg': round(related_week_chg, 2),
'former': event.former,
'forecast': event.forecast,
'fact': event.fact
'former': cleaned_former,
'forecast': cleaned_forecast,
'fact': cleaned_fact
}
# 可选：添加搜索匹配标记
@@ -4948,8 +5246,8 @@ def api_calendar_events():
# 11. 投资日历-数据接口
@app.route('/api/calendar/data', methods=['GET'])
@token_required
@pro_or_max_required
def api_calendar_data():
"""投资日历数据接口"""
try:
@@ -5136,8 +5434,8 @@ def extract_concepts_from_concepts_field(concepts_text):
@app.route('/api/calendar/detail/<int:item_id>', methods=['GET'])
@token_required
@pro_or_max_required
def api_future_event_detail(item_id):
"""未来事件详情接口 - 连接 future_events 表 (修正数据解析) - 仅限 Pro/Max 会员"""
try:
@@ -5372,8 +5670,8 @@ def api_future_event_detail(item_id):
# 13-15. 筛选弹窗接口（已有，优化格式）
@app.route('/api/filter/options', methods=['GET'])
@token_required
@pro_or_max_required
def api_filter_options():
"""筛选选项接口"""
try:
@@ -5952,7 +6250,7 @@ if __name__ == '__main__':
port=5002,
debug=True,
ssl_context=(
'/home/ubuntu/dify/docker/nginx/ssl/fullchain.pem',
'/home/ubuntu/dify/docker/nginx/ssl/privkey.pem'
'/etc/letsencrypt/live/api.valuefrontier.cn/fullchain.pem',
'/etc/letsencrypt/live/api.valuefrontier.cn/privkey.pem'
)
)

5956
app_vx.py.backup Normal file
View File

File diff suppressed because it is too large Load Diff

6318
app_vx.py.optimized_backup Normal file
View File

File diff suppressed because it is too large Load Diff

97
change_pct_fix.py Normal file
View File

@@ -0,0 +1,97 @@
# 这是要替换的涨跌幅计算逻辑（第3729-3738行）
# 计算基于事件时间的涨跌幅（参考 app.py /api/stock/quotes 的实现）
change_pct = None
change_amount = None
current_price = None
try:
# 获取事件时间和当前时间
evt_time = event.start_time if event.start_time else event.created_at
cur_time = datetime.now()
# 获取交易日和时间范围
evt_date = evt_time.date()
evt_time_only = evt_time.time()
market_open = dt_time(9, 30)
market_close = dt_time(15, 0)
# 检查是否是交易日
is_trading_day_result = db.session.execute(text("""
SELECT 1 FROM trading_days WHERE EXCHANGE_DATE = :date
"""), {"date": evt_date}).fetchone()
trading_day = None
start_time = None
end_time = None
if is_trading_day_result:
# 是交易日
if evt_time_only < market_open:
# 盘前 - 使用当日开盘
trading_day, start_time, end_time = evt_date, market_open, market_close
elif evt_time_only > market_close:
# 盘后 - 使用下一交易日
next_day_result = db.session.execute(text("""
SELECT EXCHANGE_DATE FROM trading_days
WHERE EXCHANGE_DATE > :date ORDER BY EXCHANGE_DATE LIMIT 1
"""), {"date": evt_date}).fetchone()
if next_day_result:
trading_day, start_time, end_time = next_day_result[0].date(), market_open, market_close
else:
# 盘中 - 从事件时间到收盘
trading_day, start_time, end_time = evt_date, evt_time_only, market_close
else:
# 非交易日 - 获取下一交易日
next_day_result = db.session.execute(text("""
SELECT EXCHANGE_DATE FROM trading_days
WHERE EXCHANGE_DATE > :date ORDER BY EXCHANGE_DATE LIMIT 1
"""), {"date": evt_date}).fetchone()
if next_day_result:
trading_day, start_time, end_time = next_day_result[0].date(), market_open, market_close
# 如果有有效的交易日且不在未来，查询涨跌幅
if trading_day and trading_day <= cur_time.date():
start_dt = datetime.combine(trading_day, start_time)
end_dt = datetime.combine(trading_day, end_time)
# 查询第一个bar和最后一个bar的价格
price_data = client.execute("""
WITH first_price AS (
SELECT close FROM stock_minute
WHERE code = %(code)s AND timestamp >= %(start)s AND timestamp <= %(end)s
ORDER BY timestamp LIMIT 1
),
last_price AS (
SELECT close FROM stock_minute
WHERE code = %(code)s AND timestamp >= %(start)s AND timestamp <= %(end)s
ORDER BY timestamp DESC LIMIT 1
)
SELECT first_price.close as first_price,
last_price.close as last_price,
(last_price.close - first_price.close) / first_price.close * 100 as change_pct
FROM last_price CROSS JOIN first_price
WHERE EXISTS (SELECT 1 FROM first_price) AND EXISTS (SELECT 1 FROM last_price)
""", {'code': stock.stock_code, 'start': start_dt, 'end': end_dt})
if price_data and price_data[0] and price_data[0][0] is not None:
first_price = float(price_data[0][0])
current_price = float(price_data[0][1])
