560 lines
19 KiB
Python
560 lines
19 KiB
Python
"""趋势突破三阶段扫描管道
|
||
|
||
Phase 1: 全市场批量预筛 — get_daily_all × 5 天 + get_daily_basic → ~300 候选
|
||
Phase 2: 资金流批量过滤 — get_moneyflow_batch → ~80 候选
|
||
Phase 3: 逐股深度分析 — get_stock_daily(120d) + 入场信号 + 供需评分 → ~20 推荐
|
||
"""
|
||
|
||
import logging
|
||
from datetime import datetime, timedelta
|
||
|
||
import numpy as np
|
||
import pandas as pd
|
||
|
||
from app.config import settings
|
||
from app.data.tushare_client import tushare_client
|
||
from app.analysis.technical import add_all_indicators
|
||
from app.analysis.breakout_signals import (
|
||
classify_entry_signal,
|
||
score_supply_demand,
|
||
analyze_volume_pattern,
|
||
EntrySignal,
|
||
)
|
||
|
||
logger = logging.getLogger(__name__)
|
||
|
||
|
||
async def scan_trend_breakout(
|
||
trade_date: str = None,
|
||
market_temp=None,
|
||
hot_sectors: list = None,
|
||
intraday: bool = False,
|
||
) -> list[dict]:
|
||
"""统一趋势突破扫描 — 三阶段管道
|
||
|
||
Returns:
|
||
list[dict] — 每个字典包含:
|
||
ts_code, name, sector, entry_signal_type, entry_signal_score,
|
||
trend_timing_score, supply_demand_score, capital_score,
|
||
main_net_inflow, inflow_ratio, turnover_rate, volume_ratio,
|
||
circ_mv, pe, pb, volume_trend, demand_supply_ratio
|
||
"""
|
||
if not trade_date:
|
||
trade_date = tushare_client.get_latest_trade_date()
|
||
trade_date = _resolve_daily_basic_trade_date(trade_date)
|
||
|
||
logger.info(f"=== 趋势突破扫描 (trade_date={trade_date}) ===")
|
||
|
||
# Phase 1: 批量预筛
|
||
candidates = _bulk_pre_filter(trade_date)
|
||
if candidates.empty:
|
||
logger.info("Phase 1 预筛无候选")
|
||
return []
|
||
|
||
logger.info(f"Phase 1 预筛: {len(candidates)} 只候选")
|
||
|
||
# Phase 2: 资金流过滤
|
||
candidates = _bulk_capital_filter(candidates, trade_date)
|
||
if candidates.empty:
|
||
logger.info("Phase 2 资金过滤后无候选")
|
||
return []
|
||
|
||
logger.info(f"Phase 2 资金过滤: {len(candidates)} 只候选")
|
||
|
||
# Phase 3: 逐股深度分析
|
||
results = await _deep_analysis(candidates, trade_date, market_temp, hot_sectors or [], intraday)
|
||
|
||
logger.info(f"Phase 3 深度分析: {len(results)} 只推荐")
|
||
return results
|
||
|
||
|
||
def _bulk_pre_filter(trade_date: str) -> pd.DataFrame:
|
||
"""Phase 1: 批量预筛
|
||
|
||
利用 get_daily_all × 5 天构建迷你时序，用 5 日趋势初筛。
|
||
合并 get_daily_basic 过滤市值/换手率。
|
||
排除 ST 和次新股。
|
||
"""
|
||
# 获取交易日历
|
||
dates = tushare_client.get_trade_dates()
|
||
if not dates:
|
||
logger.warning("Phase 1: 无法获取交易日历")
|
||
return pd.DataFrame()
|
||
|
||
# 取目标日之前最近 5 个交易日，避免早盘把尚无日频行情的当天纳入窗口。
|
||
eligible_dates = [d for d in dates if d <= trade_date]
|
||
recent_dates = eligible_dates[-5:] if len(eligible_dates) >= 5 else eligible_dates
|
||
if trade_date not in recent_dates:
|
||
recent_dates = recent_dates[-4:] + [trade_date] if len(recent_dates) >= 4 else [trade_date]
|
||
|
||
logger.info(f"Phase 1: 使用交易日 {recent_dates}")
|
||
|
||
# 获取多日全市场数据
|
||
daily_frames = []
|
||
for d in recent_dates:
|
||
df = tushare_client.get_daily_all(d)
|
||
if not df.empty:
|
||
daily_frames.append(df)
|
||
logger.debug(f"Phase 1: {d} 获取到 {len(df)} 只股票数据")
|
||
else:
|
||
logger.warning(f"Phase 1: {d} 无数据")
|
||
|
||
if len(daily_frames) < 2:
|
||
# 至少需要 2 天数据做趋势判断
|
||
if daily_frames:
|
||
logger.info("Phase 1: 仅有1天数据，跳过趋势筛选")
|
||
return _filter_daily_basic(daily_frames[0], trade_date)
|
||
logger.warning("Phase 1: 完全无数据")
|
||
return pd.DataFrame()
|
||
|
||
# 拼接多日数据
|
||
all_daily = pd.concat(daily_frames, ignore_index=True)
|
||
logger.info(f"Phase 1: 拼接后总记录数 {len(all_daily)}")
|
||
|
||
# 计算每只股票的 5 日趋势
|
||
stock_groups = all_daily.groupby("ts_code")
|
||
|
||
# 分步统计
|
||
total_stocks = len(stock_groups)
|
||
count_has_data = 0
|
||
count_main_board = 0
|
||
count_no_limit = 0
|
||
count_positive_return = 0
|
||
count_above_avg = 0
|
||
count_near_high = 0
|
||
|
||
trend_data = []
|
||
for ts_code, group in stock_groups:
|
||
if len(group) < 2:
|
||
continue
|
||
count_has_data += 1
|
||
|
||
group = group.sort_values("trade_date")
|
||
latest = group.iloc[-1]
|
||
|
||
# 排除非主板（只保留 00/30/60 开头）
|
||
code_prefix = ts_code[:2]
|
||
if code_prefix not in ("00", "30", "60"):
|
||
continue
|
||
count_main_board += 1
|
||
|
||
# 排除涨跌停
|
||
pct = latest["pct_chg"]
|
||
if pd.isna(pct) or abs(pct) > 9.5:
|
||
continue
|
||
count_no_limit += 1
|
||
|
||
# 5 日涨幅
|
||
first_close = group.iloc[0]["close"]
|
||
last_close = latest["close"]
|
||
if first_close <= 0:
|
||
continue
|
||
ret_5d = (last_close - first_close) / first_close * 100
|
||
|
||
# 条件：5 日涨幅 > -2%（允许小幅回调） 且 < 20%（未过热）
|
||
if ret_5d <= -2 or ret_5d >= 20:
|
||
continue
|
||
count_positive_return += 1
|
||
|
||
# 5 日均价
|
||
avg_close = group["close"].mean()
|
||
if avg_close <= 0:
|
||
continue
|
||
|
||
# 收盘价 >= 5 日均价 * 0.98（允许略低于均线）
|
||
if last_close < avg_close * 0.98:
|
||
continue
|
||
count_above_avg += 1
|
||
|
||
# 距 5 日高点 < 8%（放宽，不要求紧贴高点）
|
||
high_5d = group["high"].max()
|
||
if high_5d <= 0:
|
||
continue
|
||
dist_from_high = (high_5d - last_close) / high_5d * 100
|
||
if dist_from_high > 8:
|
||
continue
|
||
count_near_high += 1
|
||
|
||
trend_data.append({
|
||
"ts_code": ts_code,
|
||
"close": last_close,
|
||
"pct_chg_today": pct,
|
||
"return_5d": round(ret_5d, 2),
|
||
"dist_from_high_5d": round(dist_from_high, 2),
|
||
"vol_today": latest["vol"],
|
||
})
|
||
|
||
logger.info(
|
||
f"Phase 1 趋势筛选: "
|
||
f"总{total_stocks} → 有数据{count_has_data} → 主板{count_main_board} → "
|
||
f"非涨跌停{count_no_limit} → 5日涨幅>-2%{count_positive_return} → "
|
||
f"近均线{count_above_avg} → 近高点{count_near_high}"