change_pct = float(price_data[0][2])
change_amount = current_price - first_price
except Exception as e:
print(f"计算事件涨跌幅失败 {stock.stock_code}: {e}")
# 如果ClickHouse没有数据，fallback到原来的逻辑
if change_pct is None:
if latest_trade and prev_trade:
if prev_trade.F007N and prev_trade.F007N != 0:
change_amount = float(latest_trade.F007N) - float(prev_trade.F007N)
change_pct = (change_amount / float(prev_trade.F007N)) * 100
elif latest_trade and latest_trade.F010N:
change_pct = float(latest_trade.F010N)
change_amount = float(latest_trade.F009N) if latest_trade.F009N else None
# 如果还没有当前价格，使用latest_trade
if current_price is None and latest_trade and latest_trade.F007N:
current_price = float(latest_trade.F007N)

540
clickhouse_optimization_guide.py Normal file
View File

@@ -0,0 +1,540 @@
"""
ClickHouse 查询优化方案 - 针对 /api/event/<int:event_id>/related-stocks-detail
问题分析：
1. N+1 查询问题：每只股票执行 3 次独立查询（共 30+ 次）
2. 重复扫描：first_price 和 last_price 需要扫描表两次
3. 缺少批量查询优化
优化方案对比：
┌─────────────┬──────────────┬──────────────┬────────────┐
│ 方案 │ 查询次数 │ 性能提升 │ 实现难度 │
├─────────────┼──────────────┼──────────────┼────────────┤
│ 当前代码 │ N * 3 │ 基准 │ - │
│ 方案1 批量 │ 1 │ 80-90% │ 中等 │
│ 方案2 并行 │ N * 3 (并行)│ 40-60% │ 简单 │
│ 方案3 缓存 │ 减少重复 │ 20-40% │ 简单 │
└─────────────┴──────────────┴──────────────┴────────────┘
"""
# ============================================================================
# 方案 1: 批量查询（推荐）- 将所有股票的查询合并为一次
# ============================================================================
def get_batch_stock_prices_optimized(client, stock_codes, start_datetime, end_datetime):
"""
批量获取多只股票的价格数据（一次查询）
性能对比：
- 旧方案：10 只股票 = 20 次查询（first + last）
- 新方案：10 只股票 = 1 次查询
- 性能提升：约 20 倍
Args:
client: ClickHouse 客户端
stock_codes: 股票代码列表 ['600519.SH', '601088.SH', ...]
start_datetime: 开始时间
end_datetime: 结束时间
Returns:
dict: {
'600519.SH': {
'first_price': 1850.0,
'last_price': 1860.0,
'change_pct': 0.54,
'open': 1850.0,
'high': 1865.0,
'low': 1848.0,
'volume': 1234567,
'amount': 2345678900.0
},
...
}
"""
if not stock_codes:
return {}
# 构建批量查询 SQL（使用 IN 子句）
query = """
SELECT
code,
-- 第一个价格（事件发生时）
anyIf(close, rownum_asc = 1) as first_price,
-- 最后一个价格（当前时间）
anyIf(close, rownum_desc = 1) as last_price,
-- 涨跌幅
(last_price - first_price) / first_price * 100 as change_pct,
-- 涨跌额
last_price - first_price as change_amount,
-- 其他价格信息（取最后一条记录）
anyIf(open, rownum_desc = 1) as open_price,
anyIf(high, rownum_desc = 1) as high_price,
anyIf(low, rownum_desc = 1) as low_price,
anyIf(volume, rownum_desc = 1) as volume,
anyIf(amt, rownum_desc = 1) as amount
FROM (
SELECT
code,
timestamp,
close,
open,
high,
low,
volume,
amt,
-- 正序排名（用于获取第一个价格）
ROW_NUMBER() OVER (PARTITION BY code ORDER BY timestamp ASC) as rownum_asc,
-- 倒序排名（用于获取最后一个价格）
ROW_NUMBER() OVER (PARTITION BY code ORDER BY timestamp DESC) as rownum_desc
FROM stock_minute
WHERE code IN %(codes)s
AND timestamp >= %(start)s
AND timestamp <= %(end)s
)
GROUP BY code
"""
try:
# 执行批量查询
data = client.execute(query, {
'codes': tuple(stock_codes), # ClickHouse IN 需要 tuple
'start': start_datetime,
'end': end_datetime
})
# 格式化结果为字典
result = {}
for row in data:
code = row[0]
result[code] = {
'first_price': float(row[1]) if row[1] else None,
'last_price': float(row[2]) if row[2] else None,
'change_pct': float(row[3]) if row[3] else None,
'change_amount': float(row[4]) if row[4] else None,
'open_price': float(row[5]) if row[5] else None,
'high_price': float(row[6]) if row[6] else None,
'low_price': float(row[7]) if row[7] else None,
'volume': int(row[8]) if row[8] else None,
'amount': float(row[9]) if row[9] else None,
}
print(f"批量查询完成，获取了 {len(result)} 只股票的数据")
return result
except Exception as e:
print(f"批量查询失败: {e}")
return {}
def get_batch_minute_chart_data(client, stock_codes, start_datetime, end_datetime):
"""
批量获取多只股票的分时图数据
Args:
client: ClickHouse 客户端
stock_codes: 股票代码列表
start_datetime: 开始时间
end_datetime: 结束时间
Returns:
dict: {
'600519.SH': [
{'time': '09:30', 'close': 1850.0, 'volume': 12345, ...},
{'time': '09:31', 'close': 1851.0, 'volume': 12346, ...},
...
],
...
}
"""
if not stock_codes:
return {}
query = """
SELECT
code,
timestamp,
open,
high,
low,
close,
volume,
amt
FROM stock_minute
WHERE code IN %(codes)s
AND timestamp >= %(start)s
AND timestamp <= %(end)s
ORDER BY code, timestamp
"""
try:
data = client.execute(query, {
'codes': tuple(stock_codes),
'start': start_datetime,
'end': end_datetime
})
# 按股票代码分组
result = {}
for row in data:
code = row[0]
if code not in result:
result[code] = []
result[code].append({
'time': row[1].strftime('%H:%M'),
'open': float(row[2]) if row[2] else None,
'high': float(row[3]) if row[3] else None,
'low': float(row[4]) if row[4] else None,
'close': float(row[5]) if row[5] else None,
'volume': float(row[6]) if row[6] else None,
'amount': float(row[7]) if row[7] else None
})
print(f"批量获取分时数据完成，获取了 {len(result)} 只股票的数据")
return result
except Exception as e:
print(f"批量获取分时数据失败: {e}")
return {}
# ============================================================================
# 使用示例：替换原来的 for 循环
# ============================================================================
def api_event_related_stocks_optimized(event_id):
"""优化后的端点实现"""
try:
from datetime import datetime
event = Event.query.get_or_404(event_id)
related_stocks = event.related_stocks.order_by(RelatedStock.correlation.desc()).all()
if not related_stocks:
return jsonify({'code': 200, 'data': {'related_stocks': []}})
# 获取 ClickHouse 客户端
client = get_clickhouse_client()
# 计算时间范围（省略交易日计算逻辑，与原代码相同）
event_time = event.start_time if event.start_time else event.created_at
trading_day, start_time, end_time = get_trading_day_and_times(event_time)
start_datetime = datetime.combine(trading_day, start_time)
end_datetime = datetime.combine(trading_day, end_time)
# ✅ 批量查询所有股票的价格数据（只查询 1 次）
stock_codes = [stock.stock_code for stock in related_stocks]
prices_data = get_batch_stock_prices_optimized(
client, stock_codes, start_datetime, end_datetime
)
# ✅ 批量查询所有股票的分时图数据（只查询 1 次）
minute_data = get_batch_minute_chart_data(
client, stock_codes, start_datetime, end_datetime
)
# 组装返回数据
stocks_data = []
for stock in related_stocks:
# 从批量查询结果中获取数据（无需再次查询）
price_info = prices_data.get(stock.stock_code, {})
chart_data = minute_data.get(stock.stock_code, [])
# 获取股票基本信息（这里可以考虑也批量查询）
stock_info = StockBasicInfo.query.filter_by(SECCODE=stock.stock_code).first()
if not stock_info:
base_code = stock.stock_code.split('.')[0]
stock_info = StockBasicInfo.query.filter_by(SECCODE=base_code).first()
stock_data = {
'id': stock.id,
'stock_code': stock.stock_code,
'stock_name': stock.stock_name,
'sector': stock.sector,
'relation_desc': stock.relation_desc,
'correlation': stock.correlation,
'momentum': stock.momentum,
'listing_date': stock_info.F006D.isoformat() if stock_info and stock_info.F006D else None,
'market': stock_info.F005V if stock_info else None,
# 交易数据（从批量查询结果获取）
'trade_data': {
'latest_price': price_info.get('last_price'),
'first_price': price_info.get('first_price'),
'open_price': price_info.get('open_price'),
'high_price': price_info.get('high_price'),
'low_price': price_info.get('low_price'),
'change_amount': round(price_info['change_amount'], 2) if price_info.get('change_amount') else None,
'change_pct': round(price_info['change_pct'], 2) if price_info.get('change_pct') else None,
'volume': price_info.get('volume'),
'amount': price_info.get('amount'),
'trade_date': trading_day.isoformat(),
},
# 分时图数据
'minute_chart': chart_data
}
stocks_data.append(stock_data)
return jsonify({
'code': 200,
'message': 'success',
'data': {
'event_id': event_id,
'event_title': event.title,
'related_stocks': stocks_data,
'total_count': len(stocks_data)
}
})
except Exception as e:
print(f"Error in api_event_related_stocks_optimized: {e}")
return jsonify({'code': 500, 'message': str(e)}), 500
# ============================================================================
# 方案 2: 异步并行查询（适用于无法批量查询的场景）
# ============================================================================
import asyncio
from concurrent.futures import ThreadPoolExecutor
def get_stock_price_async(client, stock_code, start_datetime, end_datetime):
"""单个股票的查询函数（线程安全）"""
# 与原代码相同的查询逻辑
try:
data = client.execute("""
WITH first_price AS (
SELECT close FROM stock_minute WHERE code = %(code)s ...