|
||
)
|
||
|
||
if not trend_data:
|
||
return pd.DataFrame()
|
||
|
||
candidates = pd.DataFrame(trend_data)
|
||
|
||
# 合并 daily_basic 过滤（使用更宽松的换手率）
|
||
return _filter_daily_basic(candidates, trade_date)
|
||
|
||
|
||
def _resolve_daily_basic_trade_date(preferred: str) -> str:
|
||
"""盘中当日日频数据未更新时，回退到最近可用交易日。"""
|
||
dates = tushare_client.get_trade_dates()
|
||
if preferred and preferred not in dates:
|
||
dates = sorted([*dates, preferred])
|
||
candidates = [d for d in dates if d <= preferred] if dates else [preferred]
|
||
for date in reversed(candidates[-8:]):
|
||
basic = tushare_client.get_daily_basic(date)
|
||
daily = tushare_client.get_daily_all(date)
|
||
if not basic.empty and not daily.empty:
|
||
if date != preferred:
|
||
logger.info("趋势扫描日频数据回退: %s -> %s", preferred, date)
|
||
return date
|
||
return preferred
|
||
|
||
|
||
def _filter_daily_basic(candidates: pd.DataFrame, trade_date: str) -> pd.DataFrame:
|
||
"""使用 daily_basic 过滤市值、换手率，排除 ST 和次新"""
|
||
basic = tushare_client.get_daily_basic(trade_date)
|
||
if basic.empty:
|
||
logger.warning("Phase 1: daily_basic 无数据")
|
||
return candidates
|
||
|
||
# 股票基本信息（排除 ST 和次新）
|
||
stock_basic = tushare_client.get_stock_basic()
|
||
exclude_codes = set()
|
||
if not stock_basic.empty:
|
||
# ST
|
||
st_codes = set(stock_basic[stock_basic["name"].str.contains("ST", na=False)]["ts_code"])
|
||
exclude_codes.update(st_codes)
|
||
# 次新
|
||
cutoff = (datetime.now() - timedelta(days=settings.min_list_days)).strftime("%Y%m%d")
|
||
new_codes = set(stock_basic[stock_basic["list_date"] > cutoff]["ts_code"])
|
||
exclude_codes.update(new_codes)
|
||
|
||
# Tushare circ_mv 单位是万元，转为亿元
|
||
basic["circ_mv"] = basic["circ_mv"] / 10000
|
||
|
||
# Phase 1 使用更宽松的换手率门槛（2%）以保留更多候选
|
||
min_tr = min(settings.min_turnover_rate, 2.0)
|
||
basic_filtered = basic[
|
||
(basic["circ_mv"] >= settings.min_circ_mv) &
|
||
(basic["circ_mv"] <= settings.max_circ_mv) &
|
||
(basic["turnover_rate"] >= min_tr) &
|
||
(basic["turnover_rate"] <= settings.max_turnover_rate) &
|
||
(~basic["ts_code"].isin(exclude_codes))
|
||
].copy()
|
||
|
||
logger.info(
|
||
f"Phase 1 daily_basic: 全市场{len(basic)}只 → "
|
||
f"过滤后{len(basic_filtered)}只 "
|
||
f"(circ_mv {settings.min_circ_mv}-{settings.max_circ_mv}亿, "
|
||
f"turnover >={min_tr}%)"
|
||
)
|
||
|
||
if basic_filtered.empty:
|
||
return pd.DataFrame()
|
||
|
||
# 合并候选和 basic 数据
|
||
if "circ_mv" in candidates.columns:
|
||
merged = candidates
|
||
else:
|
||
# candidates 没有 basic 列，通过 ts_code 合并
|
||
basic_subset = basic_filtered[["ts_code", "turnover_rate", "volume_ratio",
|
||
"circ_mv", "pe", "pb"]].drop_duplicates("ts_code")
|
||
# 如果 candidates 有 ts_code 列
|
||
if "ts_code" in candidates.columns:
|
||
merged = candidates.merge(basic_subset, on="ts_code", how="inner")
|
||
else:
|
||
# candidates 是 get_daily_all 返回的原始 df
|
||
merged = basic_subset
|
||
|
||
logger.info(f"Phase 1 合并后: {len(merged)} 只候选")
|
||
return merged
|
||
|
||
|
||
def _bulk_capital_filter(candidates: pd.DataFrame, trade_date: str) -> pd.DataFrame:
|
||
"""Phase 2: 资金流批量过滤
|
||
|
||
硬条件：主力净流入 > 0（特大单 + 大单净买为正）
|
||
"""
|
||
mf = tushare_client.get_moneyflow_batch(trade_date)
|
||
if mf.empty:
|
||
return candidates
|
||
|
||
# 计算主力净流入 = (特大单买入-特大单卖出) + (大单买入-大单卖出)
|
||
mf["main_net_inflow"] = (
|
||
(mf["buy_elg_amount"] - mf["sell_elg_amount"]) +
|
||
(mf["buy_lg_amount"] - mf["sell_lg_amount"])
|
||
)
|
||
|
||
# 计算流入比例
|
||
total = (
|
||
mf["buy_elg_amount"] + mf["sell_elg_amount"] +
|
||
mf["buy_lg_amount"] + mf["sell_lg_amount"] +
|
||
mf["buy_md_amount"] + mf["sell_md_amount"] +
|
||
mf["buy_sm_amount"] + mf["sell_sm_amount"]
|
||
)
|
||
mf["inflow_ratio"] = mf["main_net_inflow"] / total.replace(0, np.nan) * 100
|
||
mf["inflow_ratio"] = mf["inflow_ratio"].fillna(0)
|
||
|
||
# 只保留主力净流入 > 0 的
|
||
mf_positive = mf[mf["main_net_inflow"] > 0][["ts_code", "main_net_inflow", "inflow_ratio"]]
|
||
|
||
# 合并候选
|
||
if "ts_code" in candidates.columns:
|
||
merged = candidates.merge(mf_positive, on="ts_code", how="inner")
|
||
else:
|
||
merged = mf_positive
|
||
|
||
# 按主力净流入排序，取 top 100
|
||
if not merged.empty:
|
||
merged = merged.sort_values("main_net_inflow", ascending=False).head(100)
|
||
|
||
return merged.reset_index(drop=True)
|
||
|
||
|
||
async def _deep_analysis(
|
||
candidates: pd.DataFrame,
|
||
trade_date: str,
|
||
market_temp,
|
||
hot_sectors: list,
|
||
intraday: bool,
|
||
) -> list[dict]:
|
||
"""Phase 3: 逐股深度分析
|
||
|
||
对每只候选获取 120 日 K 线，计算技术指标，
|
||
分类入场信号，评分供需关系。
|
||
"""
|
||
import asyncio
|
||
from app.analysis.signals import generate_signals
|
||
from app.analysis.capital_flow import _score_valuation
|
||
|
||
# 获取股票名称映射
|
||
stock_basic = tushare_client.get_stock_basic()
|
||
name_map = {}
|
||
industry_map = {}
|
||
if not stock_basic.empty:
|
||
for _, row in stock_basic.iterrows():
|
||
name_map[row["ts_code"]] = row["name"]
|
||
industry_map[row["ts_code"]] = row.get("industry", "")
|
||
|
||
# 获取多日资金流（用于资金持续性评分）
|
||
mf_history = _get_multi_day_moneyflow(candidates, trade_date)
|
||
|
||
results = []
|
||
total = len(candidates)
|
||
signal_counts = {"breakout": 0, "breakout_confirm": 0, "pullback": 0, "launch": 0, "reversal": 0, "none": 0}
|
||
|
||
for idx, row in candidates.iterrows():
|
||
ts_code = row["ts_code"] if "ts_code" in candidates.columns else ""
|
||
|
||
if not ts_code:
|
||
continue
|
||
|
||
name = name_map.get(ts_code, ts_code)
|
||
sector = industry_map.get(ts_code, "")
|
||
|
||
try:
|
||
# 获取 120 日 K 线
|