)
...
""", {'code': stock_code, 'start': start_datetime, 'end': end_datetime})
return stock_code, data
except Exception as e:
return stock_code, None
def get_all_stocks_parallel(client, stock_codes, start_datetime, end_datetime):
"""
并行查询多只股票（使用线程池）
性能对比：
- 串行：10 只股票 * 0.1 秒 = 1 秒
- 并行：max(0.1 秒) = 0.1 秒（10 倍提速）
"""
with ThreadPoolExecutor(max_workers=10) as executor:
# 提交所有查询任务
futures = [
executor.submit(get_stock_price_async, client, code, start_datetime, end_datetime)
for code in stock_codes
]
# 等待所有任务完成
results = {}
for future in futures:
stock_code, data = future.result()
results[stock_code] = data
return results
# ============================================================================
# 方案 3: 添加缓存层（Redis）
# ============================================================================
import redis
import json
redis_client = redis.Redis(host='localhost', port=6379, db=0)
def get_stock_price_with_cache(client, stock_code, start_datetime, end_datetime):
"""
带缓存的查询（适用于历史数据）
缓存策略：
- 历史数据（非当日）：缓存 24 小时
- 当日数据：缓存 1 分钟
"""
from datetime import datetime
# 生成缓存键
cache_key = f"stock_price:{stock_code}:{start_datetime.date()}:{end_datetime.date()}"
# 尝试从缓存获取
cached_data = redis_client.get(cache_key)
if cached_data:
print(f"从缓存获取 {stock_code} 数据")
return json.loads(cached_data)
# 缓存未命中，查询数据库
print(f"从 ClickHouse 查询 {stock_code} 数据")
data = client.execute("""...""", {
'code': stock_code,
'start': start_datetime,
'end': end_datetime
})
# 格式化数据
result = {
'first_price': float(data[0][2]) if data else None,
'last_price': float(data[0][0]) if data else None,
# ...
}
# 写入缓存
is_today = start_datetime.date() == datetime.now().date()
ttl = 60 if is_today else 86400 # 当日数据缓存 1 分钟，历史数据缓存 24 小时
redis_client.setex(cache_key, ttl, json.dumps(result))
return result
# ============================================================================
# 方案 4: ClickHouse 查询优化（索引提示）
# ============================================================================
def get_stock_price_with_hints(client, stock_code, start_datetime, end_datetime):
"""
使用 ClickHouse 特性优化查询
优化点：
1. PREWHERE 子句（提前过滤，减少数据扫描）
2. FINAL 修饰符（如果使用了 ReplacingMergeTree）
3. 分区裁剪（如果表按日期分区）
"""
query = """
SELECT
code,
anyLast(close) as last_price,
any(close) as first_price,
(last_price - first_price) / first_price * 100 as change_pct
FROM stock_minute
PREWHERE code = %(code)s -- 使用 PREWHERE 提前过滤（比 WHERE 快）
WHERE timestamp >= %(start)s
AND timestamp <= %(end)s
GROUP BY code
SETTINGS max_threads = 2 -- 限制线程数（避免资源竞争）
"""
data = client.execute(query, {
'code': stock_code,
'start': start_datetime,
'end': end_datetime
})
return data
# ============================================================================
# 数据库层面优化建议
# ============================================================================
"""
1. 确保 stock_minute 表有以下索引：
- PRIMARY KEY (code, timestamp) -- 主键索引
- INDEX idx_timestamp timestamp TYPE minmax GRANULARITY 3 -- 时间索引
2. 表分区策略（如果数据量大）：
CREATE TABLE stock_minute (
code String,
timestamp DateTime,
...