||
df = tushare_client.get_stock_daily(ts_code, 120)
|
||
if df.empty or len(df) < 30:
|
||
continue
|
||
|
||
# 添加技术指标
|
||
df = add_all_indicators(df)
|
||
|
||
# 入场信号分类
|
||
entry_signal = classify_entry_signal(df)
|
||
signal_type = entry_signal["signal_type"]
|
||
if signal_type == EntrySignal.NONE:
|
||
signal_counts["none"] += 1
|
||
continue
|
||
signal_counts[signal_type.value] += 1
|
||
|
||
# 趋势评分（内联简化版）
|
||
trend_score = _simple_trend_score(df)
|
||
|
||
# 供需评分
|
||
sd_score = score_supply_demand(df)
|
||
|
||
# 量价模式分析
|
||
vol_pattern = analyze_volume_pattern(df)
|
||
|
||
# 技术信号（复用现有 generate_signals）
|
||
tech_signal = generate_signals(ts_code, name)
|
||
|
||
# 资金流评分
|
||
capital_score = _score_capital(
|
||
row, mf_history.get(ts_code, pd.DataFrame())
|
||
)
|
||
|
||
# 估值评分（作为辅助参考）
|
||
pe = row.get("pe")
|
||
pb = row.get("pb")
|
||
valuation_score = _score_valuation(pe, pb)
|
||
|
||
results.append({
|
||
"ts_code": ts_code,
|
||
"name": name,
|
||
"sector": sector,
|
||
"entry_signal_type": entry_signal["signal_type"].value,
|
||
"entry_signal_score": round(entry_signal["signal_score"], 1),
|
||
"entry_signal_details": entry_signal.get("details", {}),
|
||
"trend_timing_score": round(trend_score, 1),
|
||
"supply_demand_score": round(sd_score, 1),
|
||
"capital_score": round(capital_score, 1),
|
||
"valuation_score": round(valuation_score, 1),
|
||
"technical_score": round(tech_signal.score, 1),
|
||
"position_score": round(tech_signal.position_score, 1),
|
||
"main_net_inflow": row.get("main_net_inflow", 0),
|
||
"inflow_ratio": round(row.get("inflow_ratio", 0), 2),
|
||
"turnover_rate": row.get("turnover_rate"),
|
||
"volume_ratio": row.get("volume_ratio"),
|
||
"circ_mv": row.get("circ_mv"),
|
||
"pe": pe,
|
||
"pb": pb,
|
||
"volume_trend": vol_pattern["volume_trend"],
|
||
"demand_supply_ratio": vol_pattern["demand_supply_ratio"],
|
||
# 技术信号详情（用于生成推荐理由）
|
||
"tech_signal": tech_signal,
|
||
})
|
||
|
||
if len(results) >= settings.top_stock_count:
|
||
break
|
||
|
||
except Exception as e:
|
||
logger.debug(f"深度分析 {ts_code} 失败: {e}")
|
||
continue
|
||
|
||
# 让出控制权，避免阻塞事件循环
|
||
if idx % 10 == 0:
|
||
await asyncio.sleep(0)
|
||
|
||
logger.info(
|
||
f"Phase 3 入场信号分布: "
|
||
f"突破={signal_counts['breakout']} 确认={signal_counts['breakout_confirm']} "
|
||
f"回踩={signal_counts['pullback']} 启动={signal_counts['launch']} "
|
||
f"反转={signal_counts['reversal']} 无信号={signal_counts['none']} "
|
||
f"(共分析{total}只)"
|
||
)
|
||
|
||
return results
|
||
|
||
|
||
def _get_multi_day_moneyflow(candidates: pd.DataFrame, trade_date: str) -> dict[str, pd.DataFrame]:
|
||
"""获取候选股票近 5 日资金流数据（用于资金持续性评分）
|
||
|
||
利用 get_stock_moneyflow 逐只获取（仅对 Phase 2 过滤后的 ~80 只）。
|
||
"""
|
||
result = {}
|
||
|
||