) ENGINE = MergeTree()
PARTITION BY toYYYYMM(timestamp) -- 按月分区
ORDER BY (code, timestamp)
SETTINGS index_granularity = 8192;
3. 使用物化视图预计算（适用于固定查询模式）：
CREATE MATERIALIZED VIEW stock_minute_summary
ENGINE = AggregatingMergeTree()
PARTITION BY toYYYYMMDD(timestamp)
ORDER BY (code, timestamp)
AS SELECT
code,
toStartOfMinute(timestamp) as minute,
anyLast(close) as last_close,
any(close) as first_close,
...
FROM stock_minute
GROUP BY code, minute;
4. 检查表统计信息：
SELECT
table,
partition,
rows,
bytes_on_disk
FROM system.parts
WHERE table = 'stock_minute';
"""
# ============================================================================
# 性能对比测试
# ============================================================================
def benchmark_query_methods():
"""
性能对比测试
测试场景：查询 10 只股票的价格数据
预期结果：
- 原方案（串行 N+1）：~1000ms
- 方案 1（批量查询）：~50ms（20 倍提速）
- 方案 2（并行查询）：~200ms（5 倍提速）
- 方案 3（带缓存）：~10ms（100 倍提速，第二次请求）
"""
import time
stock_codes = ['600519.SH', '601088.SH', '600276.SH', '000001.SZ', ...]
# 测试方案 1：批量查询
start = time.time()
result1 = get_batch_stock_prices_optimized(client, stock_codes, start_dt, end_dt)
print(f"批量查询耗时: {(time.time() - start) * 1000:.2f}ms")
# 测试方案 2：并行查询
start = time.time()
result2 = get_all_stocks_parallel(client, stock_codes, start_dt, end_dt)
print(f"并行查询耗时: {(time.time() - start) * 1000:.2f}ms")
# 测试原方案（串行）
start = time.time()
result3 = {}
for code in stock_codes:
result3[code] = get_stock_price_original(client, code, start_dt, end_dt)
print(f"串行查询耗时: {(time.time() - start) * 1000:.2f}ms")
# ============================================================================
# 总结与建议
# ============================================================================
"""
推荐实施顺序：
第一步（立即实施）：方案 1 - 批量查询
- 实现难度：中等
- 性能提升：80-90%
- 风险：低
- 时间：1-2 小时
第二步（可选）：方案 3 - 添加缓存
- 实现难度：简单
- 性能提升：额外 20-40%
- 风险：低
- 时间：30 分钟
第三步（长期）：方案 4 - 数据库优化
- 实现难度：中等
- 性能提升：20-30%
- 风险：中（需要测试）
- 时间：2-4 小时
监控指标：
- 查询时间：目标 < 200ms（当前 > 1000ms）
- ClickHouse 查询次数：目标 1-2 次（当前 30+ 次）
- 缓存命中率：目标 > 80%（如果使用缓存）
"""

465
fix_related_stocks_performance.py Normal file
View File

@@ -0,0 +1,465 @@
"""
性能优化补丁 - 修复 /api/event/<int:event_id>/related-stocks-detail 的 N+1 查询问题
使用方法：
1. 将下面的两个函数复制到 app_vx.py 中
2. 替换原来的 api_event_related_stocks 函数
预期效果：
- 查询时间：从 1000-3000ms 降低到 100-300ms
- ClickHouse 查询次数：从 30+ 次降低到 2 次
- 性能提升：约 80-90%
"""
def get_batch_stock_prices(client, stock_codes, start_datetime, end_datetime):
"""
批量获取多只股票的价格数据（只查询一次 ClickHouse）
Args:
client: ClickHouse 客户端
stock_codes: 股票代码列表 ['600519.SH', '601088.SH', ...]
start_datetime: 开始时间
end_datetime: 结束时间
Returns:
dict: {
'600519.SH': {
'first_price': 1850.0,
'last_price': 1860.0,
'change_pct': 0.54,
'change_amount': 10.0,
'open': 1850.0,
'high': 1865.0,
'low': 1848.0,
'volume': 1234567,
'amount': 2345678900.0
},
...
}
"""
if not stock_codes:
return {}
try:
# 批量查询 SQL - 使用窗口函数一次性获取所有股票的数据
query = """
SELECT
code,
first_price,
last_price,
(last_price - first_price) / nullIf(first_price, 0) * 100 as change_pct,
last_price - first_price as change_amount,
open_price,
high_price,
low_price,
volume,
amount
FROM (
SELECT
code,
-- 使用 anyIf 获取第一个和最后一个价格
anyIf(close, rn_asc = 1) as first_price,
anyIf(close, rn_desc = 1) as last_price,
anyIf(open, rn_desc = 1) as open_price,
-- 使用 max 获取最高价
max(high) as high_price,
-- 使用 min 获取最低价
min(low) as low_price,
anyIf(volume, rn_desc = 1) as volume,
anyIf(amt, rn_desc = 1) as amount
FROM (
SELECT
code,
timestamp,
close,
open,
high,
low,
volume,
amt,
-- 正序行号（用于获取第一个价格）
row_number() OVER (PARTITION BY code ORDER BY timestamp ASC) as rn_asc,
-- 倒序行号（用于获取最后一个价格）
row_number() OVER (PARTITION BY code ORDER BY timestamp DESC) as rn_desc
FROM stock_minute
WHERE code IN %(codes)s
AND timestamp >= %(start)s
AND timestamp <= %(end)s
)
GROUP BY code
)
"""
# 执行查询
data = client.execute(query, {
'codes': tuple(stock_codes), # ClickHouse IN 需要 tuple
'start': start_datetime,
'end': end_datetime
})
# 格式化结果
result = {}
for row in data:
code = row[0]
result[code] = {
'first_price': float(row[1]) if row[1] is not None else None,
'last_price': float(row[2]) if row[2] is not None else None,
'change_pct': float(row[3]) if row[3] is not None else None,
'change_amount': float(row[4]) if row[4] is not None else None,
'open_price': float(row[5]) if row[5] is not None else None,
'high_price': float(row[6]) if row[6] is not None else None,
'low_price': float(row[7]) if row[7] is not None else None,
'volume': int(row[8]) if row[8] is not None else None,
'amount': float(row[9]) if row[9] is not None else None,
}
print(f"✅ 批量查询完成，获取了 {len(result)}/{len(stock_codes)} 只股票的数据")
return result
except Exception as e:
print(f"❌ 批量查询失败: {e}")
import traceback
traceback.print_exc()
return {}
def get_batch_minute_chart_data(client, stock_codes, start_datetime, end_datetime):
"""
批量获取多只股票的分时图数据
Args:
client: ClickHouse 客户端
stock_codes: 股票代码列表
start_datetime: 开始时间
end_datetime: 结束时间
Returns:
dict: {
'600519.SH': [
{'time': '09:30', 'open': 1850.0, 'close': 1851.0, 'volume': 12345, ...},
{'time': '09:31', 'open': 1851.0, 'close': 1852.0, 'volume': 12346, ...},
...
],
...
}
"""
if not stock_codes:
return {}
try:
query = """
SELECT
code,
timestamp,
open,
high,
low,
close,
volume,
amt
FROM stock_minute
WHERE code IN %(codes)s
AND timestamp >= %(start)s
AND timestamp <= %(end)s
ORDER BY code, timestamp
"""
data = client.execute(query, {
'codes': tuple(stock_codes),
'start': start_datetime,
'end': end_datetime
})
# 按股票代码分组
result = {}
for row in data:
code = row[0]
if code not in result:
result[code] = []
result[code].append({
'time': row[1].strftime('%H:%M'),
'open': float(row[2]) if row[2] is not None else None,
'high': float(row[3]) if row[3] is not None else None,
'low': float(row[4]) if row[4] is not None else None,
'close': float(row[5]) if row[5] is not None else None,
'volume': float(row[6]) if row[6] is not None else None,
'amount': float(row[7]) if row[7] is not None else None
})
print(f"✅ 批量获取分时数据完成，获取了 {len(result)}/{len(stock_codes)} 只股票的数据")
return result
except Exception as e:
print(f"❌ 批量获取分时数据失败: {e}")
import traceback
traceback.print_exc()
return {}
# ============================================================================
# 优化后的端点函数（替换原来的 api_event_related_stocks）
# ============================================================================
@app.route('/api/event/<int:event_id>/related-stocks-detail', methods=['GET'])
def api_event_related_stocks(event_id):
"""事件相关标的详情接口 - 仅限 Pro/Max 会员（已优化性能）"""
try:
from datetime import datetime, timedelta, time as dt_time
from sqlalchemy import text
import time as time_module
# 记录开始时间
start_time = time_module.time()
event = Event.query.get_or_404(event_id)
related_stocks = event.related_stocks.order_by(RelatedStock.correlation.desc()).all()