if "ts_code" not in candidates.columns:
|
||
return result
|
||
|
||
for ts_code in candidates["ts_code"].values[:80]:
|
||
try:
|
||
df = tushare_client.get_stock_moneyflow(ts_code, 5)
|
||
if not df.empty:
|
||
result[ts_code] = df
|
||
except Exception:
|
||
pass
|
||
|
||
return result
|
||
|
||
|
||
def _simple_trend_score(df: pd.DataFrame) -> float:
|
||
"""简化趋势评分 (0-100)，用于回退扫描路径"""
|
||
import numpy as np
|
||
score = 0
|
||
last = df.iloc[-1]
|
||
|
||
ma_cols = [c for c in ["ma5", "ma10", "ma20", "ma60"] if c in df.columns]
|
||
if len(ma_cols) >= 4 and not any(pd.isna(last[c]) for c in ma_cols):
|
||
if last["ma5"] > last["ma10"] > last["ma20"] > last["ma60"]:
|
||
score += 50
|
||
elif last["ma5"] > last["ma10"] > last["ma20"]:
|
||
score += 35
|
||
elif last["ma5"] > last["ma20"]:
|
||
score += 20
|
||
|
||
if len(df) >= 20:
|
||
recent = df.tail(20)
|
||
if recent["high"].iloc[10:].max() > recent["high"].iloc[:10].max():
|
||
score += 30
|
||
if recent["low"].iloc[10:].min() > recent["low"].iloc[:10].min():
|
||
score += 20
|
||
|
||
return min(score, 100)
|
||
|
||
|
||
def _score_capital(stock_row: dict | pd.Series, mf_history: pd.DataFrame) -> float:
|
||
"""资金流评分 (0-100)
|
||
|
||
维度：
|
||
- 当日主力净流入规模 (35)
|
||
- 资金持续性 (35): 近 N 日中主力净流入为正的天数
|
||
- 流入比 (15)
|
||
- 量比 (15)
|
||
"""
|
||
score = 0
|
||
|
||
main_net = stock_row.get("main_net_inflow", 0) or 0
|
||
inflow_ratio = stock_row.get("inflow_ratio", 0) or 0
|
||
volume_ratio = stock_row.get("volume_ratio")
|
||
|
||
# 当日主力净流入规模 (35)
|
||
if main_net > 10000:
|
||
score += 35
|
||
elif main_net > 5000:
|
||
score += 28
|
||
elif main_net > 2000:
|
||
score += 20
|
||
elif main_net > 500:
|
||
score += 12
|
||
elif main_net > 0:
|
||
score += 5
|
||
|
||
# 资金持续性 (35)
|
||
if not mf_history.empty and len(mf_history) >= 2:
|
||
positive_days = 0
|
||
for _, r in mf_history.iterrows():
|
||
net = (r.get("buy_elg_amount", 0) - r.get("sell_elg_amount", 0) +
|
||
r.get("buy_lg_amount", 0) - r.get("sell_lg_amount", 0))
|
||
if net > 0:
|
||
positive_days += 1
|
||
total_days = len(mf_history)
|
||
if total_days >= 3 and positive_days >= total_days * 0.7:
|
||
score += 35
|
||
elif total_days >= 2 and positive_days >= total_days * 0.5:
|
||
score += 25
|
||
elif positive_days >= 1:
|
||
score += 12
|
||
else:
|
||
# 只有一天数据且为正（已经通过 Phase 2 过滤）
|
||
score += 15
|
||
|
||
# 流入比 (15)
|
||
if inflow_ratio > 15:
|
||
score += 15
|
||
elif inflow_ratio > 10:
|
||
score += 12
|
||
elif inflow_ratio > 5:
|
||
score += 8
|
||
elif inflow_ratio > 0:
|
||
score += 4
|
||
|
||
# 量比 (15)
|
||
if volume_ratio:
|
||
if volume_ratio > 2.0:
|
||
score += 15
|
||
elif volume_ratio > 1.5:
|
||
score += 10
|
||
elif volume_ratio > 1.0:
|
||
score += 5
|
||
|
||
return min(score, 100)
|