if not related_stocks:
return jsonify({
'code': 200,
'message': 'success',
'data': {
'event_id': event_id,
'event_title': event.title,
'related_stocks': [],
'total_count': 0
}
})
# 获取ClickHouse客户端
client = get_clickhouse_client()
# 获取事件时间和交易日（与原代码逻辑相同）
event_time = event.start_time if event.start_time else event.created_at
current_time = datetime.now()
# 定义交易日和时间范围计算函数（与原代码完全一致）
def get_trading_day_and_times(event_datetime):
event_date = event_datetime.date()
event_time_only = event_datetime.time()
market_open = dt_time(9, 30)
market_close = dt_time(15, 0)
with engine.connect() as conn:
is_trading_day = conn.execute(text("""
SELECT 1
FROM trading_days
WHERE EXCHANGE_DATE = :date
"""), {"date": event_date}).fetchone() is not None
if is_trading_day:
if event_time_only < market_open:
return event_date, market_open, market_close
elif event_time_only > market_close:
next_trading_day = conn.execute(text("""
SELECT EXCHANGE_DATE
FROM trading_days
WHERE EXCHANGE_DATE > :date
ORDER BY EXCHANGE_DATE LIMIT 1
"""), {"date": event_date}).fetchone()
return (next_trading_day[0].date() if next_trading_day else None,
market_open, market_close)
else:
return event_date, event_time_only, market_close
else:
next_trading_day = conn.execute(text("""
SELECT EXCHANGE_DATE
FROM trading_days
WHERE EXCHANGE_DATE > :date
ORDER BY EXCHANGE_DATE LIMIT 1
"""), {"date": event_date}).fetchone()
return (next_trading_day[0].date() if next_trading_day else None,
market_open, market_close)
trading_day, start_time_val, end_time_val = get_trading_day_and_times(event_time)
if not trading_day:
return jsonify({
'code': 200,
'message': 'success',
'data': {
'event_id': event_id,
'event_title': event.title,
'event_desc': event.description,
'event_type': event.event_type,
'event_importance': event.importance,
'event_status': event.status,
'event_created_at': event.created_at.strftime("%Y-%m-%d %H:%M:%S"),
'event_start_time': event.start_time.isoformat() if event.start_time else None,
'event_end_time': event.end_time.isoformat() if event.end_time else None,
'keywords': event.keywords,
'view_count': event.view_count,
'post_count': event.post_count,
'follower_count': event.follower_count,
'related_stocks': [],
'total_count': 0
}
})
start_datetime = datetime.combine(trading_day, start_time_val)
end_datetime = datetime.combine(trading_day, end_time_val)
print(f"📊 事件时间: {event_time}, 交易日: {trading_day}, 时间范围: {start_datetime} - {end_datetime}")
# ✅ 批量查询所有股票的价格数据（关键优化点 1）
stock_codes = [stock.stock_code for stock in related_stocks]
print(f"📈 开始批量查询 {len(stock_codes)} 只股票的价格数据...")
query_start = time_module.time()
prices_data = get_batch_stock_prices(client, stock_codes, start_datetime, end_datetime)
query_time = (time_module.time() - query_start) * 1000
print(f"⏱️ 价格查询耗时: {query_time:.2f}ms")
# ✅ 批量查询所有股票的分时图数据（关键优化点 2）
print(f"📈 开始批量查询 {len(stock_codes)} 只股票的分时数据...")
chart_start = time_module.time()
minute_data = get_batch_minute_chart_data(client, stock_codes, start_datetime, end_datetime)
chart_time = (time_module.time() - chart_start) * 1000
print(f"⏱️ 分时数据查询耗时: {chart_time:.2f}ms")
# 组装返回数据（不再需要循环查询）
stocks_data = []
for stock in related_stocks:
# 从批量查询结果中获取数据（O(1) 查找）
price_info = prices_data.get(stock.stock_code, {})
chart_data = minute_data.get(stock.stock_code, [])
# 获取股票基本信息
stock_info = StockBasicInfo.query.filter_by(SECCODE=stock.stock_code).first()
if not stock_info:
base_code = stock.stock_code.split('.')[0]
stock_info = StockBasicInfo.query.filter_by(SECCODE=base_code).first()
# 如果批量查询没有返回数据，尝试使用 TradeData 作为降级方案
if not price_info or price_info.get('last_price') is None:
try:
latest_trade = None
search_codes = [stock.stock_code, stock.stock_code.split('.')[0]]
for code in search_codes:
latest_trade = TradeData.query.filter_by(SECCODE=code) \
.order_by(TradeData.TRADEDATE.desc()).first()
if latest_trade:
break
if latest_trade:
price_info = {
'last_price': float(latest_trade.F007N) if latest_trade.F007N else None,
'first_price': float(latest_trade.F002N) if latest_trade.F002N else None,
'open_price': float(latest_trade.F003N) if latest_trade.F003N else None,
'high_price': float(latest_trade.F005N) if latest_trade.F005N else None,
'low_price': float(latest_trade.F006N) if latest_trade.F006N else None,
'volume': float(latest_trade.F004N) if latest_trade.F004N else None,
'amount': float(latest_trade.F011N) if latest_trade.F011N else None,
'change_pct': float(latest_trade.F010N) if latest_trade.F010N else None,
'change_amount': float(latest_trade.F009N) if latest_trade.F009N else None,
}
except Exception as fallback_error:
print(f"⚠️ 降级查询失败 {stock.stock_code}: {fallback_error}")
stock_data = {
'id': stock.id,
'stock_code': stock.stock_code,
'stock_name': stock.stock_name,
'sector': stock.sector,
'relation_desc': stock.relation_desc,
'correlation': stock.correlation,
'momentum': stock.momentum,
'listing_date': stock_info.F006D.isoformat() if stock_info and stock_info.F006D else None,
'market': stock_info.F005V if stock_info else None,
# 交易数据（从批量查询结果获取）
'trade_data': {
'latest_price': price_info.get('last_price'),
'first_price': price_info.get('first_price'),
'open_price': price_info.get('open_price'),
'high_price': price_info.get('high_price'),
'low_price': price_info.get('low_price'),
'change_amount': round(price_info['change_amount'], 2) if price_info.get('change_amount') is not None else None,
'change_pct': round(price_info['change_pct'], 2) if price_info.get('change_pct') is not None else None,
'volume': price_info.get('volume'),
'amount': price_info.get('amount'),
'trade_date': trading_day.isoformat() if trading_day else None,
},
# 分时图数据
'minute_chart': chart_data
}
stocks_data.append(stock_data)
# 计算总耗时
total_time = (time_module.time() - start_time) * 1000
print(f"✅ 请求完成，总耗时: {total_time:.2f}ms (价格: {query_time:.2f}ms, 分时: {chart_time:.2f}ms)")
return jsonify({
'code': 200,
'message': 'success',
'data': {
'event_id': event_id,
'event_title': event.title,
'event_desc': event.description,
'event_type': event.event_type,
'event_importance': event.importance,
'event_status': event.status,
'event_created_at': event.created_at.strftime("%Y-%m-%d %H:%M:%S"),
'event_start_time': event.start_time.isoformat() if event.start_time else None,
'event_end_time': event.end_time.isoformat() if event.end_time else None,
'keywords': event.keywords,
'view_count': event.view_count,
'post_count': event.post_count,
'follower_count': event.follower_count,
'related_stocks': stocks_data,
'total_count': len(stocks_data),
# 性能指标（可选，调试用）
'performance': {
'total_time_ms': round(total_time, 2),
'price_query_ms': round(query_time, 2),
'chart_query_ms': round(chart_time, 2)
}
}
})
except Exception as e:
print(f"❌ Error in api_event_related_stocks: {e}")
import traceback
traceback.print_exc()
return jsonify({'code': 500, 'message': str(e)}), 500
# ============================================================================
# 使用说明
# ============================================================================
"""
1. 将上面的 3 个函数复制到 app_vx.py 中：
- get_batch_stock_prices()
- get_batch_minute_chart_data()
- api_event_related_stocks()（替换原函数）
2. 重启 Flask 应用：
python app_vx.py
3. 测试端点：
curl http://localhost:5001/api/event/18058/related-stocks-detail
4. 观察日志输出：
✅ 批量查询完成，获取了 10/10 只股票的数据
⏱️ 价格查询耗时: 45.23ms
⏱️ 分时数据查询耗时: 78.56ms
✅ 请求完成，总耗时: 234.67ms
5. 性能对比（10 只股票）：
- 优化前：1000-3000ms（30+ 次查询）
- 优化后：100-300ms（2 次查询）
- 提升：80-90%
6. 如果还是慢，检查：
- ClickHouse 表是否有索引：SHOW CREATE TABLE stock_minute;
- 数据量是否过大：SELECT count() FROM stock_minute WHERE code = '600519.SH';
- 网络延迟：ping ClickHouse 服务器
"""

7
requirements.txt
View File

@@ -1,11 +1,12 @@
Flask==2.3.3
Flask>=3.0.0
Flask-CORS==4.0.0
Flask-SQLAlchemy==3.0.5
Flask-Login==0.6.3
Flask-SQLAlchemy>=3.1.0
Flask-Login>=0.6.3
Flask-Compress==1.14
Flask-SocketIO==5.3.6
Flask-Mail==0.9.1
Flask-Migrate==4.0.5
Flask-Session>=0.5.0
pandas==2.0.3
numpy==1.24.3
requests==2.31.0